Outside, the whole day, rain has been falling from a monochromatic gray smudge which is both a fog rising from the ground and the clouds overhead.

And there are brown dirt roads with puddled pot holes. Brown slugs are crawling up on the furniture, litter bins and whatever is left lying on the ground

And I only did sleep for three hours, my mind thoughts spinning out of control throughout the night

But then, a big rainbow and a big yellow sun

Green almost yellow illuminated grass

And my friend came visiting with his girlfriend and a small baby child they’d made.

And he was tall just like me with gray streaked beard just like me, long hair in a ponytail and round glasses, just like me

But he’d made these big earrings, you know, and he’s got some tattoos too, and today he was wearing a Slayer t shirt. It’s never been my taste, slayer.

So objectively he was a little bit cooler with this wild style, some would say

And then, as it turns out, we were both running Manjaro Linux

And then they left

There are some people you’ve not talked much for years but it’s still the same when you meet, you know? He’s like that

I think

Thank you for your time

JOURNAL 1er septembre 2026

Je fais une pause. On a eu hier plusieurs inscriptions nouvelles au kendo d'élèves anciens de to sensei qui le suivent, c'est bien ça. Pas mal de nouvelles inscriptions aussi pour notre côté, il faut que je restructure tous les groupes et notamment les débutants. C’est très chouette quand même.

Et puis aujourd'hui j'ai eu la surprise de voir arriver mon Américain, cette fois s'exprimant dans un japonais très correct malgré son foutu accent de Boston. Il a commencé à me remercier de l'avoir forcé à travailler la langue, il est comme ça en règle avec les nouvelles exigences des cinglés de takaichi. Il a bien assimilé les formules de politesse à employer avec un sensei. 😆 Bref il me demande si je suis toujours ok pour le prendre comme élève. Je tiens toujours mes promesses, je l'ai donc inscrit dans un groupe débutant. Pour les cours particuliers on verra après, je peux pas le prendre à son niveau. Voilà donc inscrit Mr Elijah qui se fait appeler Tom en japonais, c’est vrai c’est plus simple 😅. On verra

#dojo

by the wonderful Scott Lava

“This is your last chance. After this, there is no turning back. You take the blue pill – the story ends, you wake up in your bed and believe whatever you want to believe. You take the red pill – you stay in Wonderland and I show you how deep the rabbit hole goes.”

“Father Poseidon, they blinded me. Vengeance, father!“

“I'm letting life hit me until it gets tired. Then I'll hit back. It's a classic rope-a-dope.”

“I want my $200!”

“Clever Girl.”

“Rocky Grace save stars.”

“Here’s ta swimmin’ with bowlegged wemin’.”

“Bah-nahn-ah!”

“Tumble down the rabbit hole.”

“Come in, friend.”

“You’re a good boy, Charlie.”

“Ahhhhhhhhhh!”

“Aim for the legs!”

“Super-man!”

“What a glorious creature.”

Intermission

The internet has changed the way businesses communicate with customers, sell products and provide services. Almost every business needs a website, which has created a growing demand for people who can design attractive, responsive and user-friendly websites.

If you are a student, fresher, working professional or someone planning a career change, web designing can be a useful skill to learn. The good news is that you do not need to be an expert programmer to get started.

With the right learning path, regular practice and practical projects, you can gradually develop the skills needed to work as a web designer or move towards front-end development.

What Is Web Designing?

Web designing is the process of creating the visible and user-facing parts of a website. It includes the layout, structure, colours, typography, spacing, navigation and responsive behaviour of web pages.

A web designer focuses on making a website easy to understand, visually appealing and comfortable to use on different devices.

Modern websites need to work properly on desktops, laptops, tablets and smartphones. Therefore, responsive design is an important part of learning web designing.

Can a Beginner Learn Web Designing?

Yes. Web designing can be learned by beginners with no professional experience in the IT industry.

The best approach is to start with the basics and gradually move towards more advanced technologies.

A simple learning path can look like this:

HTML → CSS → Responsive Design → Bootstrap → JavaScript → Projects → Advanced Front-End Technologies

Starting with fundamentals helps you understand how websites actually work instead of simply copying ready-made designs.

Important Skills for a Web Designer

To become a good web designer, you need a combination of technical, creative and problem-solving skills.

1. HTML

HTML is used to create the basic structure of a web page.

You can use HTML to create headings, paragraphs, images, links, lists, forms, tables and other website elements.

Understanding HTML properly is one of the first steps for anyone beginning a Web Designing Course.

2. CSS

CSS controls the visual appearance of a website.

With CSS, you can work with:

• Colours
• Fonts
• Spacing
• Borders
• Backgrounds
• Layouts
• Animations
• Responsive designs

A strong understanding of CSS can help you create cleaner and more professional website interfaces.

3. Responsive Web Design

People use different devices to browse the internet. A website that looks good on a desktop may not automatically look good on a smartphone.

Responsive web design teaches you how to create layouts that adapt to different screen sizes.

This is an essential skill for modern web designers.

4. JavaScript

JavaScript adds interaction and dynamic behaviour to websites.

For example, JavaScript can be used for menus, sliders, form validation, interactive buttons and other website functions.

Beginners should first become comfortable with HTML and CSS before moving deeply into JavaScript.

5. UI and Visual Design

Technical knowledge alone is not enough.

A web designer should also understand basic design principles such as:

• Colour combinations
• Typography
• Visual hierarchy
• Alignment
• Spacing
• Consistency
• User-friendly layouts

These principles help make websites easier and more enjoyable to use.

6. Problem-Solving

Web designing involves solving problems regularly.

A layout may not look correct on mobile. A button may not work properly. An image may break the design. A section may need to be reorganised.

Learning how to identify and fix these problems is an important professional skill.

Tools and Technologies Web Designers Should Know

The tools you learn can depend on your career goals and the type of websites you want to create.

Some commonly used technologies include:

HTML and CSS

These form the foundation of front-end web design.

Bootstrap

Bootstrap provides ready-to-use components and responsive layout features that can help developers and designers build websites more efficiently.

JavaScript

JavaScript is widely used to add interactive features to web pages.

jQuery

jQuery can simplify certain JavaScript-based front-end tasks and is still useful when working with existing projects.

Tailwind CSS

Tailwind CSS provides a utility-based approach to styling web interfaces and can be useful for creating customised responsive designs.

React JS

React JS is a popular technology for building component-based user interfaces. Learners who want to move beyond basic web designing can explore React after developing a good foundation.

Angular

Angular is another front-end technology used for building structured web applications.

GitHub

GitHub can help you store, manage and collaborate on coding projects. Learning basic version-control workflows can be valuable when working on professional projects.

SITHUB's current Web Designing Course curriculum includes technologies such as HTML, CSS3, Bootstrap 5, JavaScript, React JS, jQuery, Sass, Tailwind, Angular, JSON, AJAX, GitHub and cPanel, with the exact modules varying by course level.

Why Practical Projects Are Important

One of the biggest mistakes beginners make is spending too much time watching tutorials without actually building anything.

Reading about HTML or CSS can help you understand the concepts, but building a website gives you practical experience.

For example, you can create:

• Personal portfolio websites
• Business websites
• Restaurant websites
• Landing pages
• Blog layouts
• Product pages
• Registration forms
• Responsive multi-page websites

Projects help you understand how different skills work together.

They can also become part of your portfolio when you start applying for internships, jobs or freelance projects.

How to Learn Web Designing Step by Step

If you are completely new, follow a structured learning process.

Step 1: Learn HTML Fundamentals

Understand how web pages are structured and practise creating simple pages.

Step 2: Learn CSS

Start creating layouts and learn how to control the visual appearance of your pages.

Step 3: Practise Responsive Design

Make your websites work properly on mobile, tablet and desktop screens.

Step 4: Learn Bootstrap

Use responsive frameworks to understand grids, components and faster website development.

Step 5: Start JavaScript

Learn the basics of programming and use JavaScript to add interaction to your websites.

Step 6: Build Projects

Create complete websites instead of only practising individual topics.

Step 7: Explore Advanced Technologies

Once your foundation is strong, you can explore technologies such as React JS, Angular, Tailwind and other front-end tools.

If you prefer structured learning with trainer guidance and practical project work, you can explore the Web Designing Course in Delhi offered by SITHUB. The course is available in different levels and includes classroom as well as live-online training options.

Web Designing Course: What Should You Look For?

There are many courses available online and offline, so beginners should compare them carefully.

Before choosing a Web Designing Course, check:

• Course syllabus
• Technologies included
• Practical training
• Project work
• Trainer experience
• Classroom or online options
• Doubt-solving support
• Course duration
• Certification
• Career or placement assistance
• Demo class availability

Do not choose a course only because it has a long list of technologies. A good learning experience should also give you enough time to practise and build projects.

Career Opportunities After Learning Web Designing

Learning web designing can open several career paths.

Web Designer

A web designer works on website layouts, visual presentation, responsive design and user-facing elements.

Junior Front-End Developer

With stronger HTML, CSS and JavaScript skills, you can work towards entry-level front-end development roles.

UI Designer

If you enjoy visual design and user interfaces, you can explore UI design and related roles.

WordPress Website Designer

Web designing skills can also be combined with WordPress and other content-management systems to create and customise websites.

Freelance Web Designer

Freelancing allows skilled designers to work with clients on website projects, landing pages and other digital work.

Front-End Developer

If you continue learning JavaScript and modern frameworks such as React or Angular, you can move further towards front-end development.

Your actual career opportunities will depend on your skills, portfolio, experience, communication and the requirements of individual employers.

Web Designing for Students and Freshers

Students and freshers can start learning web designing to develop practical digital skills alongside their education.

A portfolio can be particularly useful for beginners because it allows them to demonstrate what they can actually build.

Instead of simply writing “I know HTML and CSS” on a resume, you can show a responsive website, landing page or portfolio project that demonstrates your ability.

Web Designing for Working Professionals

Working professionals may also learn web designing to add a new technical skill to their existing experience.

For example, a digital marketer, content creator, business owner or graphic designer may benefit from understanding how websites are structured and designed.

Learning web designing can also make it easier to communicate with developers and understand the technical side of website projects.

How to Build a Web Design Portfolio

A portfolio should demonstrate your skills clearly.

Start with a few quality projects instead of creating dozens of unfinished websites.

Your portfolio can include:

1. Personal portfolio website

2. Business website

3. Product landing page

4. Responsive blog design

5. Restaurant or service website

6. Interactive JavaScript project

For each project, explain what you created, which technologies you used and what problems you solved.

Make sure your portfolio is responsive and works properly on mobile devices.

Is Web Designing a Good Career Choice?

Web designing can be a good career option for people who enjoy both creativity and technology.

It offers different directions. You can focus on visual design, front-end development, freelance work, website creation or continue learning advanced technologies.

However, completing a course alone is not enough.

Your practical skills, projects, portfolio and ability to solve real problems are important when looking for professional opportunities.

Tips to Become a Better Web Designer

Follow these simple habits while learning:

• Practise every day.
• Build projects instead of only watching tutorials.
• Learn responsive design properly.
• Study professional websites for inspiration.
• Keep your portfolio updated.
• Learn basic SEO and website performance concepts.
• Understand how users interact with websites.
• Keep learning new technologies.
• Ask questions when you get stuck.
• Review and improve your old projects.

Final Thoughts

Becoming a web designer is a step-by-step process. You do not need to learn every technology at once.

Start with HTML and CSS, understand responsive design, learn JavaScript, build practical projects and gradually explore advanced front-end technologies.

The most important thing is consistent practice.

Whether you are a student, fresher, working professional or career changer, a structured Web Designing Course can help you follow a clear learning path and develop practical skills.

Start with the basics, build real projects and keep improving your portfolio. With the right combination of learning and practice, you can create a strong foundation for a career in web designing and front-end technologies.

#webdesignerskills #webdesign #webdesigning

Le barbecue en inox à poser séduit de plus en plus les amateurs de cuisson au charbon de bois et au feu de bois. Compact, robuste et simple à installer, il permet de profiter d’une cuisson authentique en extérieur sans devoir construire un barbecue fixe ni entreprendre des travaux importants.

Il peut être installé sur un plan de travail adapté, dans une cuisine extérieure, sur une structure maçonnée ou sur un support suffisamment stable et résistant à la chaleur. Cette polyvalence en fait une solution intéressante aussi bien pour une terrasse que pour un jardin ou une cuisine d’été.

Mais quels sont réellement les avantages d’un barbecue en acier inoxydable à poser ? Comment choisir sa dimension et quelles caractéristiques faut-il examiner avant de l’acheter ?

Qu’est-ce qu’un barbecue en inox à poser ?

Un barbecue en inox à poser est un appareil de cuisson extérieur conçu pour être installé directement sur une surface stable, plane et incombustible.

À la différence d’un barbecue fixe en pierre ou en brique, il ne possède généralement ni socle maçonné ni cheminée. Il offre donc davantage de liberté au moment de concevoir ou de réorganiser son espace extérieur.

Les modèles les plus performants ne se limitent pas à une simple cuve métallique. Ils peuvent intégrer :

• un foyer protégé par des matériaux réfractaires ;
• une grille de cuisson en acier inoxydable ;
• plusieurs hauteurs de cuisson ;
• un tiroir ou un système de récupération des cendres ;
• des protections latérales et arrière ;
• une deuxième grille pour maintenir les aliments au chaud.

Cette conception permet d’associer la simplicité d’un barbecue compact aux performances thermiques d’un foyer plus traditionnel.

1. Une installation simple, sans travaux importants

Le premier avantage du barbecue à poser est sa facilité d’installation.

Il n’est pas nécessaire de construire une structure complète en pierre ou en brique. Il suffit de prévoir un support suffisamment solide, stable et résistant à la chaleur. Le barbecue peut ainsi être intégré à un plan de travail extérieur existant ou posé sur une structure préparée spécialement pour lui.

Cette solution convient particulièrement :

• aux terrasses où l’espace est limité ;
• aux cuisines extérieures modulaires ;
• aux maisons de vacances ;
• aux utilisateurs qui ne souhaitent pas réaliser une construction définitive ;
• aux projets où le barbecue doit pouvoir être déplacé ultérieurement.

Il faut toutefois tenir compte de son poids. Un véritable barbecue en inox avec intérieur réfractaire est beaucoup plus lourd qu’un petit barbecue métallique portable. Son support doit donc être correctement dimensionné et installé sur un sol stable.

2. L’acier inoxydable facilite l’entretien

L’acier inoxydable est apprécié pour sa résistance, son aspect moderne et sa facilité d’entretien. Il convient particulièrement bien aux équipements de cuisine et de cuisson, car sa surface est relativement simple à nettoyer.

Après utilisation, il est généralement conseillé de :

1. laisser le barbecue refroidir complètement ;

2. retirer les cendres ;

3. brosser la grille de cuisson ;

4. nettoyer les surfaces avec de l’eau tiède et un produit doux ;

5. sécher l’inox avec un chiffon propre.

Les éponges métalliques agressives, l’eau de Javel et les produits fortement chlorés sont à éviter. Ils peuvent altérer la finition de l’inox et provoquer des marques superficielles.

L’acier inoxydable ne signifie pas pour autant qu’aucun entretien n’est nécessaire. En extérieur, les dépôts de graisse, les cendres humides, le sel présent dans l’air et la pollution peuvent laisser des traces. Un nettoyage régulier et l’utilisation d’une housse adaptée contribuent à préserver durablement l’aspect du barbecue.

3. Un foyer réfractaire pour mieux conserver la chaleur

Tous les barbecues en inox à poser ne présentent pas les mêmes performances. L’un des éléments les plus importants à vérifier est la présence d’un revêtement réfractaire à l’intérieur du foyer.

Les briques ou plaques réfractaires absorbent une partie de la chaleur produite par les braises et la restituent progressivement. Elles contribuent ainsi à :

• stabiliser la température de cuisson ;
• mieux répartir la chaleur ;
• protéger les parois métalliques du contact direct avec le feu ;
• limiter les pertes thermiques ;
• prolonger la durée de vie de la structure.

Un foyer réfractaire offre également une cuisson plus régulière qu’une simple cuve métallique très fine. Cette différence devient particulièrement perceptible lors des cuissons prolongées ou lorsque plusieurs séries d’aliments doivent être grillées.

Les matériaux réfractaires peuvent développer, au fil des utilisations, de petites fissures superficielles dues aux variations de température. Ce phénomène est généralement normal et ne compromet pas nécessairement leur fonctionnement.

4. Une cuisson possible au charbon de bois ou au bois

Un barbecue à poser bien conçu peut fonctionner avec du charbon de bois, mais aussi avec de petites bûches adaptées aux dimensions du foyer.

Le charbon de bois est pratique lorsqu’on recherche :

• un allumage relativement rapide ;
• une chaleur régulière ;
• une gestion simple des braises ;
• une cuisson traditionnelle des viandes, poissons et légumes.

Le bois apporte quant à lui une expérience plus proche de la cuisson au feu. Il produit des flammes puis des braises, avec des arômes qui peuvent varier selon l’essence utilisée.

Il est préférable de choisir du bois sec, non traité et adapté à la cuisson alimentaire. Les bois peints, vernis, agglomérés ou issus de meubles ne doivent jamais être brûlés dans un barbecue.

Avant de commencer la cuisson, il faut attendre que les flammes les plus fortes diminuent et qu’un lit de braises suffisamment stable se soit formé.

5. Plusieurs dimensions selon le nombre de convives

La largeur du barbecue doit être choisie en fonction du nombre de personnes, du type d’aliments préparés et de l’espace disponible.

Les barbecues inox à poser BRASAOVENS sont notamment proposés dans plusieurs dimensions, généralement autour de 60, 70 ou 80 cm de largeur.

Barbecue de 60 cm

Le format de 60 cm est adapté aux espaces plus compacts et aux repas familiaux. Il permet de préparer plusieurs aliments simultanément tout en restant relativement facile à intégrer dans une petite cuisine extérieure.

Barbecue de 70 cm

Le format de 70 cm représente un bon compromis entre encombrement et capacité. Il convient aux familles qui reçoivent régulièrement des invités et souhaitent disposer d’une surface de cuisson plus confortable.

Barbecue de 80 cm

Le format de 80 cm est destiné aux repas plus importants. Sa grande grille permet de mieux organiser les différentes zones de cuisson et de préparer davantage d’aliments en même temps.

Il faut également vérifier la profondeur du barbecue. Une grille large mais trop peu profonde peut être moins pratique qu’une surface de cuisson correctement proportionnée.

6. Une ou deux grilles : quelle configuration choisir ?

Certains barbecues à poser sont disponibles avec une seule grille de cuisson, tandis que d’autres proposent une seconde grille placée à un niveau supérieur.

Une grille unique suffit pour une utilisation classique. Elle permet de cuire directement les aliments au-dessus des braises et offre une grande simplicité d’utilisation.

La deuxième grille apporte toutefois un confort supplémentaire. Elle peut servir à :

• maintenir les aliments cuits au chaud ;
• éloigner certaines préparations de la chaleur directe ;
• cuire plus doucement des légumes ou du pain ;
• organiser plusieurs niveaux de cuisson ;
• augmenter la capacité utile du barbecue.

Cette configuration est particulièrement intéressante lorsqu’on prépare un repas pour plusieurs personnes et que tous les aliments ne nécessitent pas le même temps de cuisson.

7. Une intégration facile dans une cuisine extérieure

Le design sobre de l’inox s’intègre facilement à différents styles d’aménagement. Il peut être associé à un plan de travail en pierre, en béton, en granit, en céramique ou en acier inoxydable.

Avant l’installation, il faut néanmoins prévoir :

• une surface parfaitement stable ;
• des matériaux incombustibles autour du foyer ;
• une ventilation suffisante ;
• un espace de sécurité par rapport aux murs, meubles et végétaux ;
• une zone dégagée au-dessus du barbecue ;
• un accès pratique pour retirer les cendres.

Le barbecue ne doit pas être utilisé dans un espace fermé ou insuffisamment ventilé. La combustion du charbon de bois ou du bois produit du monoxyde de carbone, un gaz invisible et dangereux.

Il est également déconseillé de placer le barbecue directement sous une pergola inflammable, un store, des branches ou tout autre élément susceptible d’être affecté par la chaleur et les fumées.

8. Une alternative compacte au barbecue fixe

Le barbecue en inox à poser et le barbecue maçonné ne répondent pas exactement aux mêmes besoins.

Le modèle à poser est particulièrement intéressant lorsque l’on recherche :

• une installation rapide ;
• un encombrement réduit ;
• un design contemporain ;
• une intégration dans un meuble ou un plan de travail ;
• une solution pouvant être déplacée ou réinstallée.

Le barbecue fixe en pierre ou en brique est davantage adapté à un projet permanent. Il peut comporter une cheminée, de grands plans de travail, un évier ou même un four à bois.

Pour un aménagement plus traditionnel et définitif, il est possible de consulter les barbecues fixes en brique et en pierre de FR-Barbecue.

Le choix dépend donc principalement de l’espace disponible, du style recherché et du niveau d’intégration souhaité.

9. Quels critères vérifier avant l’achat ?

Avant de choisir un barbecue en inox à poser, il est utile de comparer plusieurs caractéristiques.

La qualité de la grille

La grille est directement en contact avec les aliments. Une grille robuste en inox, avec des barreaux suffisamment épais, offre généralement une meilleure stabilité et une plus grande durabilité.

Le revêtement réfractaire

Un foyer protégé par des matériaux réfractaires conserve mieux la chaleur et protège davantage la structure métallique.

La stabilité de l’ensemble

Le barbecue doit être suffisamment rigide et lourd pour rester stable pendant son utilisation. Les poignées, grilles et éléments amovibles doivent également être solidement conçus.

La récupération des cendres

Un tiroir à cendres facilite considérablement le nettoyage. Il permet de retirer les résidus sans devoir démonter tout l’intérieur du foyer.

Les dimensions réelles

Il faut vérifier les dimensions extérieures, mais aussi la surface utile de la grille. L’espace nécessaire autour du barbecue doit également être pris en compte dès la conception du support.

La garantie

Une garantie clairement indiquée constitue un bon repère sur le niveau de confiance accordé au produit. Les modèles proposés par BRASAOVENS bénéficient d’une garantie pouvant atteindre cinq ans.

Questions fréquentes sur les barbecues en inox à poser

Peut-on laisser un barbecue en inox dehors toute l’année ?

L’inox résiste bien à une utilisation extérieure, mais il est préférable de protéger le barbecue de la pluie, de l’humidité persistante et des dépôts atmosphériques. Une housse respirante et adaptée à ses dimensions est recommandée.

Peut-on poser le barbecue sur une table en bois ?

Ce n’est pas conseillé. Le barbecue doit être installé sur une surface stable, résistante à la chaleur et non combustible. La chaleur transmise par le foyer peut endommager ou enflammer un support inadapté.

L’inox peut-il changer de couleur avec la chaleur ?

Oui. Une exposition répétée à des températures élevées peut provoquer des nuances dorées, bleutées ou brunâtres sur certaines zones. Cette coloration thermique est un phénomène naturel et n’empêche pas le fonctionnement du barbecue.

Faut-il nettoyer les briques réfractaires ?

Il n’est généralement pas nécessaire de les laver avec de l’eau ou des produits chimiques. Une fois le foyer froid, les résidus secs peuvent être retirés délicatement à l’aide d’une brosse.

Quelle dimension choisir pour une famille ?

Un modèle de 60 cm convient généralement à un usage familial. Les formats de 70 ou 80 cm offrent davantage de confort lorsque l’on reçoit régulièrement plusieurs invités ou que l’on souhaite créer différentes zones de cuisson.

En conclusion

Le barbecue en inox à poser constitue une solution polyvalente pour cuisiner au charbon de bois ou au feu de bois sans construire immédiatement un barbecue fixe.

Son installation simple, son entretien accessible, son foyer réfractaire et les différentes dimensions disponibles lui permettent de s’adapter à de nombreux espaces extérieurs. Installé sur un support adapté, il peut devenir l’élément central d’une cuisine d’été compacte, fonctionnelle et durable.

Pour comparer les dimensions, les configurations à une ou deux grilles et les modèles actuellement disponibles, découvrez la sélection de barbecues en inox à poser BRASAOVENS.

I’m a little bit sad that I even Got to this point at all. I found myself searching up what makes someone a high value man, and everything that came up was something about red pill ideology. I honestly don’t really know too much about red pill or any of the pill stuff, but I know that it’s one of those like Andrew Tate or incel adjacent schools of thought. I thought about how at the gym I see all of these people around me and I feel like all of these really beautiful women wouldn’t be interested in getting to know me or even getting my foot in the door because I do not look like someone who is on steroids. I also do not outwardly have flashy things that show wealth. And I guess this thought has started to slowly whittle at me. I kind of felt like I need to do something else to change that, like spend more money in stupid status symbols. I even watched a video today that was recommended to me about different signifiers of status, and I think the fact that that was even recommended to me even though it was a satire channel that I like to watch feels a bit worrying. I also felt like maybe I need to update my wardrobe more, or maybe I need to worry about my appearance more and start to die diet or lose weight even though I am already at a pretty low body fat percentage for the norm. I feel like this is just a step towards dysmorphia again, and I don’t even know if this changes anything. I guess I just feel really frustrated because I have an incredibly good job, I own a house, I drive a nice car, I am attractive, I am physically very fit, I am emotionally mature, I am kind, I am popular, I am well liked, and outgoing, I feel like I am all of these wonderful things that I have strive to be, and it isn’t really working. It feels really frustrating to see random science or people talking about how what really matters to them is a man that is kind and funny and even having women that thirst over me when I don’t try to. But I haven’t really been able to find a relationship with someone that I feel matches me and is a good fit. And I know that I have been rejecting a lot of different people on dates and I’ve had to warn the person that I’m currently friends with benefits With to not catch feelings for me or let me know if they do so that we can address that, because I am not interested in dating them. And I guess that is success, but I wish that instead I was finding a really nice partner. And the more I think about it the more I feel like I have explored a lot of these paths and really the most important thing I can do is put myself in situations where I am exposed to more people. And pass this I really should stop trying to optimize our focus towards this because it is such a sparse goal that there really isn’t much I can hope for.

Hello,

I am a researcher with an architectural background, exploring how to build socially sustainable neighbourhoods and cities. As part of this journey, I studied differences in how older men and women engage in intergenerational interactions in community spaces in Seoul, Korea.

I'm interested in how to create social infrastructure that connects people from different backgrounds and strengthens local community networks. I believe this requires planning physical social infrastructure and community initiatives together, which is the kind of work I hope to pursue.

I would like to share my thoughts on this and what I've learned from the field.

Happy to be here and connect with you guys.

Best, Seunghyun Dany Song

Starcraft Co 21- Defection

Citadel had given StarCraft Co. an army. It had not given them a mandate. In the weeks after Restoration, most of the surviving Terran military accepted the emergency command structure because somebody had to coordinate the war and StarCraft Co. had demonstrated that it could. That agreement became considerably less certain once the immediate panic subsided. Planetary governments began asking how long the emergency arrangement would last. Admirals who had readily accepted operational orders during the Citadel campaign became less enthusiastic when those orders involved transferring ships away from their home sectors. Industrial consortiums wanted guarantees regarding contracts and mineral concessions. Mercenary companies wanted to know which government would actually be paying them six months from now. Former UIC officials who had survived the decapitation began assembling their own networks and quietly discussing whether StarCraft Co. had exceeded the authority anyone had actually granted it. None of those groups necessarily wanted the renegades to win. They simply were not convinced StarCraft Co. would.

The renegades understood that distinction and exploited it aggressively. Their diplomats traveled wherever their armies could not. They offered autonomy to fringe governors, mineral rights to corporations, pardons to UIC officers, contracts to mercenary organizations and protected smuggling routes to criminal syndicates willing to supply them. They presented themselves less as insurgents than as the inevitable successor to a failed government. Their argument was simple: the Council was dead, the UIC had collapsed, and StarCraft Co. was a mercenary corporation pretending that battlefield competence amounted to political legitimacy. Why sacrifice another million people resisting what would eventually happen anyway? Some factions rejected them immediately. Others listened. Several opened negotiations while continuing to publicly recognize StarCraft Co.'s emergency command. The commander responded first with diplomacy. Delegations were sent. Contracts were honored. Governors were promised considerable local autonomy once the war ended. Mercenary groups were offered standardized pay and access to StarCraft Co. logistics. Corporations received assurances that nationalization was not planned. Independent Protoss tribes were approached through Ara intermediaries rather than treated automatically as hostile. It worked often enough that StarCraft Co. kept doing it. It did not work universally.

The most dangerous gathering took place on Callisto's Reach, a sparsely populated fringe world whose primary virtue was distance. It possessed little industry, no major fleet base and only a handful of settlements scattered across enormous mineral deserts broken by canyon systems kilometers deep. Choosing it was deliberate. Nearly every substantial StarCraft Co. formation was already committed elsewhere, either holding the post-Citadel coalition together or containing renegade offensives that had accelerated while the Terran command structure remained unsettled. A conventional intervention on Callisto's Reach would require withdrawing ships from somewhere more important, precisely the calculation the renegades expected StarCraft Co. to make.

The conference was less a unified political movement than a collection of people who had all reached the same conclusion for different reasons. The largest former-UIC contingent called itself the Continuity Directorate, a grand title for seven surviving administrators, three generals and Admiral Hector Vale, whose Seventh Reserve Fleet had avoided the fighting at Citadel because most of its ships had been undergoing refit. Vale did not consider himself a traitor. He considered StarCraft Co. dangerous. In his view, the UIC had lost its leadership, not its legitimacy, and handing emergency command to a private military corporation merely replaced one crisis with another. His officers brought disciplined marines, several Goliath platoons and four siege tanks painted in the old blue-and-white markings of the UIC fleet. Their equipment was maintained immaculately. Their problem was that half of them were still waiting for somebody to explain whether they were attending a peace conference or preparing for another war.

The criminal delegations were considerably less conflicted. The Orion Passage Cooperative controlled smuggling routes through three sparsely policed systems and had spent twenty years moving everything from contraband stimulants to stolen reactor components. “Cooperative” was mostly a joke. Its crews flew patched-together Wraiths with mismatched armor panels, drove Vultures modified for cargo hauling and carried weapons from half a dozen manufacturers. Its leader, Mara Venn, had no ideological attachment to either side. The UIC had occasionally arrested her people; the renegades were offering exclusive transit rights and protection from Zerg predation. Beside them camped the Red Knives, a mercenary outfit descended from colonial security contractors whose reputation came from being inexpensive, aggressive and willing to work in places respectable companies would not. Their infantry wore crimson markings across otherwise standardized armor, and their officers favored rapid Vulture attacks supported by Goliaths instead of the slower combined-arms doctrine used by regular armies. They were experienced enough to be dangerous and undisciplined enough to believe experience made them better than they were.

The more serious mercenary force was Helix Security, which had once competed with StarCraft Co. for government contracts before the war elevated one company and nearly bankrupted the other. Helix had survived by becoming the private army of whoever could pay. Its commander, Colonel Darius Pell, had spent twenty-two years in the UIC Marines before leaving over a promotion dispute and had never forgiven institutions for failing to recognize his own estimation of his talents. His troops were professionals. They maintained overlapping fields of fire, fielded medics at proper ratios and traveled with their own Goliaths, Wraiths and mobile artillery. Pell had not yet signed with the renegades, but he had entered the conference convinced that StarCraft Co.'s rise had made continued neutrality impossible. If he was going to serve somebody else's new order, he intended to choose the side with the better chance of surviving it.

The corporate delegations brought forces built around entirely different assumptions. Caldera Extraction Group owned mines, refineries and processing stations across six systems, along with a security arm designed primarily to suppress piracy and labor unrest. Their equipment was excellent because Caldera could afford excellent equipment. Their soldiers were not. Most had never fought anything more dangerous than raiders. Their armored battalion possessed new-model Goliaths, pristine Vultures and imported targeting suites that looked impressive during demonstrations. Its commanders had been selected as much for corporate reliability as battlefield experience. Sagan Interstellar Logistics, by contrast, had almost no conventional ground force but operated a small private aerospace wing because its shipping network had spent years defending freighters from pirates. Its pilots were good. Its Wraiths were old. Its mechanics were exceptional. Sagan's executives were not particularly sympathetic to the renegades, but their business depended upon functioning trade routes, and they were willing to negotiate with anybody who could keep those routes open.

Several independent Protoss tribes had come as observers rather than allies. The largest, the Khal-Tor, possessed only a few dozen warriors but controlled a chain of isolated settlements near Terran space and had spent decades avoiding entanglement in human politics. They distrusted the Ara almost as much as they distrusted the renegades. Another group, the Serevai, had already fought alongside Terran mercenaries elsewhere in the war and believed the collapse of the UIC might finally create room for permanent Protoss enclaves without Terran interference. Neither group had pledged anything. Both wanted to see what the renegades actually controlled before deciding whether negotiation was safer than opposition.

The renegades came last and made no attempt to appear conciliatory. Their delegation arrived behind several companies of veteran infantry supported by siege tanks, Goliaths and Wraiths, enough force to dominate the conference militarily without appearing to occupy it. Their commander, Colonel Jorek Malk, had defected from the UIC early in the war and possessed the unnerving calm of someone who believed history had already vindicated him. He spoke of reconciliation, regional autonomy and an end to unnecessary bloodshed. Then he brought the delegates outside and showed them the real argument.

He had brought a Zerg brood.

Overlords descended over the desert while the delegates watched. Drones established creep beyond the conference perimeter. Hatcheries swelled from the ground. Hydralisk dens and spawning pools followed. Thousands of zerglings gathered in the valleys, with hydralisks behind them and mutalisks circling overhead. The demonstration was carefully staged. Renegade officers walked through the growing brood without being attacked. Zerg organisms moved when ordered and stopped when ordered. A hydralisk formation turned in unison at a hand signal relayed through the renegade control system. A wave of zerglings surged toward an empty ridge, stopped before reaching the observers and settled almost motionless onto the creep. An ultralisk strain was conspicuously absent, but nobody watching required one to understand the implication. The renegades were not merely promising military protection. They were displaying an army they could reproduce anywhere given enough biomass and time.

StarCraft Co. intercepted fragments of the negotiations through several sources. The discussions were farther along than expected. The Red Knives had tentatively agreed to a renegade contract. Pell was negotiating Helix's incorporation as an autonomous brigade rather than merely hired security. Caldera wanted mineral rights to two systems the renegades did not yet control and, remarkably, Malk appeared willing to grant them. Admiral Vale and the Continuity Directorate were discussing something more dangerous: creation of a provisional Terran authority that would recognize the renegades as a legitimate belligerent, suspend offensive operations and open negotiations over a permanent partition of Terran space. Vale believed he was preventing civilizational collapse. Malk understood perfectly well that a recognized partition would give the renegades time to consolidate every world they already held. If the conference ended successfully, the agreements themselves would matter less than the precedent. Every undecided faction in Terran space would know that negotiations with the renegades were possible, respectable and perhaps even profitable.

The commander looked at the available force list. Almost everything useful had somebody else's crisis attached to it. Pull a battlecruiser group from the western systems and three colonies would lose their only deterrent. Move the fleet from Citadel and the renegades might retake the station. Redirect planetary formations and governors who had only just accepted StarCraft Co.'s authority would conclude that their own worlds were expendable. The operation had to be small enough that nobody noticed what was missing until it was already over.

A fast cruiser became the answer. The ship was not large enough to fight a fleet and was never intended to. Its internal volume had been stripped for speed, electronic warfare equipment and troop capacity. It carried a newly graduated commando company, one of the first formations produced after the Aiur deployment proved the enhancement program could be expanded beyond a few dozen experimental soldiers. These commandos were not identical to the original twenty-four. The augmentation process had been standardized, several dangerous procedures removed, armor manufacturing simplified and weapons redesigned for production beyond laboratory quantities. They remained enormously expensive, but StarCraft Co. could now create them in numbers that mattered. The commander assigned them a mission that required exactly what the program had been designed to provide: arrive somewhere conventional forces could not, destroy something larger than themselves and leave before the enemy could respond properly.

The Ara contributed a small force as well. Their involvement was less politically delicate than at Citadel; there was no need to conceal it this time. Zealots, dragoons and several high templar joined the cruiser aboard dedicated transports. The independent Protoss tribes attending the renegade conference complicated matters, but the Ara were unsympathetic to tribes willing to make accommodations with Terrans controlling Zerg. One Ara commander examined intelligence images showing the renegade demonstration and gave his assessment without hesitation. “They have made the same error as the rebels on Aiur.” The StarCraft Co. officer beside him asked, “Which one?” “They believe control makes the abomination less offensive.”

The cruiser entered the system without announcing itself and descended behind the planet's larger moon before making its final approach. The conference occupied a broad plateau surrounded by canyon country. Malk's sensor stations covered the open desert, Sagan Wraiths patrolled the upper atmosphere and Red Knife scouts regularly swept the easiest ground approaches. StarCraft Co. ignored all of them. Transports descended into a canyon nearly forty kilometers from the meeting site, flying below the surrounding ridgelines until the final kilometers. The commandos landed in darkness, advanced through the canyon network on foot and reached the plateau through a narrow ravine that reconnaissance maps considered impassable to conventional infantry. By dawn they were less than three kilometers from the Orion Passage camp.

The first indication anyone had that StarCraft Co. was present was one of Mara Venn's Goliaths falling over. Its cockpit erupted under concentrated commando fire before its pilot could rotate the weapons platform. Two modified Vultures beside it accelerated toward the source and disappeared beneath grenades fired into the canyon mouth. Commandos emerged behind them in squads, moving far faster than conventional armored infantry could comfortably track. Orion crews reacted like smugglers rather than soldiers. Some rushed toward their ships. Others opened fire from cargo stacks or tried to move contraband vehicles out of the line of attack. Several pilots attempted to launch their Wraiths without waiting for ground crews, and one clipped another aircraft while lifting from the improvised strip. The cruiser fired precision missiles into the runway only after the first fighters began climbing. Three Wraiths were destroyed on the ground. Two escaped rather than fight.

The commandos did not chase them. They turned immediately on the Red Knives, whose camp sat less than a kilometer away and had already begun deploying. The difference was obvious. Red Knife marines formed actual fire teams, Goliaths moved into supporting positions, and Vultures raced toward side canyons looking for the commandos' flanks. Their commander, Edda Rourke, responded to the surprise with exactly the aggression that had made her company useful for twenty years. She ordered a counterattack before she knew how many enemies she faced. StarCraft Co. exploited it. One commando platoon pinned the Red Knife infantry outside effective rifle range while another dropped into the ravine beneath the western flank. Rourke's Vultures accelerated through the canyon expecting to emerge behind the attackers and instead found commandos waiting above them. Armor-piercing fire killed the lead vehicles. The wreckage blocked the others. Grenades finished the column.

Rourke attempted to reorganize around a prefab warehouse complex, but her troops had been deployed to attack rather than defend it. Goliaths found their lines of fire obstructed by their own vehicles. Marines fired from cargo barriers never designed as fortifications. The commandos kept moving, concentrating entire squads against individual sections instead of fighting the whole camp simultaneously. Within forty minutes the Orion Passage security force had scattered and the Red Knives had ceased functioning as an organized company. StarCraft Co. had several wounded but no dead. Rourke escaped aboard a transport with perhaps a quarter of her personnel. Mara Venn did not even attempt to rally her crews. She transmitted a profanity-laced notice that she was withdrawing from “everyone's damn civil war” and fled toward orbit.

StarCraft Co. did not stop to occupy either camp. Weapons depots were destroyed. Communications systems were disabled. Vehicles that could not immediately be used were sabotaged. Fuel dumps were punctured rather than detonated because the commandos had no interest in advertising their exact position with a column of fire visible across the plateau. Representatives who fled were allowed to flee unless they attempted to organize resistance. The objective was not to kill everyone attending the conference. It was to remove the factions most likely to provide Malk with additional troops once the real battle began.

The Ara landed as the commandos approached the Zerg perimeter. They arrived directly into combat. Mutalisks spotted the descending transports and attacked before they reached the ground. Dragoons fired upward from open ramps while the ships were still moving. One transport took enough damage that its pilot put it down several kilometers short of the intended landing zone, disgorging zealots into the desert while the surviving aircraft continued toward the commandos. By the time the two forces linked, the renegades understood that the raid was not aimed at one delegation. StarCraft Co. was heading directly for the brood.

That was deliberately reckless. The brood was the centerpiece of the conference and the physical proof behind every renegade promise. Leaving it intact while attacking lesser allies would accomplish almost nothing. Destroying it in front of the delegates would undermine the entire argument they had traveled to hear. The combined formation therefore went straight into the Zerg concentration before Malk could fully reorganize it. Commandos opened corridors through zergling packs with controlled automatic fire while zealots occupied anything that reached close combat. Dragoons concentrated on hydralisk groups and biological structures. High templar remained protected inside the formation until the Zerg massed heavily enough to justify exposing them, then dropped storms across entire approach lanes. The first storm annihilated several hundred zerglings before they could close. A second struck hydralisks packed between two hatcheries. Commandos pushed through the resulting gap and began firing directly into the structures.

The Zerg responded violently. Mutalisks descended into the formation and killed three commandos before dragoons and cruiser-launched Wraiths drove them away. Burrowed hydralisks emerged inside the advancing line and caught one squad from two directions, killing four more and wounding nearly a dozen before Ara zealots reached them. Another commando died pulling a wounded medic out of the path of a zergling rush. By the time the formation entered the central creep field, where the brood could reinforce faster than the attackers could safely advance, eight commandos were dead and more than twenty had suffered wounds serious enough to require eventual evacuation. Nobody suggested slowing down. StarCraft Co. had not sent a company because it expected the mission to be safe. It had sent a company because the hive had to stop functioning before Malk could bring the rest of the conference forces into the battle.

StarCraft Co. did not attempt to clear every organism. Intelligence teams had identified three primary hive clusters and the largest concentrations of overlords, and those became the only objectives that mattered. The first hatchery complex fell to concentrated dragoon fire while commandos held a semicircle around it. The second required the high templar. Zerg organisms were packed so densely across the approach that conventional fire could not create enough space. Two storms opened the line, and the commandos sprinted through before it closed. They planted shaped charges directly against the biological structure and withdrew while zealots kept the surviving defenders occupied. The hatchery ruptured from within. The third cluster attempted to evacuate. Overlords began moving away from the battlefield with drones and other organisms beneath them. The cruiser descended just far enough above the canyon rim to acquire them and opened fire. Wraiths followed. Overlords fell across the desert.

The effect was not instantaneous, but it was obvious. With the central hive structures destroyed and much of the overlord network gone, the brood's coordination degraded. Zerglings continued attacking whatever was nearest. Hydralisks reacted to immediate threats but stopped maneuvering as coherent groups. Mutalisks scattered toward distant contacts. Malk's controllers still possessed influence over portions of the brood, but the demonstration army he had brought to impress the conference was becoming exactly what his enemies had always claimed it was: a mass of dangerous organisms whose usefulness depended on a fragile command structure. The diplomats watching from the plateau did not need a military education to recognize the transformation. Twenty minutes earlier the renegades had possessed an obedient alien army. Now zerglings were chasing abandoned vehicles, hydralisks were firing at distant Wraiths and mutalisks were circling terrain with no apparent objective.

That was when Malk committed the renegade regulars. They came from the northern plateau in disciplined formation, veteran marines at the front with Goliaths moving among them and siege tanks following along prepared lanes. Unlike the Red Knives, they did not rush. Unlike Caldera's troops, they did not hesitate. These were soldiers who had fought through the same war as StarCraft Co., many of them former UIC personnel who had defected years earlier and accumulated experience commanding Zerg alongside conventional forces. Their armor was mismatched but heavily modified, with additional plating over joints most vulnerable to hydralisk spines and heat shielding adapted from campaigns against Protoss weaponry. Their tank crews knew enough not to deploy until they had secure observation. Their Wraith pilots came in low over the ridges rather than presenting themselves to the cruiser.

Pell's Helix Security joined them almost immediately. That surprised nobody. Pell had spent the entire conference demanding assurances that his company would retain operational autonomy under the renegades; the attack gave him an opportunity to demonstrate why that autonomy was worth granting. His formations were smaller than Malk's but just as orderly. Helix medics moved behind advancing infantry. Two Goliath platoons formed a mobile anti-air screen. A battery of portable missile turrets began deploying around the conference approaches. Pell personally took command of the western flank and ordered his troops to hold rather than chase the commandos. He understood what StarCraft Co. wanted because he had competed against them long enough to recognize their habits.

Caldera's private army joined for an entirely different reason. Its executives had already signed preliminary mineral agreements and feared losing them if the renegades were driven off-world. Their battalion rolled out with immaculate equipment and almost no combat cohesion. New-model Goliaths advanced in parade-perfect spacing until the first damaged Zerg organisms wandered across their route and forced the column to break formation. Vulture crews transmitted constantly asking for revised orders. Infantry officers kept stopping to re-form their squads. Their commander requested instructions from the corporate delegation twice during the opening five minutes of combat. The machines were first-rate. The organization underneath them was not.

Sagan Interstellar refused. Its aerospace commander informed both Malk and StarCraft Co. that Sagan personnel had come to negotiate shipping access, not die in somebody else's succession crisis. Their Wraiths climbed above the battlefield and established a defensive orbit around their own transports. Admiral Vale did something more consequential. Malk requested that the Continuity Directorate commit its troops to the counterattack. Vale refused as well. “We came to discuss an armistice,” he transmitted. “We did not come here to become your auxiliary force.” His marines remained in their camp with weapons ready but did not move. The distinction mattered. Callisto's Reach was no longer one Terran army fighting another. It was a coalition beginning to fracture while the shooting was still underway.

Even without Vale and Sagan, StarCraft Co. and the Ara were badly outnumbered. The obvious move was retreat. The commander ordered an advance. Malk expected them to pull back toward the southern canyon where their transports could extract them. His tanks began deploying to dominate that route. Helix Goliaths spread outward to prevent air evacuation. Infantry formations widened into a closing arc. StarCraft Co. instead pushed north toward the conference complex itself, where the terrain narrowed between two escarpments and the opposing forces had not yet positioned their artillery.

The fast cruiser joined the battle properly for the first time. It could not remain exposed for long, but it did not need to. Electronic warfare systems flooded Malk's command network while the ship made a single low pass over the battlefield. Renegade Wraiths attempted to intercept it, and unlike the Orion pilots they pressed the attack. One reached firing range and put missiles into the cruiser's aft armor before StarCraft Co. escorts drove it away. The cruiser ignored the fighters long enough to fire into the lead siege-tank formation. Two tanks died. A third lost a track assembly. Then the ship climbed before Helix Goliaths could concentrate on it.

The ground force reached the escarpments first and immediately occupied the high ground. Ara dragoons established firing positions beneath them while commandos converted natural rock shelves into overlapping fields of fire. The first renegade infantry formations entered the narrow approach and met concentrated fire from above. Armor penetrators punched through powered suits at distances where ordinary marines would have struggled to hit reliably. Medics kept wounded commandos in the line. When Malk's Goliaths attempted to push forward, dragoons focused on one at a time until the machines collapsed and blocked the route behind them. Malk reacted intelligently and ordered his siege tanks onto surrounding ridges rather than feeding more units through the center. The Ara high templar moved before the tanks finished deploying. Psionic storms struck crews and infantry escorts, forcing vehicles either to undeploy or remain inside the storms long enough to be destroyed by concentrated fire.

Pell recognized the trap sooner than Malk did. Rather than attack the center, Helix infantry began working through side ravines in small units, searching for routes onto the commandos' flanks. They found one. A Helix platoon emerged unexpectedly above the western position and poured fire into a commando fireteam before the rest of the company could react. Three commandos died holding the ravine while their squad fell back. Pell pushed harder, correctly judging that StarCraft Co. could not afford to leave the flank open, and brought a Goliath close enough to fire directly into the reverse slope. The commandos killed the machine, but not before it forced them off another position. StarCraft Co. shortened its perimeter. Helix advanced again. The movement looked increasingly like a retreat, but each contraction pulled Pell farther from Malk's artillery and compressed both enemy formations into the same narrowing piece of terrain.

Caldera's battalion provided the opening. Its commander saw the withdrawing commandos and interpreted the movement literally. Against Pell's advice, he ordered the western ridge seized before StarCraft Co. could establish another line. Caldera marines advanced behind three Goliaths and reached the crest with almost no resistance. The commandos withdrew down the reverse slope. Caldera followed.

Ara zealots were waiting beneath them.

The first Goliath fired downward and killed one zealot before two others reached it. The machine disappeared beneath plasma blades. Caldera's infantry tried to back away, only to discover that the troops behind them were still advancing. The ridge became a compression point. Zealots tore into the front while commandos fired from both sides. Officers shouted contradictory orders over the corporate channel. Some squads attempted to retreat. Others continued firing. One Goliath backed over two marines trying to get clear. The fight lasted less than four minutes. When the survivors finally broke, they did not retreat toward the renegades. They ran toward the Caldera landing field.

Malk committed his reserve to prevent the western flank from collapsing. That was the point StarCraft Co. had been waiting for. The cruiser had spent the previous twenty minutes moving around the far side of the canyon system below enemy sensor horizons. Now it reappeared directly behind the renegade artillery line. Wraith escorts hit Goliaths first. The cruiser fired everything else into the tanks. At almost the same moment, the commandos stopped withdrawing and counterattacked from the escarpments. Dragoons advanced beneath them. The remaining high templar dropped storms behind the lead renegade formations rather than directly on them, cutting reinforcement and retreat at the same time.

The counterattack cost StarCraft Co. heavily. A commando squad moving through a side ravine encountered one of Pell's remaining Helix units at almost point-blank range. Both sides opened fire before either could maneuver. Two commandos were killed immediately, and several more wounded before zealots forced Helix back. Elsewhere, a renegade siege tank fired at a StarCraft Co. position less than a hundred meters away. The shell struck short, killing several of Malk's own marines but also catching the edge of a commando team moving between rock formations. One died outright. By that point the company's losses had reached fourteen killed. Thirty-one more were seriously wounded, although several remained in the fight under medication and the support of their enhancements because evacuation was impossible until the enemy line broke.

For several minutes the battle became close enough that nobody had a useful picture of it. Terran infantry fought among damaged vehicles and Zerg corpses. Goliaths fired at Wraiths overhead while commandos shot through their leg assemblies from the ground. Helix squads withdrew by bounds rather than routing, covering one another with a professionalism that made them considerably harder to kill than Caldera's soldiers. Pell personally organized a defensive line around two damaged Goliaths and held it long enough for dozens of his wounded to escape. Malk's veterans attempted to re-form around their remaining tanks, but the command network had been disrupted by the cruiser and their formations had become tangled with retreating corporate troops. Zealots moved through smoke and wreckage into soldiers trained to fight Zerg and conventional Terrans but far less comfortable dealing with Protoss at arm's length. The renegades still had more troops. They no longer had a coherent battle line. StarCraft Co. did.

That decided it. Malk recognized the collapse before most of his subordinates and ordered a withdrawal toward the northern landing field. His regulars obeyed. Helix followed several minutes later, Pell refusing to sacrifice the company simply to make Malk's retreat look orderly. StarCraft Co. did not pursue far enough to lose formation. Fourteen dead commandos were already fourteen more than the company could casually replace, and thirty-one seriously wounded represented an enormous investment in personnel, augmentation and training that would take months to recover. The operation had succeeded because the commandos could absorb losses without losing cohesion, not because those losses did not matter. For an ordinary infantry company, the casualty rate might have produced combat ineffectiveness. The surviving commandos still held their heavy weapons, maintained squad communications and were prepared to continue fighting if ordered. The commander had no intention of testing how much farther that resilience extended when the strategic objective was already within reach.

The cruiser destroyed several renegade transports before they could launch but ignored others once Malk's withdrawal became irreversible. Pell's Helix aircraft were deliberately left alone after his ground formations disengaged; StarCraft Co. wanted the company beaten, not exterminated. Caldera Security mostly surrendered. Their executives initially demanded protection for corporate property before somebody reminded them that their private army had just participated in an armed attack against StarCraft Co. One of the executives stopped arguing after that. Sagan's Wraiths remained in their neutral orbit until the firing subsided, then requested permission to land medical shuttles for wounded personnel from all factions. Permission was granted immediately.

By the time the firing around the central plateau ended, the political gathering had ceased to exist in any meaningful sense. Malk fled with what remained of his security force. Mara Venn was already gone. Rourke and the surviving Red Knives disappeared off-world without waiting to discover whether their preliminary renegade contract was still valid. Caldera's representatives remained because most of their transport crews had surrendered. Pell did something nobody expected. He walked into the StarCraft Co. perimeter himself with two officers, removed his sidearm and handed it grip-first to the nearest commando. “Contract's terminated,” he said. The commando looked at the weapon, then at him. “With them?” Pell glanced toward the retreating renegade transports. “With anyone who needs a Zerg brood to make his argument.”

Admiral Vale arrived shortly afterward. He had not fired on StarCraft Co., which meant StarCraft Co. had not fired on him. The distinction was deliberate on both sides. Vale looked across the battlefield at burning Caldera vehicles, collapsing Zerg structures and commandos carrying their own wounded toward the extraction point. “Fourteen?” he asked after hearing the casualty report. “Fourteen dead,” the officer answered. “Thirty-one serious.” Vale looked toward the remnants of the brood. “For this?” The officer understood what he meant. “For everyone watching this.”

The commander joined the surviving delegates only by secure transmission. There was no theatrical demand for unconditional loyalty and no speech about treason. He offered three choices. They could leave Terran space entirely and take no further part in the war. They could return home and remain neutral, provided they supplied no material support to the renegades. Or they could join the emergency Terran command and fight. Vale asked what happened to anyone who chose the renegades. The StarCraft Co. officer standing among the delegates looked toward the battlefield. Burning Goliaths surrounded the plateau. What remained of Malk's demonstration brood wandered through the desert without coherent direction. Renegade transports were disappearing toward the horizon. “You saw what happens.”

Most of them joined, although the manner of joining revealed almost as much as the decision. Vale agreed to place the Seventh Reserve Fleet under unified operational command while explicitly reserving the Continuity Directorate's right to contest the postwar political settlement. The commander accepted without argument. Sagan Interstellar volunteered its shipping network and aerospace escorts in exchange for guaranteed neutrality in future disputes over commercial policy. StarCraft Co. accepted that too. Caldera's executives swore loyalty with considerably more enthusiasm than credibility and immediately began negotiating to keep their existing mineral concessions. Pell requested a formal StarCraft Co. contract for Helix Security, including guarantees that his personnel would retain their own command structure. The request was taken under review. The Khal-Tor Protoss departed without making any pledge, but their representative privately informed the Ara that they would no longer consider agreements with factions using controlled Zerg. The Serevai went farther and offered scouts for future operations against renegade brood sites.

Some of those decisions came from conviction. Most came from arithmetic. The renegades had brought a Zerg army to Callisto's Reach to demonstrate that resistance was futile. StarCraft Co. had arrived with one fast cruiser, one commando company and a small Protoss detachment and destroyed it. The factions present had watched Orion Passage scatter almost immediately, the Red Knives discover that aggression was not a substitute for discipline, Caldera learn that expensive hardware did not manufacture soldiers, Helix fight well and still lose, and Malk's veterans lose a battle despite overwhelming numerical superiority. They had also watched StarCraft Co. distinguish between enemies instead of indiscriminately destroying everyone present. Vale had remained neutral and survived untouched. Sagan had refused the renegades and been allowed to provide medical aid. Pell had broken contact and been permitted to withdraw. StarCraft Co. had not come to Callisto's Reach to massacre dissenters. It had come to establish that joining the enemy was a military decision with military consequences.

News of the battle spread faster than StarCraft Co.'s diplomats could have arranged. Survivors carried it. Corporate communications carried it. Mercenary channels carried it, and the mercenary version was particularly effective because nobody in that world believed official casualty reports but everyone believed Pell when he told former colleagues that the new commandos were “the worst infantry problem I've ever seen.” The independent Protoss observers carried their own version. Within days, groups that had been quietly negotiating with the renegades began contacting StarCraft Co. instead. Several mercenary companies requested formal contracts. Two governors publicly rejected ceasefire proposals they had been considering the week before. A shipping consortium volunteered transport capacity to the emergency command. Former UIC officers who had been attempting to create independent regional commands began discussing integration.

StarCraft Co. continued offering diplomacy, contracts, autonomy agreements and guarantees to factions willing to cooperate. The offers were considerably more persuasive now. There was never an official proclamation regarding Callisto's Reach. Nobody announced that neutrality was forbidden or that disagreement with StarCraft Co. constituted treason. In fact, Vale's continued political opposition made that impossible to claim even if StarCraft Co. had wanted to. The warning was narrower and therefore more credible: remaining neutral was possible; arguing about the future was possible; refusing StarCraft Co.'s political leadership was possible. Actively lending soldiers, money, ships or legitimacy to the renegades was something else.

There were now two sides with a realistic chance of winning the war. One commanded the Zerg and promised that accommodating it was safer than resistance. The other had just found a conference on a forgotten fringe world, inserted a few hundred troops through terrain everyone else considered inaccessible, destroyed the Zerg force brought there as a symbol of renegade power, defeated a coalition several times its size and left the factions that had stayed out of the fight almost entirely alone. Fourteen commandos had died proving the point, and thirty-one more had been carried home badly wounded. StarCraft Co. considered the price high. Everyone else considered the message unmistakable.

Siding with the renegades was no longer hedging against the future.

It was choosing the losing side before the war was over.

I should be heading to sleep on this last day of August. I want to, but I also want to write. I don't know what I want to write. I know that I don't want my notes to automatically hyperlink words that I write to something that I'm not linking it to. That makes me want to stop writing.

Today was not unexpected. There was both potential bad news and potential good news. I knew something along such lines would be coming. I was just unsure of what my response would be. There are times when I wonder if I am too understanding. It is a quality that I believe is very important to have, but it comes at a cost.

It is possible that it is a very positive thing and come January I am going to be extremely happy. It is also possible that I am just naive and too understanding. I think I've also had such a heavy dose of pragmatism invade my fairy tale mindset that maybe, just maybe, my ability to hope and believe is more than a little bit tarnished.

Perhaps, I have had too many curtains pulled back and have gotten to see what lay on the other side too often to still have the innocent, dreaming mindset I once had. Even while quite cynical. I really don't know anymore. I really don't know anymore...

I need to make a phone call tomorrow that I don't want to make. It reveals way too much of myself to persons that I don't particularly want to. It is a necessity, but I am digging my heels in hard. I have been for some time. I have tried to figure out why. Why am I so reluctant?

I think, perhaps, it is because when I got the opportunity to tell the only person in the world that it really had anything to do with, his response was...disappointing. I think that is often the case with such things. I am still glad that I told him. It was a regret I had carried for decades.

Tomorrow is different. Tomorrow I need to acknowledge it to strangers. If I'm being extremely honest with myself, I'm still debating whether or not I'm going to follow through. I recognize that I really do need to, but I do not want to. Who will win out? Need or want?

I probably need to want to give a damn. I think in general, part of me gives a damn way too much, and part of me could not give a damn at all. Such is the danger in accepting “reality.”

© 2026 AnOublietteofThought. All rights reserved.

“You don’t move on from grief, you learn to live beside it.”

Grief is such a complicated thing. It doesn’t just live in loss, It hides in the quiet spaces after. It sits in the chair where someone used to sit, In the text you want to send but can’t. In the songs you skip because they still know your heart too well.

People think grief is all tears and sadness. But sometimes it’s just emptiness. A dull ache that follows you from morning to night. A feeling that something is missing, Even when you can’t name what it is anymore.

I’ve come to realize that grief isn’t just about losing someone. It’s also about losing who I was when they were still here. The version of me that believed in certain things. In forever, In forgiveness, In the idea that love could fix anything. When all of that crumbled, I had to face myself in a way I never had before.

It’s weird how grief makes you grow. Not because you want to, But because standing still hurts too much. I used to fight it, The crying,

The loneliness. I thought if I could just stay busy enough. Or keep pretending I was fine. The pain would eventually get tired and move on. But grief doesn’t leave when you ignore it. It only settles deeper into your bones.

Now I let myself feel it. I cry when I need to. I write when I can’t speak. I sit in the ache instead of running from it.

Because somewhere in that stillness. I’ve started to find something softer, Not quite peace, but something close. Maybe acceptance isn’t about saying “I’m okay with what happened.” Maybe it’s about saying, “I’m okay with me, even after what happened.”

I miss parts of my old life, And there are still days I wish I could go back and fix it all. But if I’m honest, I don’t want to be that same version of myself anymore. I’ve outgrown the places where I used to hide, And I’m slowly learning that losing people doesn’t mean losing myself.

Some days, acceptance looks like forgiveness. Other days, it’s just breathing through the hurt until it fades a little. And on the hardest days, It’s reminding myself that healing doesn’t mean forgetting, It means remembering without losing myself in the memory.

I don’t know if grief ever fully ends. I think it just changes shape. It becomes less about what I’ve lost and more about what I’ve learned, That love doesn’t disappear just because things didn’t turn out the way I hoped.

So, I keep walking forward. Carrying both the love and the loss. They belong to each other, They both belong to me. And maybe that’s what healing really is, not letting go, But learning how to hold differently.

(NEXT – CHAPTER 5)

Somewhere tonight, a person will leave a hospital holding a sheet of paper they can read perfectly and should not trust.

The paper will look right. Clean typography, the hospital's logo in the corner, sentences running in grammatical order in the reader's own language. It will have been produced in under a second at a cost that rounds to nothing, and handed over by a clinician who cannot read a word of it, to a patient with no way of comparing it to anything. Everyone will believe the document is complete, because it looks complete, and looking complete is what this technology does flawlessly.

We know what these documents can contain, because researchers have checked. In 2021, a team led by Breena Taira at the University of California, Los Angeles took twenty commonly used emergency department discharge statements, ran them through Google Translate into seven widely spoken languages, and had native speakers evaluate the output. Their paper in the Journal of General Internal Medicine contains two findings that ought to have ended the argument about unsupervised machine translation in clinical settings. The first is statistical. The second is a pair of words: in the Armenian output, ibuprofen was rendered as “anti-tank missile”, and in Chinese, the anticoagulant Coumadin came out as “soybean”.

These are not subtle failures. A bilingual reader would spot them in a second. The problem is that no bilingual reader was ever going to look, because the entire economic point of the exercise was to avoid paying one.

That is the shape of the thing. Over roughly three years, the cost of producing a translation has fallen by close to three orders of magnitude. The cost of finding out whether it is correct has not moved, because verification still requires a person who reads both languages, sitting down with both documents, and thinking. Generation collapsed. Verification did not. Everything that follows, from the restructuring of a seventy billion dollar industry to the question of who is answerable when the instructions are wrong, falls out of that single asymmetry.

The Arithmetic That Broke in Only One Direction

Google's Cloud Translation service prices standard neural machine translation at twenty United States dollars per million characters, the first half a million each month free. DeepL's professional interface has been offered at around five and a half dollars per million. A thousand words of English is roughly six thousand characters, so translating it costs about twelve cents at the top rate and about three at the bottom. Professional human translation of the same thousand words sits, in 2026, at one hundred and fifty to three hundred dollars, with rate surveys clustering human work between ten and thirty cents per source word.

Generation is effectively free. But consider what verification requires and the picture inverts. To know whether a thousand-word translation is correct, a qualified bilingual reader must read the source, read the target, and compare them for meaning, register, omission, addition and negation. That is not faster than translating. It is often slower, because the reviewer must reconstruct the translator's reasoning as well as the author's. And it requires precisely the scarce human capability the machine was bought to replace.

Verification now costs more than production by a factor of roughly a thousand. If a factory could stamp a component for a tenth of a cent but testing each one cost a hundred dollars, nobody would call the testing optional. They would redesign the process, or admit on the record that they were shipping untested parts. Translation took the third option, never written down: shipping untested parts while behaving as though they had been tested, because they look exactly like the tested ones.

A second scarcity sits beneath the first. The binding constraint is not money but qualified bilingual attention. Only so many people can competently review a discharge instruction in Armenian, a tenancy notice in Tigrinya, or a pesticide warning in Hmong. Machine translation multiplied the volume of text requiring review by orders of magnitude while leaving the pool of reviewers where it was. The bottleneck is a population, not a budget line.

Failure With No Symptoms

The asymmetry persists because its consequences are invisible at the point of delivery. Machine translation does not fail the way software fails. It does not crash, return an error code, garble characters or leave a field blank. It fails by producing a well-formed sentence that means something other than what the source said.

The canonical example comes from a 2019 study in JAMA Internal Medicine by Elaine Khoong, Eric Steinbrook, Cortlyn Brown and Alicia Fernandez, who passed one hundred sets of emergency department discharge instructions comprising 647 sentences through Google Translate into Spanish and Chinese. Eight per cent of the Spanish sentences and nineteen per cent of the Chinese sentences were inaccurate. More to the point, two per cent of the Spanish and eight per cent of the Chinese carried potential to cause clinically significant harm.

One sentence read, in English, “hold the kidney medicine until you have a chance to speak with your kidney doctor”. The Chinese output instructed the patient to keep taking the kidney medicine, and the Spanish said much the same. “Hold”, in clinical English, means stop. The machine took it in its ordinary sense of retain and produced a fluent, grammatically flawless instruction to do the opposite of what the prescriber intended. Elsewhere, “aortic aneurysm” became, in Spanish, an “evacuation of the main blood vessel”.

The pattern is not confined to hospitals. A 2010 study in Pediatrics by Iman Sharif and Julia Tse examined 76 Spanish-language prescription labels generated by the computer programmes New York pharmacies were actually using and found errors in half of them, including the notorious collision in which the English word “once” reads in Spanish as eleven. In immigration proceedings, translators with the volunteer organisation Respond Crisis Translation told Rest of World in 2023 that machine translation of Pashto and Dari was corrupting asylum claims, in one case rendering the first person singular as a plural throughout a personal narrative, so that an individual persecution story read as a collective complaint. At least one Afghan claim was rejected after such errors.

No surface signal attaches to any of these. A negation reversal reads exactly like a negation preserved: same length, same cadence, same register, same institutional wrapper. That is what distinguishes machine translation from almost every other failure mode in consumer technology. The output carries no evidence of its own unreliability, and the evidence that would reveal it exists only in a document the recipient does not have.

Consider the chain. The clinician cannot read the target language and assumes the system did its job. The organisation that licensed the engine sees an aggregate quality score, if anything. The patient sees fluent prose on hospital letterhead and assumes that if it had not been checked, it would not have been handed over. Nobody holds both halves of the comparison, and the appearance of completeness does all the work verification used to do.

The older failure mode is instructive by contrast. In January 1980, an eighteen-year-old named Willie Ramirez arrived unconscious at a Florida emergency department. His Cuban family said he was “intoxicado”, meaning ill from something he had eaten. It was taken to mean intoxicated. Ramirez was treated as a drug overdose, his intracerebellar haemorrhage went undiagnosed for two days, and he was left quadriplegic; the settlement was estimated at around seventy-one million dollars over his lifetime. That failure had a location: a room, a decision, a defendant. When the same error comes from an engine invoked automatically inside an electronic patient record, it has no location at all.

A Thirty-Nine Point Gap Inside One Product

The Taira study's statistical finding should trouble anyone who believes machine translation is solved. Across seven languages, the overall meaning survived in 82.5 per cent of cases, 330 out of 400. That headline is meaningless, because the variation underneath it is enormous.

Spanish came out at 94 per cent accuracy. Tagalog at 90. Korean at 82.5. Chinese at 81.7. Vietnamese at 77.5. Farsi at 67.5. Armenian at 55. A thirty-nine point spread between best and worst inside a single product, through a single interface, with a single set of expectations attached. Farsi had a further problem no accuracy metric captures: directional text handling rendered some of it illegible.

Nothing in the interface communicates that spread. The menu presents Armenian and Spanish as equivalent options, and output arrives with the same speed, formatting and absence of caveat. An administrator enabling automated discharge translation is not choosing between a reliable service and an unreliable one. They are enabling both at once, with no signal for which patients get which.

This is language-tier inequality, and it maps onto each language's digital footprint. Languages with vast parallel corpora perform well; those spoken largely by populations who were never a lucrative localisation market perform badly. So the patients most likely to receive an unreliable translation are those least likely to have an alternative: recent arrivals, smaller diaspora communities, speakers of languages for which the local health system has no on-call interpreter at three in the morning.

The gap is not closing, because progress concentrates where the data is. A study in JMIR Formative Research in January 2026 by a team at Beth Israel Deaconess Medical Center tested Claude Sonnet 3.5 on Spanish translations of emergency discharge instructions and found the output essentially clinically acceptable, with evaluators scoring completeness and severity at 5.0 on a five-point scale and meaning at 4.9. Genuinely impressive, and a result about Spanish. It is also no longer isolated, which is the part of this argument that has changed. Meanwhile a framework paper in npj Digital Medicine in September 2025 by Ivan Lopez, David Velasquez, Jonathan Chen and Jorge Rodriguez recommends deploying first into languages with abundant digital resources, naming Spanish and Portuguese, and warns that underrepresented languages such as Quechua and Yorùbá need further testing first. The authors are right. But notice what that implies: the languages validated first needed validating least, and those where risk is highest get served last, or with no validation at all.

The strongest evidence against the case being made here arrived in Academic Emergency Medicine in 2026, and it deserves stating at full strength rather than filing under limitations. Giovanni Rodriguez, Patricia Hernández and colleagues ran a blinded non-inferiority study on fifty-three real emergency department discharge instructions of between one hundred and five hundred words, taken exactly as the clinicians wrote them, original spelling and grammar errors left in. Google Translate and ChatGPT-4o rendered them into Spanish, Brazilian Portuguese and Simplified Chinese, against professional translations of the same text. Professional medical interpreters, blinded to which was which, scored 477 unique translations for fluency, adequacy, meaning and severity against a non-inferiority margin of half a point.

Both machines came out non-inferior to the professionals on most domains. In Spanish and Brazilian Portuguese they matched professional translation on adequacy, meaning and severity, falling short only on fluency. In Simplified Chinese they matched it on all four. The frequency of clinically significant errors did not differ significantly by translation method.

That has to be conceded without qualification, because it is true and it matters. In those three languages, against genuinely messy clinician-written source text rather than tidied research prose, current machine translation is not measurably more dangerous than a professional human being. The 2019 and 2021 numbers quoted earlier are real, and they were real about the systems they tested, but those systems are two generations back. Nobody should now cite eight per cent Spanish inaccuracy as a description of what a modern engine does to a Spanish discharge sheet. It is not what happens.

What the study does not do is rescue the position it appears to demolish, and both reasons are visible in its own design. Start with the languages. Spanish, Brazilian Portuguese and Simplified Chinese are among the largest localisation markets on earth and among the best-resourced pairs in existence. The study says nothing whatever about Armenian at 55 per cent, or Farsi at 67.5, or any language whose speakers were never worth building a corpus for. It does not narrow the thirty-nine point spread. It measures the top of it, carefully, one study later.

The second reason goes to what non-inferiority establishes. That clinically significant errors did not differ by translation method is not a finding that machine translation is safe. It is a finding that professional human translation also carries a clinically significant error rate, and that the machine has drawn level with it. Nobody who commissions professional translation of consequential documents believes a single pass is sufficient, which is why the profession has review stages, standards describing them, and indemnity behind them. Both arms of that study were then read by trained evaluators. Every translation in it was checked. The subject here is the document nobody checks at all, and non-inferiority to an unverified human baseline is not a warrant for skipping verification. It moves the argument off provenance and onto verification, which is exactly where it belongs. The question was never whether a machine or a person produced the text. It is whether anyone competent read it afterwards.

There is direct evidence on both counts, from a study that measured what the non-inferiority design left out. In npj Digital Medicine in October 2025, Ryan Brewster and a large multidisciplinary team took paediatric discharge instructions into six languages, Arabic, Armenian, Bengali, simplified Chinese, Somali and Spanish, and compared three routes: ChatGPT-4o alone, professional linguist translation, and human-in-the-loop, meaning a machine draft post-edited by a professional linguist. Forty-two evaluators scored the output, twelve professional linguists, sixteen clinicians and fourteen bilingual family caregivers.

The tier gap turned up intact in a frontier model. For Spanish and Bengali, ChatGPT-4o came close to professional quality. For Armenian it scored 2.4 on overall quality against 3.6 for professional translation, a deficit of more than a full point on a five-point scale, and Somali and simplified Chinese also fell significantly below the professional standard. Whatever has been fixed since 2021, this has not been. Note that the two studies disagree about Chinese, non-inferior on every domain in one and materially below professional standard in the other, on different document types, against different comparison standards, judged by different evaluators, a year apart. That disagreement is not a scandal. It is a measurement of how thin the evidence is.

The other finding is the one this whole argument turns on. Human-in-the-loop did not merely rescue Armenian, it beat the professionals, scoring 3.9 against their 3.6. Across all six languages it was also the faster route, averaging 7.1 minutes per document against 16.8 for professional translation. Post-editing a machine draft was better than either route alone and roughly twice as fast as translating from scratch. That is what working looks like, and it is worth being precise about what it costs. It still requires the qualified bilingual reader. It makes that person faster. It does not make them unnecessary.

Every such study is itself an act of verification, and inherits the same cost structure. Beth Israel validated one language pair, one document type, one institution, which tells you nothing about Armenian, consent forms, or next quarter's model update. The evidence base is built at human speed against a deployment surface expanding at machine speed.

The Instruments That Cannot See the Errors That Matter

The obvious response is to automate the checking. If a machine can produce the translation, surely a machine can grade it. That is the promise of quality estimation, and what most enterprise deployments rely on when they route some segments to review and let others through.

Two pieces of 2026 research suggest the hope is misplaced exactly where it matters. In August, Serge Gladkoff, Angelika Vaasa, Sue Ellen Wright, Ingemar Strandvik and Lifeng Han released a peer-reviewed paper titled “Looking under the Wrong Lamppost”, accepted for the ninth International Conference on Natural Language and Speech Processing in September, arguing that automated quality estimation suffers structural rather than incidental limitations. Their conclusion is unusually blunt: segment-level scores should not be used as a standalone basis for routing, release or review bypass in production. They document overfitting, distribution collapse, failure to generalise across domains, and blindness to cohesion and coherence, which are not properties of individual segments at all. Their proposed direction is not better scoring but making human review cheaper rather than pretending it is unnecessary.

The clinical version arrived two months earlier. At the American Medical Informatics Association's Amplify conference in June 2026, William Mundo, Elizabeth Goldberg and Yanjun Gao presented work on AI-generated translation of emergency discharge instructions and found direct disagreement between automated evaluation and clinician judgement. Automated metrics suggested high quality. Clinicians found errors in precisely the meaning-sensitive areas carrying clinical consequence: mistranslated medication instructions, omitted return precautions, negation errors of the “take with food” variety, confusion between daily and twice daily, mild discomfort turned into severe pain. Automated metrics alone, they concluded, should not drive deployment.

Read together, the escape route closes. The class of error automated estimation misses is not a random sample. It is disproportionately the class that reverses instructions, drops warnings and changes doses, because those errors are small, local, fluent and invisible to systems trained to reward fluency. The scoring machinery is best at catching failures a human would catch instantly, and worst at the failures only a human can catch at all. You cannot build a verification layer out of the competence that produced the problem.

What the Market Did With the Savings

To see where value goes when a commodity becomes free, read the 2026 Nimdzi 100, the annual ranking of the world's largest language industry providers. It sizes the market at 72.6 billion dollars and reports growth figures that read like a diagram of stratification.

The hundred largest providers collectively grew by 1.1 per cent. The top ten grew by 3.6 per cent. Providers ranked fifty-first to hundredth contracted by 4.3 per cent, the first negative growth recorded for any segment since 2021. Ten providers entered the ranking who were not there the year before, largely through acquisition; Propio Language Services vaulted into third place by acquiring the interpreting company CyraCom. Industry optimism, on Nimdzi's own confidence rating, fell from 6.5 to 6.2.

That distribution is the signature of a market whose commodity has moved to machines. If you sell volume translation, your product competes with something priced at twelve cents per thousand words, and no cost base survives that. If you sell verification, liability, domain expertise, certification and the ability to put a name and an indemnity policy behind a document, you are selling the one thing machines did not make cheaper. The middle contracts because the middle sold the thing that became free.

From inside the profession, the same restructuring reads differently. A study submitted to arXiv in June 2026 by Yujun Wang, Ehud Reiter, Shimei Pan, Steffen Eger and Wei Zhao analysed 79,286 social media posts from Reddit, Facebook, Bluesky and Mastodon between 2019 and 2025 across four stakeholder groups. The arc in translator discussion tells the story in three data points. In 2021 they were talking about computer-assisted translation infrastructure, translation memories and termbases. By 2023 it was integrating AI into those workflows. By 2025 it was automation and quality assurance, which is to say they had stopped discussing how to translate and started discussing how to check machines. Translators held the most critical stance of any group. The communities are not even arguing about the same variables: where developers frame reliability as a hallucination rate, everyone else frames it as verification and oversight.

The professional consequence is a job description inverted without being renegotiated. Translators are increasingly employed not to produce text but to verify it, the more cognitively demanding half, and are paid at post-editing rates set on the assumption the machine already did most of the job. The standards infrastructure sees the distinction even where the market does not. ISO 17100 covers human translation services; ISO 18587 covers post-editing of machine output and requires post-editors to meet the same competence levels as translators under ISO 17100. Same competence, different pay grade, different assumption about how much of the output anyone read.

Nobody Ever Decided This

The most striking feature of this migration is that almost nowhere can you find the meeting where it was approved.

Organisations did not convene a risk committee, evaluate error rates by language, define a threshold above which human review became mandatory, and sign off. They enabled a feature. A content management system offered automatic localisation; a support platform added a translate button. A free allowance covered the first half a million characters a month, enough to translate a small organisation's entire public documentation set without generating an invoice that needed a signature.

Research commissioned by DeepL and reported in 2026 found enterprises investing heavily in language AI while still running critical global operations through manual, unintegrated workflows, which tells you the adoption pattern is not a strategy but a scattering of local decisions. Marketing copy, support documentation, product listings, internal policy, contract summaries and patient-facing instructions crossed the same threshold at different moments, authorised by different people, none asked to consider the whole. A system nobody decided to build is one nobody feels responsible for.

The Chain in Which Everyone Is a Bystander

Try to identify the responsible party in a concrete case and the exercise becomes a study in evaporation.

A patient is discharged with machine-translated instructions containing a reversed medication direction. Is the model provider liable? Its terms disclaim high-stakes use and it never knew the text was clinical. The record vendor? It integrated a service at a customer's request. The hospital? Arguably yes, the most promising thread, but it will point to the impossibility of reviewing every document in every language, and to the absence of any regulatory instruction telling it where the line sits. The clinician cannot read Armenian. The patient received a document from a hospital in their own language and read it correctly. The document was wrong.

Occasionally an institution refuses to accept this diffusion. In June 2018, a United States district court in Kansas suppressed evidence in United States versus Cruz-Zamora, in which an officer used Google Translate on a patrol car laptop to obtain consent to search a vehicle from a driver who spoke very little English. The officer typed “can I search the car?” The Spanish output, translated back, asked something closer to “can I find the car?” The court found the literal but nonsensical rendering meant the driver had not given unequivocal consent. That is a court locating responsibility with the party who chose to deploy the technology, rather than the person who received its output.

Cruz-Zamora is instructive precisely because it is unusual. It required an adversarial process, a defence lawyer, a back-translation and a judge willing to inspect the mechanism. Almost no machine-translated document gets that treatment. The vast majority arrive where there is no adversary and nobody whose job it is to check: a discharge sheet, a benefits letter, a tenancy notice, a warning label, a consent form. The error surfaces, if ever, only through the harm it causes, by which point the chain back to a mistranslated clause is nearly impossible to reconstruct.

The industry sells professional indemnity insurance precisely because translation carries assignable liability when a professional does it. A named translator working under ISO 17100 is insurable because the responsible party is identifiable. A model invocation inside a content pipeline is not, and the risk does not disappear when it becomes uninsurable. It transfers to the person holding the paper.

The Law Points Two Ways at Once

For a case study in regulatory incoherence, consider the current United States position on language access.

Section 1557 of the Affordable Care Act, whose final implementing rule took effect on 5 July 2024, requires covered health entities to provide qualified interpretation and translation, and specifically requires that machine-translated material be reviewed by a qualified human translator where accuracy is essential or the source is complex or technical. That addresses the verification gap directly: you may use the machine, but not unsupervised where it matters.

Eight months later, on 1 March 2025, Executive Order 14224 designated English the official language of the United States and revoked Executive Order 13166, the 2000 order requiring federal agencies to plan for meaningful access for people with limited English proficiency. It directed the Attorney General to rescind guidance issued under it, and the resulting Department of Justice memorandum of 14 July 2025 encouraged agencies to rescind their own guidance, to consider English-only services, and, explicitly, to use artificial intelligence to reduce translation costs. In July 2026, formal rescission notices for Title VI language access guidance appeared in the Federal Register.

The policy environment now simultaneously requires human review of machine translation in health settings and recommends machine translation as a cost-reduction strategy across federal programmes. Underlying civil rights law has not changed: Title VI still prohibits national origin discrimination, and recipients of federal funding still owe meaningful access. What has changed is the guidance telling them how, so the question of what meaningful access means when the translation is machine-generated is unanswered by anyone with authority to answer it.

Europe's failure has a different texture. The General Product Safety Regulation, applicable since 13 December 2024, requires products to be accompanied by instructions and safety information in a language easily understood by consumers in the member state where they are sold. That is an outcome requirement, usually preferable to a process requirement. But it is silent on how the translation is produced and on who must verify the text is in fact understood rather than merely present. A machine-translated warning that reverses a prohibition satisfies the formal requirement while defeating its purpose. The EU Artificial Intelligence Act adds transparency obligations from August 2026, including machine-readable marking of synthetic content, though whether a translation engine running over patient instructions falls inside its high-risk regime is a matter for lawyers rather than settled fact. A small joke is buried here: some widely consulted online reference versions of the AI Act's own articles carry a notice explaining that the translations are machine-generated and not the official European Parliament versions.

In the United Kingdom, NHS England published an improvement framework for community language translation and interpreting on 27 May 2025 that is unusually honest about the gap. It records annual NHS spend at around 75.5 million pounds against an estimated 250 to 300 million needed to meet actual demand, and warns that translation apps, while convenient, carry risks to accuracy and patient safety. It commits NHS England to national guidance specifying when AI is suitable, which tools are approved, how accuracy is verified across languages, and the governance frameworks including indemnity and responsibility. That last clause is an admission: who is responsible when an AI translation harms an NHS patient had not been settled. It is being written now, after the tools are already on the wards.

What Recipients Are Actually Being Handed

Strip away the institutional framing and answer the first question plainly. What does it mean to receive information machine-translated at a fraction of human cost and never checked by anyone who reads both languages?

It means being handed a document carrying the full social signalling of institutional assurance and none of the substance. Every cue on that page, the letterhead, the formatting, the fluency, evolved in a world where producing text in a language required a person who knew it. Those cues were reliable because they were expensive. Fluency was a costly signal of competence. Machine translation made fluency free, so the signal conveys nothing, and nobody has told the recipients.

That is usually asserted about lay recipients, and it would be easy to dismiss as condescension: of course a patient cannot tell, they are not a translator. But the non-inferiority study tested the claim without setting out to. Its blinded evaluators were professional medical interpreters, reading in their own working languages, paid to assess translation quality, and aware that some of what they were reading was machine output. They frequently took the machine translations for professional work. If the signal has stopped carrying information under those conditions, for those readers, there is no version of this in which it still carries information for a patient reading a discharge sheet in a corridor. Fluency is not weak evidence of human competence. It is no evidence at all.

It means epistemic risk has been transferred to the party least able to bear it. Previously the burden of ensuring correctness sat with the institution producing the text, because the institution employed the translator. It has been silently relocated to the reader, who is expected, without being told, to treat official documents in their own language as provisional. Those readers are disproportionately the people with fewest resources to seek a second opinion.

It means informed consent is being quietly hollowed out. Consent requires comprehension of accurate information. A patient who fully understands a fluent instruction that has reversed the source has comprehended something, but not the thing they were meant to consent to. Their signature attests to a document nobody on the institutional side has read.

And it means we have built, without deciding to, a two-tier information citizenship. Speak a high-resource language and the machine serves you at close to professional quality. Speak Armenian and the same interface, the same button, the same institutional promise delivers something accurate slightly more than half the time in controlled testing. Both are told they have been given information in their own language. Only one of them has.

An Accountability Regime That Would Actually Bite

The second question is harder. Who bears responsibility when the system producing the translation is more capable than any system available to tell whether it worked?

The honest starting point is that no technical fix resolves this. The verification gap is not an artefact of immature models but a structural property of a task whose correctness can only be assessed by someone holding both languages. Better models will narrow the error rate. They will not create the reviewers. So the regime must be built out of allocation of responsibility rather than engineering, starting by reversing the current default.

That default should be simple. The organisation publishing a translated document warrants it, regardless of how it was produced. Not the model provider, not the platform, not the clinician, and emphatically not the recipient. If a hospital hands a patient a discharge sheet in Armenian, it stands behind that sheet exactly as if a staff translator had written it. This single rule does most of the work, because it puts the cost of unverified translation onto the party that captured the savings rather than the recipient who absorbs the risk.

The rest follows. Risk tiering should be organised by consequence class rather than document type. A marketing headline and a dosing instruction are both text; only one can put someone in intensive care. Content that can change a medical action, alter a legal obligation, forfeit a right, or carry a safety warning belongs in a tier requiring independent bilingual review before release, attributable to a named professional. Everything else can run unverified, provided it is labelled as such.

Per-language performance disclosure should be mandatory and visible in the interface. A product performing at 94 per cent in one language and 55 per cent in another should not present those as equivalent menu options. Deployers in regulated settings should publish language-specific error rates for their own content domains, sampled by qualified reviewers, and suspend automated release for pairs below a defined floor. The Lopez framework recommends the healthcare version: extend Joint Commission oversight to machine-assisted translation, build a shared clinical translation corpus, and update Section 1557 guidance with language-specific benchmarks. Language-specific is the operative phrase, because aggregate accuracy conceals the inequality that matters.

Provenance labelling should appear in the target language, in the document, in terms a recipient can act on. Not a disclaimer in eight-point type, but a plain statement: this text was produced automatically and has not been checked by a person who reads both languages; if anything here concerns your medication, contact us. The harm mechanism is a false impression of human authorship, and labelling is the cheapest counterweight.

Procurement is the most immediately available lever. Contracts in regulated settings should specify ISO 18587 full post-editing as a floor for consequence-bearing content, name the accountable reviewer, require logging sufficient to reconstruct which engine produced which document on which date, and prohibit reliance on automated quality estimation as the sole basis for bypassing review. The Gladkoff paper gives that last clause its evidentiary basis. The Brewster results give the first clause its own: full post-editing was not a compromise between speed and safety but better than either route alone, and faster than commissioning a translation from scratch. The expensive part of translation is no longer the translating. It is the reading.

Finally there is a public-goods problem no single organisation will solve. Armenian performs at 55 per cent because of a shortage of high-quality parallel data, and that shortage persists because nobody has a commercial incentive to fix it. Building open, domain-specific corpora for lower-resource languages in the areas where errors are most dangerous, clinical instructions, legal notices, safety warnings, is unglamorous infrastructure that health systems, regulators and standards bodies should fund jointly. It is cheap relative to the litigation it would prevent, and the only intervention that closes the tier gap rather than documenting it.

None of this restores the old equilibrium, in which producing a translation and trusting one cost roughly the same. What a workable regime does instead is make the residual risk visible, priced and owned, so that an organisation choosing to publish unverified translations makes that choice explicitly, in writing, with its name attached.

The Sentence That Reads Perfectly and Says the Opposite

Return to the instruction that started this. Hold the kidney medicine until you have a chance to speak with your kidney doctor. In the source it is a stop order. In the output, in two of the world's most widely spoken languages, it became a continue order, in prose so natural that a native speaker would have no cause to question it.

That sentence is the whole problem in miniature, and it explains why the failure mode has been so hard to take seriously. Every intuition we hold about text quality formed under an assumption that no longer applies: that fluent, contextually appropriate prose in a language implies someone competent in that language was involved. It was a reliable inference for the entire history of writing. It became false around three years ago, and our institutional design has not caught up.

The organisations that migrated to machine output did nothing obviously reckless. Each decision looked like a straightforward efficiency gain, and the aggregate was never assembled by anyone. The people receiving the documents are not careless either; they read correctly and trust appropriately, given every signal available. The models are not malfunctioning. They perform the task they were built for well enough that the residual failures are undetectable by anyone in the chain.

That is what makes it a trap rather than a scandal. Nobody behaves unreasonably from their own vantage point, and the system as a whole produces outcomes nobody would defend if they could see them. The technology has become powerful enough to make the question of whether it worked unanswerable at the point of use, and we have responded by not asking. The gap will not close on its own, because the cheap thing keeps getting cheaper and the expensive thing is expensive for reasons that have nothing to do with technology. The only variable under our control is who is answerable for the difference.

At present the answer is the person holding the paper, who does not know there is a question.

Sources and References

1. Breena R. Taira, Vanessa Kreger, Aristides Orue and Lisa C. Diamond, “A Pragmatic Assessment of Google Translate for Emergency Department Instructions,” Journal of General Internal Medicine, 5 March 2021: https://link.springer.com/article/10.1007/s11606-021-06666-z
2. Elaine C. Khoong, Eric Steinbrook, Cortlyn Brown and Alicia Fernandez, “Assessing the Use of Google Translate for Spanish and Chinese Translations of Emergency Department Discharge Instructions,” JAMA Internal Medicine, 25 February 2019: https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2725080
3. Giovanni Rodriguez, Patricia Hernández, Christopher Kirwan, Lisette Dunham and Sayon Dutta, “Comparative Evaluation of Machine Translation Accuracy of Emergency Department Discharge Instructions: A Non-Inferiority Study,” Academic Emergency Medicine, volume 33, issue 4, 2026: https://onlinelibrary.wiley.com/doi/abs/10.1111/acem.70289
4. Ryan C. L. Brewster et al., “Evaluating human-in-the-loop strategies for artificial intelligence-enabled translation of patient discharge instructions: a multidisciplinary analysis,” npj Digital Medicine, 24 October 2025: https://www.nature.com/articles/s41746-025-02055-6
5. Nimdzi Insights, “The 2026 Nimdzi 100,” 2026: https://www.nimdzi.com/nimdzi-100-2026/
6. Yujun Wang, Ehud Reiter, Shimei Pan, Steffen Eger and Wei Zhao, “Beyond Accuracy: Community Perspectives on Machine Translation,” arXiv:2606.09655, 8 June 2026: https://arxiv.org/abs/2606.09655
7. Serge Gladkoff, Angelika Vaasa, Sue Ellen Wright, Ingemar Strandvik and Lifeng Han, “Looking under the Wrong Lamppost: On the Limitations of Automated Translation Quality Estimation,” arXiv:2608.03577, 4 August 2026, forthcoming in the Proceedings of the 9th International Conference on Natural Language and Speech Processing (ICNLSP 2026), Trento, Italy, September 2026: https://arxiv.org/abs/2608.03577
8. University of Colorado Anschutz Medical Campus, “Researchers Examine Safety Risks in AI-Generated Translation of Emergency Department Discharge Instructions,” 23 June 2026: https://news.cuanschutz.edu/dbmi/ai-generated-discharge-instructions
9. Carreras Tartak et al., “Evaluating Spanish Translations of Emergency Department Discharge Instructions by a Large Language Model: Tool Validation and Reliability Study,” JMIR Formative Research, 12 January 2026: https://formative.jmir.org/2026/1/e79676
10. Ivan Lopez, David E. Velasquez, Jonathan H. Chen and Jorge A. Rodriguez, “Operationalizing machine-assisted translation in healthcare,” npj Digital Medicine, 30 September 2025: https://pmc.ncbi.nlm.nih.gov/articles/PMC12485017/
11. Gail Price-Wise, “Language, Culture, And Medical Tragedy: The Case Of Willie Ramirez,” Health Affairs Forefront, 19 November 2008: https://www.healthaffairs.org/do/10.1377/forefront.20081119.000463/
12. Iman Sharif and Julia Tse, “Accuracy of Computer-Generated, Spanish-Language Medicine Labels,” Pediatrics, May 2010: https://pubmed.ncbi.nlm.nih.gov/20368321/
13. Rest of World, “AI translation jeopardizes Afghan asylum claims,” 19 September 2023: https://restofworld.org/2023/ai-translation-errors-afghan-refugees-asylum/
14. TechCrunch, “Judge says 'literal but nonsensical' Google translation isn't consent for police search,” 15 June 2018: https://techcrunch.com/2018/06/15/judge-says-literal-but-nonsensical-google-translation-isnt-consent-for-police-search/
15. The White House, “Designating English as the Official Language of the United States,” Executive Order 14224, 1 March 2025: https://www.whitehouse.gov/presidential-actions/2025/03/designating-english-as-the-official-language-of-the-united-states/
16. Federal Register, “Notice of Rescission of Guidance to Federal Financial Assistance Recipients Regarding Title VI Prohibition Against National Origin Discrimination Affecting Limited English Proficient Persons,” 14 July 2026: https://www.federalregister.gov/documents/2026/07/14/2026-14128/notice-of-rescission-of-guidance-to-federal-financial-assistance-recipients-regarding-title-vi
17. American Translators Association, “Section 1557 of the Affordable Care Act and Language Access: Who, What, How”: https://www.atanet.org/client-assistance/blog-section-1557-of-the-affordable-care-act-and-language-access-who-what-how/
18. International Organization for Standardization, “ISO 18587:2017 Translation services, Post-editing of machine translation output, Requirements”: https://www.iso.org/standard/62970.html
19. TÜV SÜD, “ISO 17100 and ISO 18587 Certifications, Translation Quality and Machine Translation Standards”: https://www.tuvsud.com/en-us/services/auditing-and-system-certification/iso-17100
20. NHS England, “Improvement framework: community language translation and interpreting services,” 27 May 2025: https://www.england.nhs.uk/long-read/improvement-framework-community-language-translation-and-interpreting-services/
21. EU-OSHA, “Regulation (EU) 2023/988 on general product safety,” applicable from 13 December 2024: https://osha.europa.eu/en/legislation/directive/regulation-2023988eu-general-product-safety
22. EU Artificial Intelligence Act, “Article 50: Transparency Obligations for Providers and Deployers of Certain AI Systems”: https://artificialintelligenceact.eu/article/50/
23. Google Cloud, “Cloud Translation pricing”: https://cloud.google.com/translate/pricing
24. The AI Journal, “Manual translation processes still stifling enterprises despite surge in AI spending, finds DeepL research,” 10 March 2026: https://aijourn.com/manual-translation-processes-still-stifling-enterprises-despite-surge-in-ai-spending-finds-deepl-research/
25. National Immigration Law Center, “Trump Administration's Attempts to Dismantle Language Access Do Not Erase Civil Rights Law”: https://www.nilc.org/articles/trump-administrations-attempts-to-dismantle-language-access-do-not-erase-civil-rights-law/

---

Tim Green UK-based Systems Theorist & Independent Technology Writer

Tim explores the intersections of artificial intelligence, decentralised cognition, and posthuman ethics. His work, published at smarterarticles.co.uk, challenges dominant narratives of technological progress while proposing interdisciplinary frameworks for collective intelligence and digital stewardship.

His writing has been featured on Ground News and shared by independent researchers across both academic and technological communities.

ORCID: 0009-0002-0156-9795 Email: tim@smarterarticles.co.uk

Listen to the free weekly SmarterArticles Podcast

Alan Parsons Project – 1976

A Dream Within A Dream The Raven The Tell-Tale Heart The Cask Of Amontillado(The System Of) Doctor Tarr And Professor Fether The Fall Of The House Of Usher: Prelude The Fall Of The House Of Usher: Arrival The Fall Of The House Of Usher: Intermezzo The Fall Of The House Of Usher: Pavane The Fall Of The House Of Usher: Fall To One In Paradise

A Dream Within A Dream (Instrumental)

“For my own part, I have never had a thought which I could not set down in words with even more distinctness than that with which I conceived it. There is, however, a class of fancies of exquisite delicacy which are not thoughts, and to which as yet I have found it absolutely impossible to adapt to language. These fancies arise in the soul, alas how rarely. Only at epochs of most intense tranquillity, when the bodily and mental health are in perfection. And at those weird points of time, where the confines of the waking world blend with the world of dreams. And so I captured this fancy, where all that we see, or seem, is but a dream within a dream.”

---

The Raven

THE clock struck midnight And through my sleeping I heard a tapping at my door I looked but nothing lay in the darkness And so I turned inside once more

To my amazement There stood a raven Whose shadow hung above my door Then through the silence It spoke the one word That I shall hear for evermore

Nevermore Thus quoth the raven, nevermore

And still the raven remains in my room No matter how much I implore No words can soothe him No prayer remove him And I must hear for evermore

Quoth the raven, nevermore Nevermore Thus quoth the raven, nevermore

---

The Tale Tell Heart

“The Tell-Tale Heart”

YOU should have seen him Lying alone in helpless silence in the night You should have seen him You would have seen his eye reflecting in the light

So for the old man Ashes to ashes, earth to earth and dust to dust No one will see me No one with guilt to share, no secret soul to trust

Louder and louder Till I could tell the sound was not within my ears You should have seen me You would have seen my eyes grow white and cold with fear

Heard all the things in Heaven and Earth I've seen many things in Hell But his vulture's eye of a cold pale blue Is the eye of the Devil himself

Take me away now But let the silence drown the beating of his heart

---

“The Cask Of Amontillado”

BY the last breath of the four winds that blow I'll have revenge upon Fortunato Smile in his face I'll say “come let us go I've a cask of Amontillado”

Sheltered inside from the cold of the snow Follow me now to the vault down below Drinking the wine as we laugh at the time Which is passing incredibly slow

(What are these chains that are binding my arm?) Part of you dies each passing day (Say it's a game and I'll come to no harm) You'll feel your life slipping away

You who are rich and whose troubles are few May come around to see my point of view What price the Crown of a King on his throne When you're chained in the dark all alone

(Spare me my life only name your reward) Part of you dies each brick I lay (Bring back some light in the name of the Lord) You'll feel your mind slipping away

---

”(The System Of) Doctor Tarr And Professor Fether”

JUST what you need to make you feel better Just what you need to make you feel Just what you need to make you feel better Just what you need to make you feel

At the far end of your tether And your thoughts won't fit together So you sleep light or whatever And the night goes on forever Then your mind change like the weather You're in need of Doctor Tarr and Professor Fether

Even clouds seem bright and breezy 'Cause the livin' is free and easy See the rat race in a new way Like you're wakin' up to a new day It's a wise thing if you're clever Take a lead from Doctor Tarr and Professor Fether

Find the end of a rainbow Fly wherever the winds blow Laugh at life like a sideshow Just what you need to make you feel better

Satisfaction altogether Guaranteed by Doctor Tarr and Professor Fether

Find the end of a rainbow Fly wherever the winds blow Laugh at life like a sideshow Just what you need to make you feel better

Don't stop bringin' the girls round Don't start havin' a showdown Keep on handin' the jug round All that you need is wine and good company

---

“The Fall Of The House Of Usher – I Prelude”

[instrumental]

[(I) Prelude]

“Shadows of shadows passing. It is now 1831, and as always I am absorbed with a delicate thought. It is how poetry has indefinite sensations, to which end music is inessential. Since the comprehension of sweet sound is our most indefinite conception, music, when combined with a pleasurable idea, is poetry. Music without the idea is simply music. Without music or an intriguing idea, colour becomes pallor, man becomes carcase, home becomes catacomb, and the dead are but for a moment motionless.”

[(II) Arrival]

[(III) Intermezzo]

[(IV) Pavane]

[(V) Fall]

---

“To One In Paradise”

IF I could see the sky above And my mind could be set free As wild white horses reached the shore I'd stand alone and oversee

And if the bush before me burns Should I turn my eyes away And still the voices I can hear As clear to me as light of day

I believed in my dreams Nothing could change my mind Now I know what they mean How could I be so blind

Cold sands of time (Winds that blow as cold as ice Sounds that come in the night) Shall hide what is left on me (Come from Paradise)

I've been through times when no one cared (Words that were mine) I've seen clouds in empty skies When one kind word meant more to me (Shall last as a memory) Than all the love in Paradise

I believed in my dreams Nothing could change my mind Till I found what they mean Nothing can save me now

---

---

Ps 34:18 Jehovah is close to the brokenhearted; He saves those who are crushed in spirit.

Hebrews 6:10 For God is not unrighteous so as to forget your work and the love you showed for his name by ministering and continuing to minister to the holy ones.

Tales of Mystery: Dark beauty. Emotion within architecture. Disciplined intensity.

Horse gates

Jeremiah 31 is a sweeping promise of restoration. Jehovah assures Israel that his love has not ended: he will gather his scattered people, comfort their grief, rebuild their land, and turn mourning into joy. Even Ephraim’s remorse is met with compassion, showing Jehovah as a Father who still deeply loves his wayward son. The chapter reaches its high point with the promise of a new covenant—one in which God’s law would be written on hearts, sins would be forgiven, and his people would truly know him. The chapter closes with an emphatic assurance that Jehovah’s purpose for his people is as dependable as the sun, moon, stars, and the foundations of the earth.

---

“At that time,” declares Jehovah, “I will become God to all the families of Israel, and they will become my people.”

This is what Jehovah says:

“The people who survived the sword found favor in the wilderness

When Israel was walking to his resting-place.”

From far away Jehovah appeared to me and said:

“I have loved you with an everlasting love.

That is why I have drawn you to me with loyal love.

Yet again I will rebuild you and you will be rebuilt.

O virgin of Israel, you will again take up your tambourines

And go forth dancing joyfully.

You will again plant vineyards in the mountains of Sa·marʹi·a;

The planters will plant and enjoy their fruit.

For the day will come when the watchmen in the mountains of Eʹphra·im will call out:

‘Rise up, let us go up to Zion, to Jehovah our God.’”

For this is what Jehovah says:

“Cry out to Jacob with rejoicing.

Shout for joy because you are over the nations.

Proclaim it; offer praise and say,

‘O Jehovah, save your people, the remnant of Israel.’

I am bringing them back from the land of the north.

I will gather them together from the remotest parts of the earth.

Among them will be the blind and the lame,

The pregnant woman and the one giving birth, all together.

As a great congregation they will return here.

They will come weeping.

I will lead them as they beg for favor.

I will guide them to streams of water,

On a level path that will not make them stumble.

For I am a Father to Israel, and Eʹphra·im is my firstborn.”

Hear the word of Jehovah, you nations,

And proclaim it among the islands far away:

“The One who scattered Israel will gather him together.

He will watch over him as a shepherd does his flock.

For Jehovah will redeem Jacob

And rescue him out of the hand of the one stronger than he is.

They will come and shout joyfully on the height of Zion

And become radiant over the goodness of Jehovah,

Over the grain and the new wine and the oil,

And over the young of the flock and the herd.

They will become like a well-watered garden,

And they will never languish again.”

“At that time the virgin will dance joyfully,

Also the young men and the old men together.

I will change their mourning into exultation.

I will comfort them and give them joy instead of their grief.

I will satisfy the priests with plenty,

And my people will be satisfied with my goodness,” declares Jehovah.

“This is what Jehovah says:

‘A voice is heard in Raʹmah, lamentation and bitter weeping:

Rachel is weeping over her sons.

She has refused to be comforted over her sons,

Because they are no more.’”

This is what Jehovah says:

“‘Hold back your voice from weeping and your eyes from tears,

For there is a reward for your activity,’ declares Jehovah.

‘They will return from the land of the enemy.

And there is a hope for your future,’ declares Jehovah.

‘Your sons will return to their own territory.’”

“I have surely heard Eʹphra·im’s moaning,

‘You have corrected me, and I have been corrected,

Like a calf that has not been trained.

Bring me back, and I will readily turn back,

For you are Jehovah my God.

For after my turning back I felt remorse;

After I was made to understand I struck my thigh in grief.

I was ashamed and humiliated,

For I bore the reproach of my youth.’”

“Is Eʹphra·im not a precious son to me, a beloved child?

For as often as I speak against him, I do remember him still.

That is why my emotions are stirred for him.

And I will surely have pity on him,” declares Jehovah.

“Set up road markers for yourself,

And put up signposts.

Pay attention to the highway, the way that you have to go.

Return, O virgin of Israel, return to these cities of yours.

How long will you waver, O unfaithful daughter?

For Jehovah has created something new in the earth:

A woman will eagerly pursue a man.”

This is what Jehovah of armies, the God of Israel, says:

“They will again say these words in the land of Judah and in its cities when I gather back their captives: ‘May Jehovah bless you, O righteous dwelling place, O holy mountain.’

“And in it Judah and all its cities will dwell all together, farmers and those who lead the flocks. For I will satisfy the weary one and fill each one who is languishing.”

At this I awoke and opened my eyes, and my sleep had been pleasurable to me.

“Look! The days are coming,” declares Jehovah, “when I will sow the house of Israel and the house of Judah with the seed of man and with the seed of livestock.

“And just as I watched over them to uproot, to pull down, to tear down, to destroy, and to do harm, so I will watch over them to build up and to plant,” declares Jehovah.

“In those days they will no longer say, ‘The fathers ate sour grapes, but the teeth of the sons were set on edge.’ But then each one will die for his own error. Any man eating sour grapes will have his own teeth set on edge.”

“Look! The days are coming,” declares Jehovah, “when I will make with the house of Israel and with the house of Judah a new covenant.

“It will not be like the covenant that I made with their forefathers on the day I took hold of their hand to lead them out of the land of Egypt, ‘my covenant that they broke, although I was their true master,’ declares Jehovah.”

“For this is the covenant that I will make with the house of Israel after those days,” declares Jehovah.

“I will put my law within them,

And in their heart I will write it.

And I will become their God,

And they will become my people.”

“And they will no longer teach each one his neighbor and each one his brother, saying, ‘Know Jehovah!’ for they will all know me, from the least to the greatest of them,” declares Jehovah.

“For I will forgive their error, and I will no longer remember their sin.”

This is what Jehovah says,

Who gives the sun for light by day,

The laws of the moon and the stars for light by night,

Who stirs up the sea and makes its waves boisterous,

Whose name is Jehovah of armies:

“‘If these regulations should ever fail,’ declares Jehovah,

‘Only then would the offspring of Israel cease as a nation before me always.’”

This is what Jehovah says:

“‘If the heavens above could be measured and the foundations of the earth below could be explored, only then could I reject all the offspring of Israel for all they have done,’ declares Jehovah.”

“Look! The days are coming,” declares Jehovah, “when the city will be built to Jehovah from the Tower of Ha·nanʹel to the Corner Gate.

“And the measuring line will go out straight ahead to the hill of Gaʹreb, and it will turn toward Goʹah.

“And all the valley of the carcasses and of ashes and all the terraces as far as the Kidʹron Valley, clear to the corner of the Horse Gate toward the east, will be something holy to Jehovah.

“It will never again be uprooted or torn down.”

---

#jeremiah #biblereading

Boundless Energy: Blessing or Curse?

For as long as I can remember, people have talked about my mother’s energy. Neighbors, friends, relatives  commented on how energetic and active she was. She was always doing something and could never seem to sit still.

Even before she was married, she had countless hobbies: sewing, making dolls and clothes, knitting, fabric painting, and everything in between. She loved to cook and took it as a personal challenge to recreate dishes made by neighbors from the various cultures and ethnicities residing in our university housing community. Heck, she even somehow coaxed a reluctant neighbor to divulge her secret Family chocolate chip cookie recipe!I honestly don’t remember seeing her sit very often.I certainly don’t remember her lounging on the sofa, relaxing and watching Tv. She was always on the go.

Growing up Asian (if you know, you know ) there was a huge emphasis on productivity. You woke up at a respectable hour. Periodt. Sleeping in was frowned upon, even if you had been up half the night. Lazy people never got anywhere in life. The Nepali word was “Alchhi”. Because apparently, all successful futures were decided before 8 a.m.!

If I was sitting quietly, I was often asked why I was “just sitting there doing nothing.”Was I supposed to explain that I was thinking? reflecting ( or day dreaming about my latest crush) , conserving energy? Existing?!

I always felt as though I needed to be occupied with something or at perpetual risk of being judged. “Hare yo keti ta”… (translation: uff this girl!). Your value was tied to what you achieved: your grades, your accomplishments, and how many talents you could  hone.

It also gave parents more material to brag about their children while subtly validating themselves as wonderful parents.Because what was the point of raising an accomplished child if nobody else knew about it? It was personal.

And listen, I understand it. There is absolutely truth in that how you are raised: the support, discipline, and encouragement you receive, shapes what you become. I am deeply grateful to my parents. They encouraged me, challenged me, and geared me to be successful and independent.

They beamed from the moment I first walked into Medical school (and maybe my first day of kindergarten once before) and continued beaming all the way through my becoming a fully fledged, experienced physician. And they had every right to be proud.I couldn’t have done it without them.

But somewhere along the way, I absorbed more than ambition.I absorbed the idea that stillness was scary and wrong.

Because if I wasn’t doing something, wasn’t I wasting time? And if I was wasting time, wasn’t I wasting my potential?

After having children, I became more unstoppable. Literally. I stopped sitting down to watch shows. I would come home after commuting and hours of work and taking care of others to care for the kids, play with them, clean something, plan something, or find another task that suddenly felt urgent.

Over the past few years, I have incorporated more “self-care” into my life but even my self-care involved activity!

The gym. Pilates. Walking. Dancing. Learning new hobbies.Coffee dates with my girlfriends, Playdate and girls nights. Planning outings. Going places.

Sadly, I decided rest was acceptable only if it came with athletic shoes and an elevated heart rate.Did it really count as self-care if I wasn’t sweating, multitasking, or checking something off a list? Creating more, writing more.Etc etc.

Friends and family began asking me the same things about me that everyone had always said about my mother: Where do you get all that energy? How do you do so much? You’re always doing something.

And I was.

I planned, overplanned, booked, and overbooked my weekdays. Then, by the weekend, I desperately needed a break, so naturally, I immediately started planning the weekend. What better way to recover from an overbooked week than by overbooking your recovery??

I chose large, comfortable, expensive sofas for my home. I created cozy little nooks for resting, both inside and outside.

And then I stood in the kitchen eating.

Always standing. Eating, cleaning, organizing, doing.

Never actually sitting on those beautiful sofas I had carefully selected for the theoretical person who lived in my house and knew how to relax.

Who was she? When was she coming? And why did she need so many decorative pillows?

I would wash my hair at midnight after everyone was asleep-and finally sit down only when the lights were off and I was ready to fall asleep.

For years, I thought this was simply my personality. I had energy. I was productive. I was efficient. I got things done.I was proud of it.

But recently, I began to understand that some of it was also a response- a traumatic  one that was probably passed down from generation to generation.

We were taught, directly and indirectly, that having fun was frivolous. Taking care of ourselves was selfish. Resting meant we were lazy. And our worth is demonstrated repeatedly through how much we could accomplish.

But who decided that? And why did we all agree without even reading the terms and conditions?

Recently, I attended an event where we were asked to name the three things we valued most in life.

Someone said, “rest”. Someone said, “Fun.”

And I remember thinking: Rest? Fun? You’re allowed to choose that?

Why shouldn’t fun be something we value? Why shouldn’t rest be important enough to claim openly instead of disguising it as recovery from productivity?

Why do we feel compelled to justify rest by checking off everything we accomplished first?

I cleaned the house, answered my emails, worked out, fed everyone, solved three minor crises, picked up the pet’s messes, reorganized a drawer…so now I may sit down for eleven minutes.

These beliefs are sneaky. But become so deeply ingrained that we start to  think this is how responsible, successful adults are supposed to live.

I want to teach my children something different. I want them to listen to their bodies. I want them to know that it is okay to have an evening with no after-school activity. It is okay to have a weekend without a bazillion plans. It is okay to sit on the sofa, watch something silly, laugh, nap, FaceTime a friend or simply do nothing.

Will an empty Saturday prevent them from becoming successful adults? Or could learning to rest actually help them become healthier, happier ones?

Rest is not a failure to use your time well.It is a gift we give ourselves. It may even be one of the greatest acts of self-love.

Boundless energy can absolutely be a blessing. It has helped me create, learn, grow, achieve, care for people, raise my children, build a career, friendships and community. And to live a very full vibrant life.

But energy becomes a burden when we believe we are never allowed to replenish it. Our worth does not need to be proven minute by minute; not to the world, not to our families, and certainly not to ourselves.

And really, what is the point of buying the plush sofa and designing the perfect coffee nook if we never actually use it?

Sometimes, sitting on the darn sofa is enough.

There is a kind of exhaustion that comes from constantly trying to make sure everyone around you is okay with you. You start measuring your decisions by other people’s reactions. If they are happy, you feel relieved. If they are disappointed, you start wondering what you did wrong. Over time, what Jesus teaches about people-pleasing becomes something we desperately need to understand.

I think many of us confuse love with responsibility. We believe that if we truly care about someone, then we should be able to keep them from being disappointed, angry, hurt, or frustrated with us. But learning to follow Jesus without living for approval means accepting something difficult: you can love someone deeply and still not be responsible for how they feel about every decision you make.

Jesus showed us this over and over again. He cared about people. He listened to them, healed them, fed them, taught them, forgave them, and gave Himself for them. But He did not allow their expectations to become His master.

In Mark 1, people were searching for Jesus after He had been healing and helping crowds. His disciples found Him and said, “Everyone is looking for you.” It would have been easy to hear that as an obligation. People wanted Him. People needed Him. People expected Him to come back.

But Jesus did not return simply because everyone wanted Him to.

He told His disciples that they needed to go to other towns so He could preach there too.

I find that deeply freeing.

Jesus did not ignore people. He did not stop caring. He simply knew that not every expectation placed on Him was an assignment from the Father.

That is something I have to remind myself of too.

Not every request is my responsibility. Not every disappointment means I failed. Not every unhappy reaction means I need to change my decision.

Sometimes people are disappointed because we were wrong. When that happens, we should admit it. We should apologize. We should make things right.

But sometimes people are disappointed simply because we said no.

Sometimes they are upset because we created a boundary.

Sometimes they are frustrated because we stopped doing something they had grown comfortable receiving from us.

Their feelings may be real, but that does not mean we have sinned.

Jesus disappointed people too.

Peter once tried to convince Jesus to avoid the suffering that was ahead of Him. Peter loved Jesus. He did not want Him to suffer. His concern was sincere.

But Peter’s preferred path was not the Father’s plan.

Jesus loved Peter, but He did not obey Peter’s emotions.

That matters.

Someone can love you and still want the wrong thing for you.

Someone can care about you and still pressure you toward a decision you should not make.

Someone can be disappointed with you even when you are trying to do what is right.

This is where people-pleasing becomes dangerous.

It can look like kindness on the surface. We tell ourselves we are just trying to keep the peace. We do not want conflict. We do not want anyone to think badly of us.

But eventually, keeping everyone happy can become more important than being honest.

We start saying yes when we mean no. We stay in situations we know are unhealthy. We avoid conversations that need to happen. We keep giving away our time and energy because guilt feels easier than disappointment.

That is not freedom.

Jesus never promised that following Him would make everyone happy with us.

In fact, His own life proves the opposite.

There were days when crowds followed Him and praised Him. There were other days when people rejected Him, argued with Him, misunderstood Him, and walked away.

Jesus did not change who He was every time the crowd changed its opinion.

He knew who He was.

He knew who His Father was.

And He knew what He had been sent to do.

I think that is where peace begins for us too.

We have to stop asking, “Is everybody happy with me?” and start asking, “Am I being faithful to Jesus?”

Those are not the same question.

You can make someone happy and still be disobedient.

You can disappoint someone and still be faithful.

That does not give us permission to become cold. Jesus was never cold. He was compassionate, patient, generous, and forgiving.

The goal is not to stop caring about people.

The goal is to stop carrying what belongs to them.

I am responsible for how I speak.

I am responsible for whether I forgive.

I am responsible for whether I act with kindness and integrity.

I am responsible for whether I tell the truth.

But I am not responsible for controlling another person’s emotions.

Neither are you.

Sometimes love sounds like, “I care about you, but I cannot do that.”

Sometimes love says, “I understand that you are disappointed, but I believe this is the right decision.”

Sometimes love means staying close.

Sometimes love means creating distance.

The important thing is that our decisions come from prayer, truth, humility, and obedience rather than fear of rejection.

So if someone is disappointed with you today, bring it to God.

Ask Him to show you whether you were wrong.

If you were, make it right.

But if you were honest, loving, prayerful, and faithful, do not automatically assume that another person’s unhappiness means you need to change course.

You cannot keep everyone happy.

Jesus did not.

What you can do is love people well, follow Christ faithfully, and trust God with the reactions you cannot control.

That burden was never yours to carry.

Your friend,

Douglas Vandergraph

Explore the complete Douglas Vandergraph Master Index: https://douglasvandergraph.com/douglas-vandergraph-master-index/

Watch Douglas Vandergraph’s faith-based videos on YouTube: https://www.youtube.com/@douglasvandergraph

I suppose it’s presumptuous to assume that my posts are of any interest to anyone, let alone worthy of a few minutes of attention to be read. But that assumes I’m writing this because I desire or crave attention. Honestly, I do not.

But I am compelled, for reasons I myself do not entirely understand, to share my thoughts and opinions here on things that probably should not be shared.

Does writing this go against my better judgment? That depends. What constitutes ‘better’ judgment? Is it the act, the process, the substance, or the consequence of such writings? Writings that would certainly be labeled as strange at least, insane at most. Or, worse perhaps, as I have a limited imagination.

As a writer, I’m a mere amateur. This means I have something to say but lack the language to make it truly compelling and memorable. But after five decades on this earth, the prose I offer here is likely the fullest and best version of my written capabilities.

But that doesn’t really answer the question of why I’m doing this. And after much soul-searching and back-and-forth in my mind, the best answer I can offer is this: Because to keep what burns inside me has begun to devour me from within.

Speaking my truth to the world is better than letting my inner world consume me whole.

I recognize the risks.

Our world has tidy boxes for what it considers untidy things, no matter how wrong the world’s boxes may be.

Human societies and civilizations have a long history of misunderstanding, preservation bias, and erasure when it comes to the weird, anomalous, or uncanny — the atopic. The system is efficient and expedient with its assigning of labels: episode, delusions, possession, psychosis, blasphemy, mania, hallucination. Pick your favorite. The thesaurus has plenty more. And modern medicine has an endless array of chemicals they’re happy to pump into you at cost, while religion has an endless array of dogmatic corrections they’re happy to also pump into you at cost.

To both I say, “I’d rather not.”

I suppose it's safer and easier to stay silent. After all, there’s no real benefit for me here other than some hope of processing, integration, and, if I’m lucky, a bit of solace. I thought I’d say peace of mind, but that seems impossible at this point. At least, not in this lifetime. Silence would certainly keep me out of the vile filing systems of this world. But.

Silence also corrodes.

If left unmanaged, some unspoken words and thoughts become toxic. They rush the mind, body, and soul like a firestorm, leaving nothing but ash and cinder in their wake.

And there’s a naive part of me that clings to the old-fashioned notion that the sort of thing that happens to one person and is never shared has no resistance against the onslaught of preservation bias and erasure the world never seems to get enough of.

No.

It folds inward. It becomes indistinguishable from reverie, invention, and madness. Writing it down, even into the void of the internet where no one listens, and no one cares, gives it the one thing silence can’t: continued existence outside the tiny vessel of the mind that now contains it.

This is all I have to offer the world.