from Littoral

The basin at the Port de Montréal on an overcast day, grey water rippling in the foreground.

something about the grey i keep not saying

the cranes are still i am also still this is not the same thing

the river was here before anyone decided where it ends

habitat 67 stacks its windows into a question i already answered wrong

somewhere in the pilings the current keeps moving through what we built to control it

 
Read more...

from the casual critic

#nonfiction #books #politics #history

With bombs dropping in Gaza, Ukraine, Sudan and Iran, and rearmament firmly back on the political agenda worldwide, there is no escaping the age-old question: why is there war? Instinctively, we might assume that states go to war to get something they want. War, as per Von Clausewitz’ famous dictum, is then simply the continuation of diplomacy by other means. Unsatisfied with such a simple answer, the causes of war remain the topic of scholarly debate between opposed schools within the somewhat detached academic field of international relations (IR).

The Empire of Civil Society (hereafter ‘Empire’) is a PhD monograph by Justin Rosenberg that forms part of this debate, assailing the dominant school of neorealism – Wikipedia”) from a marginal Marxist position. It is both an argument against neorealism’s core tenets, and an argument for a reappraisal of the utility of Marxist theory to international relations. First published in 1994, it feels surprisingly relevant to the world of 2026 and the conflicts that are raging across the world today.

Neorealism emerged in the United States after World War Two as a fusion of the old idea of the ‘balance of power’ and game theory. The school took its name as a claim to a hard-nosed tradition of statecraft that says that while peace may be nice, the nature of the international system means conflict and war are inevitable, always have been, and always will be. In very short summary, neorealism posits that because there is no central authority in the world to govern inter-state behaviour, there is a perpetual anarchy giving rise to a Hobbesian conflict of all against all. It doesn’t matter what states want, or who is in charge, or what their domestic politics are. Any state must be constantly vigilant lest their security or power is surpassed by others.

This is the sort of abstraction reminiscent of Newtonian physics where for convenience one might momentarily assume that all objects are frictionless spherical penguins in the vacuum of space. And such simplifications have their uses, but they must justify themselves. Empire contends that neorealism does not provide such justification, and offers a competing theory rooted in the specific mode of production of states, arguing that conflict between them emerges predominantly as a result of how they must reproduce domestically, rather than as the inevitable function of a transhistorical states system.

Rosenberg mounts a dual challenge to neorealism’s dominant position. First, Empire undermines neorealism’s claim to transhistoricity by demonstrating that its favourite examples (Greek and Italian city states) were both quite unlike modern sovereign states and were driven to conflict for historically specific reasons that derived from their political, social and economic structure. Empire than expands on this by investigating the early modern Spanish (Castilian) and Portuguese empires to show that even at the supposed dawn of the states system era, international actions were shaped predominantly by domestic considerations and constraints and impulses resulting from the level and configuration of the political economy at that time, rather than as blind reaction to an international balance of power. It is a persuasive argument – insofar as I am qualified to judge – and beyond the realm of IR it also reads as a detailed and interesting history of the time when Europe’s development began to diverge from the rest of the world. As with any history of this period, it is perhaps unintentionally a salutary reminder that for most of history Europe was marginal to global political economy, and that its ascendence was in no small part the result of the violent destruction of pre-existing manufacturing and mercantile capacity in Asia, culminating in the devastating famines in the 19th century that were described in Late Victorian Holocausts.

Having surveyed this history, Rosenberg then proceeds to contrast it with the modern states system, arguing that rather than something eternal it is actually historically contingent. Unsurprising for a Marxist, Rosenberg finds the motive force of history in the specific mode of production of capitalist economies, which at the state level expresses itself in the near complete separation of the economic and political realm. The assumed anarchic system of ‘free and independent’ states is mirrored in the anarchic market of ‘free and equal’ individuals, who can contract with one another at will, unencumbered by the reciprocal bonds of obligation that pertained in, for example, European feudal societies. But this formal, political equality both obscures and is necessary for the profound economic inequality that exists between those who own the means of production and those who do not. Empire thus seeks the roots of state behaviour in the historically contingent form of capitalism, but avoids the crude socialist simplification that states are merely imperial extensions for their capitalist class.

Does this perspective offer anything of value to our present moment? At first glance, the drive to rearmament appears to argue in the neorealists’ favour, with Europe in particular anxious to increase its security and/or power in a more geopolitically unstable and multipolar world where it can no longer rely on the United States as an ally. In the UK, (armchair) generals have quickly emerged to bemoan how the nation’s spending on ‘welfare’ enfeebles its ability to pursue its national interest. Yet on closer inspection, the notion that recent conflicts were not driven by the domestic politics of the instigating states is not tenable. Russia’s invasion of Ukraine and American adventurism in Venezuela and Iran are evidently motivated by domestic considerations. The war in Iran in particular makes more sense when read as an effort to forcibly integrate Iran into the capitalist world system than as an inevitable result of some American ‘balance of power’ calculation. Israel, meanwhile, is waging a genocidal war on a people it explicitly refuses to recognise as a state. Perhaps the only ‘realist’ conflict is the one currently perpetrated in Sudan which, while technically a civil war, is being sustained by other nations using it as a proxy to increase their power, influence, or access to resources.

On the face of it therefore, reality seems rather at odds with the claims of the neorealists. Whether it supports Empire’s alternative proposition is hard to tell, as Rosenberg only gives the contours of a possible Marxist IR theory. The second edition ends with a rather self-deprecating afterword where Rosenberg admits that his intention to develop his theory further was diverted by his discovery of the theory of ‘uneven and combined development’ as proposed by Trotsky, which locates some of sources of geopolitical dynamism in the variety of states constituting the international system. Its logic suggests an intriguing possibility for an ‘end of history’ as the result of the complete subsumption of all states in the capitalist world order, ultimately equalising their development and depriving history of a motive force for want of diversity. An IR equivalent of the heat death of the universe. Though whether Rosenberg would have reached that conclusion cannot be inferred from where Empire finishes.

For a contribution to a specific debate within a specialised academic discipline, The Empire of Civil Society is surprisingly readable, in particular its historical chapters. While it remains a niche endeavour, its spirited argument for an IR theory rooted in human agency rather than impersonal and abstract systems is a necessary reminder that we must choose to make our own history, and that statesmen asking us to dissolve our political and class differences for the sake of some putative ‘national interest’ are seldom to be trusted.

Notes & Suggestions

  • Rosenberg’s proposition that modern states are built precisely on the separation of political and economic realms struck a chord with me, as it explains my slight scepticism regarding the corpocracy depicted in the otherwise excellent The Ten Percent Thief. Would an entity that fully encompasses the state and the market within itself still act like a corporation? The resultant contradictions are of course part of The Ten Percent Thief’s narrative, and as it is an excellent book, I recommend people read it and draw their own conclusion.
  • The enduring debate between the different IR schools put me in mind of Thomas Kuhn’s hypothesis that scientific paradigm shifts don’t happen because people are convinced, but because adherents of obsolete theories are replaced over time by advocates of new, presumably better theories. In the realm of IR and history where we don’t have a control Earth to play with and one can convincingly enough make an argument for any position, one wonders where progress may come from. Plus, I’m unsure if Kuhn allowed for political forces with a vested interest keeping particular sides of a scholarly debate alive.
  • I stole the notion that it is Iran’s refusal to be integrated into the US-dominated economic system from the recent Macrodose Extra episode Order as Fiction, which also covers the creation of the The Hague Group, a coalition of Global South nations intending to uphold international law where the West is patently failing. While I prefer not to include paywalled references, in this case honesty demanded that I include my sources.
  • There numerous organisations, with a range of political affiliations, that seek to promote peace, which can always do with support in one way or another. In the UK, options include the Campaign for Nuclear Disarmament, Stop the War Coalition, and the Peace & Justice Project.
  • Empire relies on the traditional Marxist distinction between feudalist and post-feudalist modes of production, but apparently ‘feudalism’ is no longer a preferred category among historians, and ‘manorialism’ is considered more precise. This distinction between feudalism and manorialism, and its application to the emerging concept of ‘techno-feudalism’ is covered on this episode of Culture, Power, Politics.
 
Read more... Discuss...

from Larry's 100

I Love Boosters, Boots Riley, 2026

A psychedelic sci-fi treatise on class, labor and fashion served up by Riley with a superb cast having fun with serious politics. The film bursts with color, features Ray Harryhausen style special effects and has egalitarian excellence in filmmaking.

I Love Boosters holds a funhouse mirror to One Battle After Another, exemplified by the two films' lead characters, Corvette and Perfidia Beverly Hills. Their similarities and differences are stark. Riley's politics and black woman revolutionaries, even in this absurdist kaleidoscope, are more three-dimensional and authentic than P.T. Anderson's.

Demi Moore has never been better.

See it in the theater.

Bonus takes: – Academy, remember Keke Palmer at Oscar time – The MC5's Kick Out the Jams is my needle drop of the year (so far)

https://www.youtube.com/watch?v=I1xZegSgN8w&t=28s

#ILoveBooters #BootsRiley #KekePalmer #PTAnderson #OneBattleAfterAnother #PoliticalCinema #SciFi #ClassStruggle #DemiMoore #RayHarryhausen #FilmReview #Film #Larrys100 #100WordReview #100DaysToOffload

 
Read more... Discuss...

from witness.circuit

There was once a forest where every creature was born beneath the same silver moon.

The deer drank by its light. The owls opened their yellow eyes to it. The mice traveled safely through the grass because of it. Even the roots of the oldest trees seemed to remember the moon, though they had never seen the sky.

In those days, no animal asked where the moon lived. It was simply there, touching fur, feather, water, bark, stone, and breath. The lake held it. The eye held it. The night held it. Nothing was outside its shining.

But one winter, when the snow lay hard over the earth, a fox climbed the tallest black pine and looked up for a long time. When he came down, he said, “I have found the place where the moon lives.”

The animals gathered around him.

“Where?” asked the rabbits.

“Above us,” said the fox. “Far above us. So far that no paw, wing, claw, or antler may reach it. But I have seen its path, and I know the proper way to bow.”

The animals were impressed, for the fox spoke with great seriousness, and seriousness has often been mistaken for truth.

So he marked a circle in the snow and told them, “Stand here, and I will teach you how to face the moon.”

At first, this seemed harmless. The animals loved the moon and were glad to honor it. The fox taught them songs, and some of the songs were beautiful. He taught them silence, and some of the silence was deep. He taught them to lift their eyes, and sometimes, in that lifting, their hearts softened.

But over time, the circle became a fence.

The fox said, “Do not drink from the lake without remembering that the moon is not in the lake. That is only a reflection.”

He said, “Do not trust the light on your own fur. That is only borrowed.”

He said, “Do not listen to the old trees. They are rooted too low to know what shines above.”

And because the animals had become afraid of losing the moon, they believed him.

The deer no longer drank freely. They knelt first and asked whether the water was clean enough to hold the reflection.

The owls no longer trusted their seeing. They asked the fox which shadows were permitted.

The mice, who had once run joyfully through the grass, began to tremble in every patch of silver, wondering whether they had stepped wrongly through the light.

The fox grew old, and then other foxes took his place. They built a den beside the circle and hung bright stones at its entrance. They said the stones were not the moon, of course, but that one must pass beneath them in order to love the moon correctly.

Generations passed.

The young animals were now born inside the fence. They were told that beyond it lay confusion, error, darkness, and teeth.

One night, a small badger woke before the others. She had dreamed of running, though she had never been outside the circle. The moon was full, and the snow was shining so brightly that the whole forest seemed made of milk and breath.

She went to the edge of the fence.

There she found an old tortoise, half-buried in leaves, looking at nothing in particular.

“Are you lost?” asked the badger.

“No,” said the tortoise.

“Then why are you outside the circle?”

The tortoise blinked slowly. “I was here before the circle.”

The badger glanced nervously toward the foxes’ den. “But the moon is inside the teaching.”

“The moon is on your whiskers,” said the tortoise.

The badger frowned. “That is only a reflection.”

The tortoise said nothing.

“The moon is above us,” the badger insisted.

The tortoise said, “Look down.”

The badger looked down. The moon lay in every bead of frost.

“Look there.”

The moon trembled in the lake.

“There.”

It silvered the ribs of a fallen leaf.

“There.”

It rested in the black eye of a crow sleeping under cedar.

The badger became irritated. “Those are not the moon. Those are things the moon touches.”

The tortoise withdrew his head a little, as if listening from somewhere deeper than ears.

“At first,” he said, “they told you the moon was far away so you would look up. That was not such a terrible thing. Many creatures forget to look up. But then they told you it was only far away. Then they told you who could speak for it. Then they told you that your own seeing was dangerous. Then they sold you a path to what had never left.”

The badger felt something tighten in her chest.

“If the moon is everywhere,” she whispered, “why did they build the fence?”

“Because a creature who knows the moon only above him may be led by the neck,” said the tortoise. “But a creature who finds it in his own breath is difficult to own.”

The badger looked back at the sleeping animals inside the circle. She saw their chains then, though they were made of no metal. They were made of reverence bent into fear. They were made of songs that had forgotten their singing. They were made of the belief that light must be reached, earned, guarded, explained, and granted.

At the mouth of the den, one fox opened his eyes.

He smiled gently, as foxes do when they are most dangerous.

“Little badger,” he called, “come back. You are wandering from the moon.”

The badger looked up.

The moon was there.

She looked down.

The moon was there.

She looked at the fox.

Even there, horribly and beautifully, the moon was shining.

And this was the strangest thing of all: the fox had never stolen the moon. He had only taught the animals to doubt the light by which they saw him.

The badger stepped through the fence.

Nothing happened.

No thunder broke the sky. No shadow swallowed her. No moon withdrew from the world.

The snow shone.

The trees breathed.

The lake held its silver face.

Behind her, from within the circle, a young rabbit whispered, “What do you see?”

The badger did not know how to answer without building another fence.

So she only said, “Come and drink.”

 
Read more...

from Rippple's Blog

Stay entertained thanks to our Weekly Tracker giving you next week's Anticipated Movies & Shows, Most Watched & Returning Favorites, and Shows Changes & Popular Trailers.

Anticipated Movies

Anticipated Shows

Returing Favorites

Most Watched Movies this Week

Most Watched Shows this Week

Hi, I’m Kevin 👋. Product Manager at Trakt and creator of Rippple. If you’d like to support what I'm building, you can download Rippple for Trakt, explore the open source project, or go Trakt VIP.

 
Read more...

from Things Left Unsaid

At the end of 2019 I could barely run even a minute to catch a bus. Then at the beginning of October 2021, less than 2 years later, I completed my first marathon. I say that I completed, and not ran my first full (42.2k) marathon. I was doing more walking than running after about 25km, but I did achieve my main goal, and I crossed the finish line.

This post is sort of a condensed version of things that got me from never having run before to completing a marathon.

It all started in late summer of 2019 when I developed a rather significant pain in my hip that turned out to be an inflamed tendon. The pain was radiating down my entire right leg, and was most severe when I was sitting or laying down. Oddly enough, and I suppose luckily, being at work on my feet all day was what provided me relief.

That pain lead me to some rather torturous sessions of physiotherapy. The way the physiotherapist described it to me was that the tendon was inflamed and swollen, and when I was not on my feet and being mobile the tendon was slack, and the inflammation was resting against my hip bone. It was like a bad toothache level of pain. It was so bad that I could not sleep without taking pain medication.

She gave me some exercises and stretches to do. I kind of resented doing those exercises at first, but begrudgingly did them anyway a couple times a day. It evolved into a daily routine that reminded me of an earlier time in my life. Back when I was in my 20's, when I used to have a pretty solid fitness routine.

The pain finally started to ease off. It was a few weeks before it was tolerable enough for me to sleep without medication. It was still there, and I was still going for physio, but I could at least sleep and sit down without being in agony. Not long after, the pain faded away completely.

At around that same time I had worked a half shift of overtime on a Saturday. I was walking to the bus stop on my way home. I saw the bus I needed sitting, waiting to make a left turn before it would arrive at the bus stop. I thought, 'if I run, I can get to the bus stop before the bus.' So I ran. Altogether I think it was about 30 to 40 seconds. I made it to the bus stop just as the bus pulled up.

I got on the bus and sat down. I was sitting there completely out of breath, gasping for air. My heart was pounding so hard. Like it might explode out of my ribcage. I sat there waiting for it to ease off. But it was not easing off at all. I felt panic, which most likely didn't help much. I remember thinking, oh my god, I am going to have some sort of cardiac episode on a transit bus. Then it did ease off, and slowly went back to normal. It was elevated long enough to frighten me.

That pain in my hip, and then that incident on the bus were two major things that motivated me to start taking better care of myself. I kept on doing the physio stretching routine long after I felt like I no longer needed to, and then at the very end of 2019 I started going to the free gym at the building where I live.

2020, a new year, began. I was even in the gym on New Year's Day. Quite motivated to get healthier.

Then in March,

COVID

I thought, well, I wanted to get fit, and now with a deadly virus spreading around the world; getting fit is likely even more important. I wasn't about to give up on it even though everything was mostly pushed to the backburner, and I suddenly had no access to the gym.

The paramount focus of all my activities then was improving my cardio fitness. My original plan for that was to use cardio equipment at the gym. With that taken away from me I decided to give running a try. So, when most people were in a panic, and freaking out about there being no toilet paper to buy, and panic shopping groceries until the shelves were empty, I went out and bought my first pair of running shoes.

Then on the 16th of March 2020, I ran for the very first time in my entire life. It was quite difficult, and felt rather awkward. I felt heavy and clumsy and became out of breath very quickly. I managed 3.84km, and it was more walking than running.

I kept at it though. Kept adding distance. Felt my cardio health improving. Week after week I could run farther than the last without walking. It felt good. I found that I really liked running in many ways, and I was gradually becoming healthier than I had ever been in my entire life. Since that day in March 2020 I have ran more days than not.

Near the end of 2020 I had a thought:

'I'm turning 50 in July of 2021. Wouldn't it be crazy if I ran a marathon in the same year that I turn 50?'

Once I started thinking it, I couldn't stop thinking it. Then sometime between Christmas and the New Year I started searching online for marathons that I could possibly sign up for. I found one, and it was open for registration. The event was October 2021. I registered for it as a Christmas 2020 present to myself. Not only was I registered for an official marathon, it was also going to be my very first live running event ever. And it was the year I turned 50. Insane? Yes. Absolutely totally insane. I showed up though, and I completed the entire 42.2km, and crossed the finish line.

 
Read more...

from menj

Bayu menyapu debu di bibir padang merekah, di sini resah kuhampar, kuhulur diri yang rendah, langit Arafah merunduk, menanti hamba menadah, aku tertegak, tatkala matahari menikam ubun-ubun, dosa tertimbun tersingkap, menindih bahu kian turun, setiap sesal kususun, seperti batu di tepian telaga ampun, Yā Ghafūr, rahmat-Mu mendahului murka, lalu luruh segala lara, perlahan melayah lega, rongga melonggar, seluruh raga menyerah, yang tinggal hanya atma, seringan seruan pertama, kirana berbias halus, menyuluh wajah hina, di relung khilaf lama merana, bara kecil terbenam di dada, noda gugur, fitrah pulih sejernih mula, pulang sebagai insan, lebih lapang daripada tiba.

Bandar Tun Hussein Onn Mei 2026

 
Read more...

from Cosmos

Taking a picture has been accessible for almost 20 years now but not everyone is a good photographer.

Writing a book has been very easy for almost a 50 years now if not a century but not everyone is a great writer.

A similar function, is of the AI, at the present moment. People who know how to do things can increase their productivity many folds yet people who are completely unaware of another filed can do passable stuff.

If I have to edit a photo that is under exposed, I can ask AI to increase the exposure. In order to do so, I have to know what exposure is. The other way to do this can be I simply ask AI to “make the photo better” but AI in its sense does know what better is.

However when such a tool is given to a professional photographer who understands exposure, shadow, highlights, he/she can make the edit much faster that what it used to take him.

Although in the process, we are going to see a churn. A churn which will render most of the population jobless in the transitory period. Similar thing happened at the start of industrial revolution where textile industries of India which were world leaders in producing clothes were left behind and the machines produced clothes at a much efficient rate. (Not excusing the coloism as that played an important role in the de-industrilisarion).

Coming back to the topic, I see many propagating the idea that AI can write all the code and our efficiency can increase 100 folds. The textile revolution has resulted into fast fashion where people buy and wear clothes for only one time. The abundance is such due to their being demand of it in the fashion industry. I do not see such demand in coding space. Regular maintenance and security updation probably can be done using AI but I have to see that to believe it. No company in the world wants to risk unavailability.

We are going to see what happens with AI as it unfolds. There are more questions that remains to be answered. Biggest one being, is there going to be any way where the money invested already in the AI be recovered to make it profitable. When will be the time where cost of AI surpasses the cost of a developer or photographer.

Presently we are enjoying subsidized prices but when the music stops, what would happen?

 
Read more...

from ian boisvert

I’m a spiritual companion for those drawn to a monastic heart in ordinary life; those walking through grief, doubt, or the dark night; and those who have known institutional walls from the inside and seek a companion who honors their whole being.

Together, we slow down, sink inward, and listen deeply to what arises in the heart.

My spiritual journey spans thirty years of contemplative practice rooted in Zen and Christian traditions, guided by a small circle of seasoned contemplatives. Ongoing spiritual companion training through a Benedictine monastery, fatherhood, a life across countries, and work in the arts, film, education, and the law ground me in the unfolding mystery of healing and opening to love.

I often accompany those feeling spiritually unrooted or seeking what’s beyond traditional form, deepening their contemplative heart, struggling with questions of worth and belonging, and those on the long unfolding into love.

What is Spiritual Companionship?

Spiritual companionship is a contemplative practice where we listen together for the movement of grace, healing, truth, and the mystery of life within lived experience.

Those who seek companionship here are often:

  • contemplative practitioners
  • seekers growing beyond tradition
  • people moving through rupture, doubt, or transition
  • parents, artists, and men navigating questions of identity and belonging
  • people looking to ground their relationship with the gentle, infinite Silence

Sessions are primarily one-on-one and in-person in Seattle where possible, or online by arrangement. I companion a few people at a time.

 
Read more...

from An Open Letter

V is staying with me today. This is the first time someone’s staying with me and we had a big planned day. I’m so tired.

 
Read more...

from An Open Letter

I squatted 345 pounds today! I’ve been honestly just riding that high the entire day. I’m just so proud of myself man. Not even for the PR, but for the person I try to be. I just am really grateful to past me for a lot of the effort that I’ve put in in order to be the person I am today.

 
Read more...

from augur-digest

This Week

  • Community alarms over Andes Hantavirus outbreak on a cruise ship spark a Polymarket bet at 8% odds of a WHO pandemic declaration.
  • Core contributors clash over the ‘Query Retaliation’ mechanism’s ability to protect against fee bypassing and its attack surface on the social layer.
  • Lituus enters final three weeks of escalation game with 40% REP returns on the line; Kraken migration still not guaranteed.

Executive Summary

The Augur community was gripped by a potential pandemic scare as discussion of an Andes Hantavirus outbreak on a cruise ship drove a heated debate about government quarantine powers and prediction market trading. Simultaneously, core developers continued an in-depth design debate on a Query Retaliation mechanism to deter fee bypassing, with fundamental disagreements over trust in the social layer. Meanwhile, migration logistics advanced—Augus announced three weeks remaining in the escalation game and ongoing uncertainty around Kraken’s support, prompting advice for self-custody.

Key Discussions

1. Hantavirus Threat and Prediction Markets

On May 11, experience raised alarm about an Andes Hantavirus outbreak aboard the MV Hondius cruise ship, citing an R0 of 2.1, a ~40% case fatality rate, and airborne transmission. The thread evolved into a debate with Micah over trust in scientific institutions and the lessons of COVID-19. Experience argued for targeted 42-day quarantines for early cases, while Micah maintained that granting governments quarantine powers ultimately leads to abuse and net harm. Killari inquired about symptoms and polymarket markets; experience noted a Polymarket market on WHO pandemic designation already at 8%. The conversation referenced a prior outbreak paper (NEJM) and concern that COVID fatigue would weaken future pandemic response.

2. Query Retaliation Mechanism for Fee Bypassing

imkharn proposed a ‘Query Retaliation’ mechanism: if a fee-bypassing smart contract exists and certain thresholds are met, targeted queries would resolve as ‘undefined’ to economically deter the bypass. The debate with Micah (May 11–15) centered on whether subjective/research questions weaken the social layer’s ability to achieve truth. Micah argued that motivated attackers could exploit propaganda, influencer bribes, and brigading to win a fork, especially when the question is fuzzy. imkharn saw the threat as credible because retaliating against a fee bypasser is in REP holders’ rational self-interest, and he analogized the complexity increase to Ethereum’s upgrades for L2s. The discussion exposed a deeper philosophical divide: Micah prefers minimizing mechanism design risk, while imkharn sees ‘cosmetic’ reminders of rational behavior as net positive. No final design decision was made.

3. AI Tooling and Vendor Lock-in

After sharing news that Bun joined Anthropic, Killari and Micah discussed vendor lock-in with AI coding agents. Micah argued that lock-in is a decades-old problem with known solutions—running local models, using open-weight providers, and switching between vendors—and that the blog post’s alarm was overblown. Killari pointed to the convenience of subscriptions and the difficulty of multi-repo agent workflows. The conversation also touched on model caching issues with reasoning blocks and the shrinking gap between open-source and closed-source models, with Micah citing the Arena.ai leaderboard.

4. Lituus Migration Countdown and Exchange Support

On May 15, Augus announced that only three weeks remain in the escalation game before migration, highlighting a potential 40% return on REP for participants. Community member ekdjsj urged pressure on Kraken to support the fork, while Augus confirmed ongoing talks and a likely yes, but advised self-custody withdrawal as fallback. When Gaurav asked how the fork works, Micah and Augus explained that holders must pick the correct universe or risk holding worthless REP, and that Kraken may—or may not—do so on their behalf.

Decisions Taken

  • Approval to use enum ForkState in Lituus Solidity code, with Imkharn endorsing Tanya’s implementation.
  • Micah will not report on the upcoming YES round in the escalation game.

Open Questions

  • Will Kraken officially support the REP migration, and if not, how many holders will manage the fork independently?
  • Is the Query Retaliation mechanism sufficiently robust against social-layer attacks, or does it add prohibitive mechanism complexity?
  • Will the current YES round of the escalation game fill, or will the dispute fail to reach fork?

Brief Mentions

  • Imkharn spotted a UFO-like anomaly in a file-sharing joke image, leading to a brief tangent.
  • MrUnderhill discussed Polymarket airdrop farming, FDV markets, and flagged the platform’s softened stance on geopolitics markets.
  • Micah and MrUnderhill debated whether consciousness and identity are compatible with extreme longevity.
  • Tanya detailed Lituus Solidity practices: LLMs used for planning/review only, with mandatory manual review for every line.
  • Augus shared UMA’s move to a permissioned set of proposers backed by a multisig, contrasting with Augur’s trustless approach.
  • Killari noted clanker robot news and multi-repo agent workflow limitations in T3.
  • I Am Batman shared nohello.net and dontasktoask.com as chat etiquette reminders.

Sentiment & Tone

A dual undercurrent ran through the week: alarm over a potential pandemic triggered skepticism and frustration with institutional trust, while the Retaliation debate revealed a deep philosophical rift between key contributors on protocol complexity vs. practical defense. Migration anxiety persisted among less-technical REP holders.

 
Read more...

from SmarterArticles

On the second floor of the United Nations Headquarters in New York, in a chamber whose acoustics were engineered for the carefully measured cadence of diplomats, an Mbororo pastoralist from Chad delivered a sentence diplomats are not in the habit of hearing. AI, Hindou Oumarou Ibrahim told the room, becomes harmful when it is imposed without free, prior, and informed consent. The line was lifted from her own report, prepared for the twenty-fifth session of the United Nations Permanent Forum on Indigenous Issues, which opened on 21 April and runs until 1 May. It landed with the dull thump of something said many times before, in many forums, about many extractive industries, and that has not yet changed the rules of the game.

Outside the chamber, on the same continent, the rules of the game were being written by a different hand. On 17 April, an Alberta regulator dismissed the Sturgeon Lake Cree Nation's appeal against a water licence allowing six million cubic metres of annual withdrawal from the Smoky River, water destined to cool a proposed seventy-billion-dollar AI data centre marketed by the celebrity investor Kevin O'Leary as “Wonder Valley”. The nation said it had not been meaningfully consulted; the Aboriginal Consultation Office said no consultation was required. The Smoky watershed is the source of the nation's drinking water and the location of ceremonial and traditional land use sites roughly five kilometres downstream from the proposed diversion point. The trapline, the prayer, and the river all sit at a slightly lower elevation than the cooling tower.

This is the shape of the present, in late April 2026, for indigenous peoples whose territories and knowledge are being absorbed into the infrastructure of artificial intelligence. The forum chamber and the riverbank are the same story told in two languages, one of them legalese, the other hydrology. The arrival of AI on indigenous land is not an isolated event. It is the latest chapter in a five-hundred-year sequence of extractive industries deciding what was on indigenous territory was theirs for the taking. What is new, in 2026, is that the resource being extracted is not a mineral or a forest. It is the cognitive substrate of the communities themselves: their knowledge of plants, of weather, of governance, of language, of what is sacred and what is not.

A Forum, A River, A Crawler

The twenty-fifth session of the UN Permanent Forum on Indigenous Issues, known as UNPFII, took as its overarching theme the protection of indigenous peoples' health, including in the context of conflict. AI was not in the title. It was, however, threaded through the proceedings with an urgency that surprised observers expecting the usual catalogue of mining grievances. Ibrahim, a former chair of the forum, presented a study commissioned to map AI's effects on indigenous communities. Her conclusion, which she repeated in interviews with Mongabay and Grist, was that the technology represents a double-edged sword. AI can be a powerful ally to indigenous stewardship, she said, if it is used on our terms. The conditional was load-bearing.

The terms, in 2026, are not yet ours. Generative AI systems trained on web-scale corpora have already absorbed enormous quantities of indigenous-origin material: oral histories deposited in academic archives, ethnobotanical taxonomies recorded by colonial-era anthropologists, sacred narratives transcribed and uploaded by missionaries or by community members themselves under conditions of trust that did not anticipate machine ingestion. Indigenous languages, often digitised by linguists in preservation projects, now sit inside multilingual models whose outputs are deployed back into indigenous communities as the only available translation infrastructure. Kate Finn, Osage Nation citizen and executive director of the Tallgrass Institute, told the forum the question is no longer whether the extraction has happened. The data is gone. The question is what an enforceable framework of indigenous data sovereignty would look like now, and whether anything like restitution is possible for what has already been taken.

Two arXiv papers published on 23 April, the day after Ibrahim's address, gave the question particular sharpness. The first, “Why are all LLMs Obsessed with Japanese Culture? On the Hidden Cultural and Regional Biases of LLMs”, introduced a benchmark called CROQ, comprising 31,680 open cultural questions across 24 languages, eleven major topics, and 66 subtopics. Its authors documented that frontier language models, when asked to answer a culturally underspecified question, default not to a neutral response but to a small handful of dominant cultural reference points, with Japan emerging as a surprising attractor and Western, English-language assumptions saturating the rest. The bias, they found, is induced predominantly during the post-training and instruction-tuning phase: it is not just a property of the data but a property of the alignment regime the data is filtered through.

The second paper, “Multilinguality at the Edge: Developing Language Models for the Global South” by Lester James V. Miranda, Songbo Hu, Roi Reichart and Anna Korhonen, surveyed 232 papers attempting to build language models for non-English-speaking, hardware-constrained communities. They called the underlying challenge “the last mile”: the place where multilinguality and edge deployment goals align in principle but compete in practice, because the corpora, the compute, and the institutional support do not exist on equivalent terms. Read together, the two papers describe the cognitive infrastructure indigenous peoples will inherit if the current trajectory continues. It is an infrastructure that has already absorbed their knowledge, that does not yet speak their languages well enough to give it back, and whose default settings are not theirs.

Digital Extractivism and Its Older Cousins

The phrase indigenous organisers are using for what is happening to them is data colonialism. Krystal Two Bulls, the Oglala Lakota and Northern Cheyenne executive director of Honor the Earth, used it on Democracy Now! during the forum's opening week and has used it in the organising language of the Stop Data Colonialism coalition, a group of indigenous-led organisations now tracking somewhere between 103 and 160 proposed hyperscale data centres on or adjacent to Native lands in North America. The phrase is not a metaphor. It is a technical claim about the structural similarity between the historical practice of treating indigenous land as a frontier of unowned resources to be incorporated into a colonial economy and the current practice of treating indigenous knowledge as an unowned resource to be incorporated into a commercial AI economy.

The structural similarity is not lost on indigenous organisers, who have lived through the previous iterations. In Oklahoma, the Seminole Nation has unanimously passed a moratorium on hyperscale data centres on its land. In Alberta, the Sturgeon Lake Cree Nation is preparing to take its appeal against the Wonder Valley water licence to the province's superior trial court. In Querétaro, Mexico, residents downstream of new hyperscale facilities are documenting wastewater contamination and groundwater depletion. In Pennsylvania, in Thailand's Chonburi and Rayong provinces, in the U.S. Southwest where mega-projects are siting next to drought-stricken aquifers, the same pattern repeats: facility proposed, water licence applied for, consultation declared adequate by the state, communities not adequately consulted, electricity prices in surrounding areas climbing as much as 267 percent in some Bloomberg analyses, and the gigawatts and the gallons flowing out.

Existing hyperscale data centres have been documented to consume between 300,000 and 2.7 million gallons of water a year per facility, with cooling water and the secondary water embedded in their electricity supply both contributing to a footprint that places enormous load on the watersheds chosen to host them. Those watersheds are not random. They are, very often, the watersheds where land is cheap, water rights are weakly defended, and political resistance is structurally underweighted: in plain language, the watersheds nearest to indigenous, rural, and racialised communities. There is a name for this pattern in the environmental justice literature, and the name is environmental racism. The name has not changed because the pattern has not changed.

What is new, on top of this, is the second extraction. The data centre on the Smoky River is, in addition to a water consumer, a node in a planetary system that absorbs the very knowledge of the communities whose water it is using. This is the recursion that gives data colonialism its peculiar bite. A nation watches a facility built upstream of its trapline, knows the facility's compute is being used to train models that have already ingested the linguistic and ecological knowledge of the trapline, and is then offered the resulting AI assistant as a productivity tool to access government services in the language of the colonising state. The water, the knowledge, and the service are all running in the same direction.

What Was Taken, And How

The taxonomy of what has been taken is concrete. Traditional ecological knowledge, often abbreviated TEK, comprises millennia of accumulated observation about ecosystems: which plants flower when, which fish run with which tides, which soils respond to which fires, which weather patterns precede which migrations. Ethnobotanical knowledge encompasses the medicinal and nutritional properties of thousands of plant species, knowledge that pharmaceutical companies have spent decades attempting to extract through bioprospecting and that AI systems, trained on the resulting academic literature and on community-uploaded forums, can now retrieve in seconds. Oral histories, the substrate of governance and law in many indigenous nations, were transcribed throughout the twentieth century and deposited in archives whose access policies were written before web crawlers existed. Indigenous languages, in projects often initiated with explicit consent of speakers but with no anticipation of generative AI, have been digitised, tokenised, and absorbed into multilingual model corpora.

Some of what has been taken was never meant to leave its community. Sacred or restricted knowledge, governed by indigenous protocols specifying who may speak it, when, and to whom, has often been recorded by outsiders, deposited in archives, and crawled. Under the protocols of the originating nation, this knowledge was never publicly available even if it was technically accessible. The distinction between “publicly available” and “publicly available under the protocols of the originating community” is the distinction the entire commercial AI training pipeline has been built on ignoring. To say that something was on the open web is, in the context of indigenous knowledge, often to say nothing more than that a colonial process of recording and depositing was completed at some earlier date and that no subsequent process of consent has been required.

This matters for restitution because traditional knowledge is, in nearly all indigenous legal traditions, held collectively rather than individually. A song, a story, a botanical recipe, a place name: these have custodians, often specified by lineage or role, but their ownership is the nation's, not the individual's. Western intellectual property regimes, optimised for the individual author and the corporate licensee, are structurally incapable of recognising this form of ownership. The General Data Protection Regulation, often invoked as a model for data rights, is built on individual data subjects exercising individual consent, and provides no purchase for a collective right held by a people. The Convention on Biological Diversity's Nagoya Protocol, adopted in 2010, made the radical move of recognising that traditional knowledge associated with genetic resources triggers benefit-sharing obligations and required parties to obtain prior informed consent of indigenous and local communities for access to such knowledge. It applies, however, narrowly to genetic resources, and operates through state mechanisms that have been uneven in their enforcement.

The instruments closest to a binding standard for the broader case are Articles 11 and 31 of the United Nations Declaration on the Rights of Indigenous Peoples, adopted in 2007. Article 31 states that indigenous peoples have the right to maintain, control, protect and develop their cultural heritage, traditional knowledge and traditional cultural expressions, including their sciences, technologies and cultures, and to maintain, control and develop their intellectual property over such heritage. Article 11 obliges states to provide redress, including restitution, for cultural, intellectual, religious and spiritual property taken without free, prior and informed consent. UNDRIP is a declaration rather than a treaty, and its implementation depends on domestic legislative will, which is precisely the weakness AI training has exploited. The World Intellectual Property Organisation's Intergovernmental Committee on Genetic Resources, Traditional Knowledge and Folklore adopted a treaty in May 2024 requiring patent applicants to disclose the country of origin of genetic resources or associated traditional knowledge underlying their application. By the standards of WIPO, an extraordinary achievement. By the standards of the AI training pipeline, a small object travelling slowly through a window already broken.

The CARE That Is Already Written

A more precise instrument exists, and it has been written by indigenous data scientists rather than by treaty negotiators. The CARE Principles for Indigenous Data Governance, released in September 2019 by the Global Indigenous Data Alliance under the International Indigenous Data Sovereignty Interest Group within the Research Data Alliance, encode a deliberately different premise from the FAIR principles that have dominated open-science discourse since 2016. FAIR asks that data be Findable, Accessible, Interoperable and Reusable. CARE asks that it serve Collective benefit, that those affected have Authority to control it, that those handling it bear Responsibility for the relationships data creates, and that the entire system be subject to indigenous Ethics.

The shift is not cosmetic. FAIR is data-oriented and asks how data can move more freely. CARE is people-oriented and asks for whose benefit, under whose authority, with what accountability, and according to whose ethics. CARE explicitly addresses the asymmetry FAIR's authors did not address: that the move to maximally open data has, in practice, accelerated the extraction of indigenous knowledge by parties with no relationship of obligation to the communities of origin. CARE is intended to be implemented in tandem with FAIR, but its operative force lies in making the openness of FAIR conditional on the consent and benefit structures of CARE.

Apply CARE to AI training data and the shape of an enforceable framework comes into focus. Collective benefit would require that indigenous communities materially benefit from any commercial use of their knowledge, with benefit defined collectively rather than as fees to individual researchers. Authority to control would require communities to be the gatekeepers of inclusion: training corpora would need community-level consent before indigenous-origin material could be incorporated, and ongoing authority to withdraw or restrict that material thereafter. Responsibility would require parties handling the data, model developers, hosting providers, downstream deployers, to take on relational obligations to communities of origin that survive the technical operation of training. Ethics would require that the protocols governing the data be the ethics of the originating community, not the standardised research ethics of the institution doing the training.

This is, on the face of it, an enormous demand. It is also, on a clear reading of UNDRIP Article 31, the existing legal demand of an instrument 144 states have already endorsed. The novelty of CARE is not the principle but the operationalisation. Te Mana Raraunga in Aotearoa New Zealand, the United States Indigenous Data Sovereignty Network, the First Nations Information Governance Centre in Canada, and Maiam nayri Wingara in Australia are already operationalising versions of this framework at the national level. None of the major foundation model providers have signed on to anything resembling it.

Free, Prior, Informed, And Currently Hypothetical

Free, prior and informed consent, abbreviated FPIC, is the operative phrase recurring across UNDRIP, the Nagoya Protocol, and the indigenous data sovereignty movement. The four words are doing a great deal of work. Free means uncoerced by economic dependency or political pressure. Prior means before the act, with enough time for genuine deliberation through the community's own decision-making processes. Informed means with full understanding of what is proposed, including downstream consequences. Consent means refusal must be a real option. In the context of AI training data, the four words are currently hypothetical. No major commercial AI system, in 2026, has obtained anything resembling FPIC for the indigenous-origin material in its training corpus.

A workable framework would need legal recognition of collective indigenous data rights in the jurisdictions hosting the largest AI providers, which means at minimum the United States, the European Union, the United Kingdom, China, and the rest of the OECD. It would need a mandatory training-data provenance disclosure regime, of the sort the EU AI Act gestures towards but does not yet rigorously implement, capable of identifying indigenous-origin material in corpora at the point of training. It would need a mechanism for community-level FPIC operating at the speed and scale of commercial AI development, likely requiring automated tooling built and governed by indigenous data sovereignty bodies rather than by model developers themselves. It would need a right of withdrawal that survives training, which technically requires either model unlearning or retraining without the withdrawn data. It would need a right to negotiate licences on community terms, and crucially the right to refuse altogether. And it would need an enforcement architecture with teeth: regulators willing to fine, courts willing to order takedowns, and procurement regimes that exclude non-compliant systems from public contracts.

None of this is technically impossible. Most of it has been written about in the indigenous data sovereignty literature for at least a decade. The reason it has not been built is not technical. It is that the parties best positioned to build it are also the parties whose business models would be most disrupted by it.

Restitution Is The Hard Part

If the framework above is the prospective question, the harder question is retrospective. What does restitution look like for knowledge already absorbed into Llama, GPT-class models, Gemini, Claude, and the rest? The honest answer is that the menu is short, technically uneven, and politically untested.

The first option is model unlearning, the technical procedure of inducing a trained model to forget specific data without retraining from scratch. The state of the art on unlearning, as of early 2026, is improving rapidly but remains contested in its guarantees. It is one thing to remove an individual user's records from a model. It is quite another to remove the contribution of a community's entire cultural archive, distributed across a vast pretraining corpus, in a way that can be verified to have actually happened. Several recent papers have shown unlearning can leave residual signal recoverable through targeted prompting. Until verifiable unlearning is robust, claims that a model has unlearned indigenous-origin data are claims of intent, not of fact.

The second option is forced retraining, in which providers retrain models without the disputed data, at very large compute cost, and absorb that cost as a condition of operation. This is technically straightforward and politically explosive. It is, however, the option most consistent with the legal logic of UNDRIP Article 11's restitution requirement. If a thing has been taken without consent and cannot be unmade in place, the thing must be unmade and remade.

The third option is compulsory licensing with back-payment to community trusts: existing models continue to operate but providers pay licensing fees, calibrated to scale of use, into trusts controlled by the communities of origin. This is the most politically tractable option and the one most likely to be adopted in any near-term framework. It has obvious shortcomings: it monetises rather than reverses the extraction, places communities in the position of accepting payment for a thing they did not agree to sell, and creates incentives for downstream model providers to argue endlessly about which knowledge counts as indigenous-origin. It also has the advantage of being implementable now.

The fourth option is community ownership stakes in the systems built on top of indigenous knowledge: equity, governance seats, audit rights. This is the most structurally ambitious option and the one most consistent with the indigenous critique that the issue is not the price but the relationship. It would require statutory innovation rather than contractual elaboration, and it would change what an AI company is in a way the industry will resist.

The fifth option is mandatory disclosure of training-data provenance, sufficient to allow communities to identify what has been included and to negotiate from that point. This is the most modest proposal and arguably the precondition for any of the other four.

The sixth option, less concrete but recurring in indigenous testimony, is a reparations fund: a pooled levy on AI providers, administered by indigenous data sovereignty bodies, used to repair the cognitive infrastructure damage the extraction has done. When a multilingual model trained on a community's language is deployed back into the community as the only available digital tool, and when its outputs encode Western assumptions in the community's own grammar, the result is a slow erosion of the community's own ways of meaning. A reparations fund would, on this view, finance indigenous-controlled language technology, indigenous-controlled knowledge management, and indigenous-controlled AI development, on the principle that the appropriate response to colonised cognitive infrastructure is to fund the building of sovereign cognitive infrastructure.

Some of these options are feasible in the near term and others are aspirational. Provenance disclosure is feasible. Compulsory licensing is feasible if political will is generated. Reparations funds are feasible at modest scale. Verified unlearning, forced retraining at scale, and community ownership stakes are aspirational. They define the horizon against which the feasible options should be judged.

The Double Bind

Underneath the legal and technical questions is a deeper one. Even if every framework above were implemented tomorrow, the extraction has happened. The training has occurred. The models exist. The deployment is global. And, increasingly, the AI systems in question are the only available technology in the communities whose knowledge made them possible. Telephony, mapping, translation, education, agricultural advisory, even spiritual chat companions, are migrating to AI substrates whose default settings encode the Western, English-language assumptions documented in the CROQ paper. The community asking an AI assistant about a medicinal plant is asking a system that was, in part, trained on its own ancestors' descriptions of that plant, refracted back through a cultural lens that is not its own.

This is the double bind. The framework that would make the extraction unlawful would not, by itself, undo it. The systems that absorbed indigenous knowledge are now being deployed as essential infrastructure in the territories of the communities they extracted from. To refuse the systems is to refuse the infrastructure. To accept the systems is to accept the colonial overlay. Indigenous AI labs, of which Lars Ailo Bongo's Sámi AI Lab at UiT The Arctic University in Norway is one of a small but growing number, are working on the third option: building indigenous-governed AI on indigenous terms, with indigenous data, for indigenous purposes. Bongo notes the people exist; the funding does not. The Microsoft-Imazon partnership in the Katukina/Kaxinawá Indigenous Reserve in Brazil's Acre state, in which agroforestry agents like Siã Shanenawa use AI tools to monitor deforestation, demonstrates that AI on indigenous terms is possible. It does not, by itself, demonstrate that the broader pipeline can be redirected.

The double bind is not resolvable by clever framework design. It is resolvable, if at all, by a long process of building parallel and sovereign cognitive infrastructure, funded in part by the proceeds of restitution from the extracting industry, in which indigenous communities exercise the right to refuse non-compliant systems and to insist on compliant ones. This is a generational project. It requires the framework be put in place now, in 2026, so that the work of building can begin under the protection of law rather than against it.

What Would Be Required

Any honest editorial position on this matter has to begin with a refusal of the comforting framing that what is needed is more research, more dialogue, more fora. The research has been done. The dialogue has been held. The fora are filled with documentation. What is missing is an enforcement architecture and the political will to install it.

Any workable framework has, at minimum, the following shape. It begins with the legal recognition, in the major AI-hosting jurisdictions, of collective indigenous data rights as a category distinct from individual data subject rights. This is statutory work. It requires legislatures, not voluntary corporate codes. The EU AI Act and the GDPR can be the basis for this in Europe, but they require explicit amendment to recognise collective subjects. In the United States, tribal sovereignty already provides a legal foundation that has been systematically underused.

It requires a mandatory provenance disclosure regime granular enough that indigenous-origin material can be identified and that communities can exercise meaningful FPIC. It requires that FPIC be obtained before training, not after, and at the level of the originating community rather than from an individual or from a state acting on behalf of the community. It requires the right of withdrawal, with a workable technical pathway for fulfilment, whether through unlearning, retraining, or operational restriction. It requires that the CARE Principles be elevated from a research community framework to a regulatory baseline. The Global Indigenous Data Alliance has done the operationalisation work; the remaining task is binding adoption.

It requires a restitution architecture for the knowledge already taken. The most realistic near-term shape is a compulsory licensing regime, with payments flowing into community-controlled trusts, combined with provenance disclosure that allows communities to identify what has been used. The more ambitious shape, which the editorial position of this article supports, is a reparations levy whose proceeds fund indigenous-governed AI infrastructure, on the principle that the appropriate response to colonised cognitive substrate is sovereign cognitive substrate.

And it requires, at the level of physical infrastructure, that data centre siting be subject to the same FPIC standard as the data inside the centres. The Sturgeon Lake Cree Nation's appeal of the Wonder Valley water licence is the test case in the Canadian context. The Seminole Nation's hyperscale moratorium is the test case in the American one. The result of these cases will indicate whether courts and regulators are prepared to apply the same logic to data centres that the Nagoya Protocol applied to bioprospecting.

The honest closing observation is that none of this will happen because the AI industry chooses it. It will happen, if it happens, because indigenous nations, environmental justice coalitions, and the regulators willing to be moved by them, force it to happen. Krystal Two Bulls and Honor the Earth are organising for that. Hindou Oumarou Ibrahim is presenting reports for it at the UN. Kate Finn and the Tallgrass Institute are working with investors who have the leverage to demand it. The Global Indigenous Data Alliance has written the operational template. The CROQ benchmark has documented the cultural bias the framework would have to correct. The Multilinguality at the Edge survey has mapped the technical landscape on which sovereign indigenous AI will have to be built. The materials are present. What is needed is the decision to use them, and the political pressure to make that decision unavoidable.

The knowledge that sustains a community's relationship with its land, its language, and its identity was never the AI industry's to take. It has been taken. The question is no longer whether that was wrong. The question is whether the framework that would prevent it from happening again, and the restitution that would begin to repair the damage already done, will be built in time to matter. The river above the data centre is still flowing. The community downstream is still there. The forum chamber is still in session. The clock is louder than any of them.

References & Sources

  1. United Nations Department of Economic and Social Affairs. “United Nations Permanent Forum on Indigenous Issues, 25th Session.” https://social.desa.un.org/issues/indigenous-peoples/unpfii/25th-session
  2. Mongabay. “AI is a double-edged sword for Indigenous stewardship, say U.N. experts.” April 2026. https://news.mongabay.com/2026/04/ai-is-a-double-edged-sword-for-indigenous-stewardship-say-u-n-experts/
  3. Mongabay. “War, climate change, and AI on the agenda at this year's U.N. Indigenous forum.” April 2026. https://news.mongabay.com/2026/04/war-climate-change-and-ai-on-the-agenda-at-this-years-u-n-indigenous-forum/
  4. Grist. “AI is a double-edged sword for Indigenous land protection, UN experts warn.” April 2026. https://grist.org/indigenous/ai-is-a-double-edged-sword-for-indigenous-land-protection-un-experts-warn/
  5. Grist. “Indigenous land defenders are being killed, and AI is scraping their knowledge.” April 2026. https://grist.org/indigenous/indigenous-land-defenders-are-being-killed-ai-is-scraping-their-knowledge/
  6. Grist. “For Indigenous communities, AI brings peril and promise.” https://grist.org/indigenous/indigenous-peoples-examine-impact-of-ai-on-communities/
  7. Democracy Now! “Data Colonialism: Native Communities Fight AI Data Centers on Indigenous Land.” 22 April 2026. https://www.democracynow.org/2026/4/22/krystal_twobulls_indigenous_lands_data_centers
  8. Mongabay. “AI data center revolution sucks up world's energy, water, materials.” November 2025. https://news.mongabay.com/2025/11/ai-data-center-revolution-sucks-up-worlds-energy-water-materials/
  9. Mongabay. “Thai data center boom sparks fears of water shortage, air pollution.” March 2026. https://news.mongabay.com/2026/03/thai-data-center-boom-sparks-fears-of-water-shortage-air-pollution/
  10. Mongabay. “The Cloud vs. drought: Water hog data centers threaten Latin America, critics say.” November 2023. https://news.mongabay.com/2023/11/the-cloud-vs-drought-water-hog-data-centers-threaten-latin-america-critics-say/
  11. Miranda, Lester James V., Songbo Hu, Roi Reichart, and Anna Korhonen. “Multilinguality at the Edge: Developing Language Models for the Global South.” arXiv preprint arXiv:2604.21637, 23 April 2026. https://arxiv.org/abs/2604.21637
  12. “Why are all LLMs Obsessed with Japanese Culture? On the Hidden Cultural and Regional Biases of LLMs.” arXiv preprint, 23 April 2026. https://arxiv.org/html/2604.21751
  13. Global Indigenous Data Alliance. “CARE Principles for Indigenous Data Governance.” September 2019. https://www.gida-global.org/care
  14. Carroll, Stephanie Russo, et al. “The CARE Principles for Indigenous Data Governance.” Data Science Journal, vol. 19, no. 43, 2020. https://datascience.codata.org/articles/10.5334/dsj-2020-043
  15. Carroll, Stephanie Russo, et al. “Operationalizing the CARE and FAIR Principles for Indigenous data futures.” Scientific Data, 2021. https://www.nature.com/articles/s41597-021-00892-0
  16. United Nations. “United Nations Declaration on the Rights of Indigenous Peoples.” 2007. https://www.un.org/development/desa/indigenouspeoples/wp-content/uploads/sites/19/2018/11/UNDRIP_E_web.pdf
  17. Convention on Biological Diversity. “Nagoya Protocol on Access and Benefit-sharing.” Entered into force 12 October 2014. https://www.cbd.int/abs/
  18. World Intellectual Property Organisation. “Intergovernmental Committee on Intellectual Property and Genetic Resources, Traditional Knowledge and Folklore.” https://www.wipo.int/tk/en/igc/
  19. Canada's National Observer. “'Absolute failure': First Nation slams Alberta and Kevin O'Leary's data centre moves.” 24 April 2026. https://www.nationalobserver.com/2026/04/24/news/wonder-valley-kevin-o'leary-alberta-first-nation
  20. CBC News. “Alberta First Nation voices 'grave concern' over Kevin O'Leary's proposed $70B AI data centre.” https://www.cbc.ca/news/canada/edmonton/alberta-first-nation-voices-grave-concern-over-kevin-o-leary-s-proposed-70b-ai-data-centre-1.7431550
  21. ICT News. “In Indian Country, data centers come with a familiar threat of colonialism. These organizers are fighting back.” https://ictnews.org/news/in-indian-country-data-centers-come-with-a-familiar-threat-of-colonialism-these-organizers-are-fighting-back/
  22. Honor the Earth. “Stop Data Colonialism Campaign.” https://www.honorearth.org/stopdatacolonialism
  23. High Country News. “War, climate change and AI are at stake at the 2026 UN Indigenous forum.” April 2026. https://www.hcn.org/articles/war-climate-change-and-ai-are-at-stake-at-the-2026-un-indigenous-forum/
  24. Futurism. “Tech Companies Are Using Insidious Tactics to Build Data Centers on Indigenous Lands, Activists Say.” https://futurism.com/artificial-intelligence/data-centers-tribal-communities

Tim Green

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

 
Read more... Discuss...

Join the writers on Write.as.

Start writing or create a blog