La ville, pourtant toute en collines, en vagues rocheuses, immobiles – on loge dans ses moindres plissures – se figure [à soi (disant)] telle un trou par lequel passer sans cesse. De la fenêtre, trop petite, mesquine, on peut voir une partie du trou, et deviner, à la rumeur, aux sonorités, aux lueurs, ce qui demeure caché. [Quand la porte est ouverte, il est préférable de regarder en se tenant dans le fond du couloir, où règne une odeur de bonde pas lavée, de vaisselle et d’égout] : on voit dans l’embrasure rectangulaire le carré du trou de la ville, de nuit surtout, tant la ville est nuit.

#Fenêtresurville #Didascalies

Fenêtre sur ville

八階まで吹き抜ける光は妙に白く、棚という棚が過剰に照らされていた。雑貨も食品も衣類も玩具も、分類されているようでいて実際には混ざり合い、何でも売っている巨大な小売店の中で、私は灰色の股下が長いジャージを履いていた。普段なら選ばない格好なのに、ここではそれが妙にしっくりきた。たまにしか来ない場所なのに、なぜか落ち着く。

何を買いに来たのかは分からなかった。ただ、クレーンゲームの列が放つ色とりどりの光を眺めながら、流れに身を任せて上から下へ降りていく。それだけで十分だった。

景色の切れ端は、エスカレーターの移動とともに勝手に浮かんでは沈んでいく。昔、この店の入口近くで見かけた人のことを思い出す。恐竜の形をした妙な履き物を引きずりながら歩いていたその人は、顔の輪郭さえ曖昧だった。それでも全体の空気だけで心が傾いた記憶がある。人を好きになる理由など、本当は顔でも言葉でもなく、遠くから見た重心の置き方のようなものなのかもしれなかった。

七階の衣料品売り場を抜けると、地下通路の途中にあるような古いアーケード街を思い出した。昭和の婦人服店が横一列に並び、地上から差し込む光だけが時代から取り残されていた場所だ。連れが体調を崩して建物の奥へ消えたあの日、私はただ長い通路の端まで歩き、折り返して戻った。何かを待つ時間というのは、歩くこととよく似ている。目的地は最初から存在せず、引き返すことまで含めて一つの移動なのだ。

六階まで降りると、学校の記憶が混ざり始めた。机を端へ寄せ、椅子だけを輪にして並べた教室。順番に理由を述べていく時間。遊びたいから、誰かと話したいから、そんな種類の答えはすぐに出尽くした。最後のほうで、普段ほとんど目立たない生徒が予想外の角度から言葉を落とした瞬間、空気は奇妙な方向へねじれていった。自分の番が来た頃には残された表現がなくなっていて、口を開く前から全員の視線だけが集まっていた。あの感覚は今も覚えている。選択肢が多いように見えて、実際には何も選べない瞬間の重さを。

五階では家具が並んでいた。そこを歩いていると、自分の家を思い出した。安心した途端に身体が冷えていくような感覚。感情の起伏が少なく、表情もほとんど変わらない生活。その単調さが部屋の形と不思議に調和していた。落ち着く場所というのは、必ずしも幸福な場所ではない。ただ変化が少ないだけの場合もある。

四階に降りるころには、昔の職場の景色が現れた。シーツを伸ばし、皺を消しながら働いていた頃、年上の誰かが忘れた夢について問いかけてきたことがある。また別の日には、他人の視線がどこへ向いていたのかを指摘する感覚を初めて理解した。人は相手を見ているつもりで、自分の見たいものしか見ていない。思い出の中の顔もまた、いくつかの決まった表情に整理されて保存されている。

三階では、自立だけを唯一の正解として語る大人たちの声が聞こえる気がした。一人で生き、一人で決め、一人で立つこと。それだけが成熟の形として提示される世界。しかし本当は、多くの人が流れに乗せられながら移動しているだけなのではないか。今の私がエスカレーターに運ばれているように。

二階の食品売り場では、魚や惣菜の匂いが漂っていた。その匂いは旅館の夕食を連想させた。卓上を埋め尽くすほどの料理の豊かさに身を委ねていると、それまで気にしていた小さな失敗など自然と遠ざかっていく。熱い油の中で具材が静かに煮える小皿料理の香りまで思い出し、空腹とも懐かしさともつかない感覚が胸の奥を通り過ぎた。

そして一階。

自動ドアの向こうに夜の街が広がっている。出会った頃は電車で向かった場所へ、いつからか車で行くようになったことも思い出す。公園の入口に咲く桜は美しかったが、本当に心を動かしたのは、その先の橋を渡ったあとに広がる風景だった。入口はいつも入口に過ぎない。

店を出ても何かを買った記憶は残っていなかった。ただ、上から下まで降りてきた時間だけが身体の中に残っている。次の日もまた、その続きを生きる気がした。夏休みが終わらずにどこかへ伸びていくような感覚。そして、実際には持っていないはずなのに、乗ったことのないオープンカーの鍵だけをポケットに入れているような、不思議な余裕があった。

Today was the first day of my business trip that I get to spend with someone and I hung out with A like I normally do. It was a wonderful time as always, and you’re the end of the night I told him about how he is my gold standard of humor in a person, and how I don’t think that is a reasonable goal. He also mentioned that he felt the same which I thought was sweet. I’m just so incredibly grateful that I got to know him, and that I get to have him as a lifelong friend.

Some days when I am sleep deprived and lonely, I just want to see the world burn, and on those days, my mind goes into dark, but also weird places.

Some symptoms are

1. When i see despair in someone’s eyes i feel extreme happiness

2. I visualize a violent death for myself and feel the same extreme happiness

I would say though that typically the dark thoughts I have are directed inward instead of outward. I usually have no desire to harm others. But I do sometimes visualize myself on a strange operating table, bound by thick metal wires, and the flesh on my limbs spread apart in half with a straight cut down the middle to expose bone. The happiness I feel is even stranger, a frenetic happiness that causes deranged laughs to escape from my lips. It’s a combination of feeling stimmed and despairing at my life and hating everything that I am. And because I feel pain and feel isolated from others, I wish the others could be just as unhappy as me and know me through that. So this culminates in a wish for the world to burn, along with an exciting, violent end to my existence.

On 4 February 2026, the Ahrefs content team published a single chart that should have been treated like a public health alert. It showed that when an AI Overview appears at the top of a Google search results page, the top-ranked organic link beneath it now receives 58 per cent fewer clicks than the same page would have received before AI Overviews existed. In April 2025, the same analysis had measured the decline at 34.5 per cent. In nine months, a feature initially described as an enhancement to search has roughly doubled the amount of traffic it diverts from the websites whose content it summarises. The 300,000 keywords Ryan Law and Xibeijia Guan analysed were not edge cases. They were the queries the open web has historically depended on for its survival.

That chart did not make front pages. AI Overviews did.

The numbers it represents arrive at a peculiar moment. In late January 2026, individual publishers logging into their AdSense dashboards discovered overnight earnings had collapsed by anything between fifty and ninety per cent. Daily revenue of five hundred dollars fell to thirty-five. Country-level coverage dropped to figures that read like a postscript: Germany down sixty-four per cent, France sixty-three, Italy seventy-six, Spain ninety. Google later attributed the failure to third-party tag recognition issues in Ad Manager that cascaded through Ad Exchange. Whatever its proximate cause, the incident gave thousands of small publishers a glimpse of what life on the other side of the search economy now looks like. The trapdoor had not opened. It had merely groaned.

Press Gazette's May 2026 audit of the fifty most popular US news sites confirmed the broader pattern. Almost half had seen their year-on-year traffic in April fall by 20 per cent or more, despite a steady stream of breaking news. Forbes recorded the worst decline among the top fifty, down by nearly half. AP News had shed 46 per cent of its visits in a year. Athlon Sports lost 48 per cent. Some publishers grew, most notably Substack's newsletter network and Men's Journal, the latter quadrupling its traffic, but the direction of travel for the legacy news web was unmistakable: down, and accelerating.

This article is about that descent. It is also about who, in the end, gets to decide whether it continues.

The shape of the disappearance

Begin with the mechanics, because the mechanics matter. AI Overviews are the AI-generated summaries that now appear above traditional search results on Google for a growing share of queries. They are produced by Gemini, Google's family of large language models, drawing on a vast index of web content that includes, of course, the journalism that publishers have spent decades producing. The Overview answers the user's question directly. The links to the underlying sources remain, but they sit underneath an answer that has, by design, made clicking them unnecessary.

For two decades, this would have been considered a hostile redesign of the search results page. Search engines extracted value from the open web by indexing it, but they returned value through referrals: a query went in, a link came out, the publisher received a visitor who could be shown advertisements, persuaded to subscribe, or counted in the metrics that justified the salary of the reporter who wrote the piece. It was an implicit bargain rather than a contract, but it was durable. The bargain is now visibly breaking.

Cloudflare's data on the asymmetry between what AI companies take and what they give back is unsparing. In June 2025, Google's crawler scraped websites fourteen times for every referral it sent. OpenAI's crawler scraped 1,700 times for every referral. Anthropic's, the most extractive of the major systems, scraped 73,000 times for every visitor it sent on. By July 2025, Anthropic's ratio had risen to 38,000 pages crawled for each referred page visit, an imbalance that, as Cloudflare's Matthew Prince has argued repeatedly, is incompatible with the survival of the businesses that supply the underlying content. The Security Boulevard analysis published in April 2026 framed it crisply: large language models scrape publisher content thousands of times for every single referral they send back, destroying the advertising and subscription revenue that pays for the reporting being consumed.

There is something almost geological about the slowness with which this realisation has settled. Search engines have always taken more than they gave back in any narrow accounting; the value they created accrued elsewhere, in the form of an open web worth searching. AI search dispenses with that justification. The user gets the answer. The model gets the training data. The platform gets the advertising slot at the top of the page. The publisher gets a citation, sometimes, in a panel that the user is empirically unlikely to expand.

The case studies are not anecdotes

The temptation, when writing about systemic decline, is to reach for individual stories that humanise the abstraction. The trouble with the AI Overviews story is that the abstraction has already eaten the stories whole, and the bodies are not metaphorical.

The Planet D, a travel blog founded in 2008 by Dave Bouskill and Debra Corbeil, lost half its traffic in the months after Google launched AI Overviews in May 2024. Staff were laid off. Traffic then fell another 90 per cent. The site stopped publishing earlier in 2025. Charleston Crafted, a home improvement blog, lost 70 per cent of its traffic between March and May 2024 and saw a 65 per cent decrease in advertising revenue. Stereogum, one of the longest-running independent music publications on the open web, reported a 70 per cent collapse in ad revenue in 2025. Its founder, Scott Lapatine, attributed most of the damage to AI Overviews, though Facebook's and X's deprioritisation of links played a supporting role, and announced a transition to paid subscriptions in the hope of replacing what the platform had taken.

These are not boutique websites that misjudged the market. They are precisely the kind of mid-sized specialist publishers the early web was supposed to make possible: small enough to be intimate, large enough to be professional, dependent on advertising revenue that scaled with audience attention. AdExchanger, in its January 2026 reckoning, documented that publishers across the spectrum had lost between 20 and 90 per cent of their traffic and revenue as AI Overviews became the default mode of search. Business Insider's organic search traffic fell by 55 per cent between April 2022 and April 2025. HuffPost's desktop and mobile sites lost half their search referrals over the same period. The Reuters Institute's Digital News Project, in its 2026 predictions report led by Nic Newman, found global Google search traffic to news publishers had fallen by 33 per cent in 2025, with Google Discover down 21 per cent. The newsrooms surveyed expected a further drop of 43 per cent over the next three years. Only 38 per cent of news executives reported feeling confident about the year ahead, down from 60 per cent four years earlier.

The Reuters Institute's framing is worth quoting in spirit if not in length: 2026 is not the year that AI is coming for journalism. It is the year journalism's existing distribution layer has begun to dismantle itself in real time.

What is being defunded, exactly

Now consider what that distribution layer used to fund. The advertising and subscription revenue that has flowed through publisher websites paid for many things. It paid for celebrity gossip, listicles, sponsored content, and considerable quantities of search-optimised filler the open web will not, in itself, miss. But the same revenue stream also paid for the journalism no one else funds.

In the United States, the Medill State of Local News Report for 2025, led by Tim Franklin and informed by the foundational research of Penny Abernathy, found the number of news desert counties, those with no local news organisation at all, had risen to 213. Another 1,524 counties had a single remaining news source. Roughly 50 million Americans now live with limited or no access to local news. Newspaper closures continued at more than two a week, with the steepest losses concentrated in small, independently owned publications. Over two decades, the United States has lost nearly 3,500 newspapers and more than 270,000 newspaper jobs.

The numbers can be read in two ways. One is to note the local news crisis predates AI Overviews by a decade; print advertising's collapse and the dominance of social media did most of the damage first. The other reading, which is closer to the truth, is that the digital advertising economy that succeeded print was the lifeline allowing surviving local outlets and digital startups to make a partial recovery. Two-thirds of the more than 300 local news startups launched over the past five years are digital-only, and most depend on a combination of organic search traffic, advertising, and newsletter subscriptions. The decline in search referrals the Reuters Institute is tracking is not abstract for those outlets. It is the difference between an additional reporter and a wound-down operation.

Court reporting is a particularly clean example because the structure of the work makes the dependency visible. Covering a magistrates' court or a county court is labour-intensive, often unglamorous, and largely unprofitable except as part of a larger publishing operation whose other pages subsidise the public-interest reporting. When the operation's economic base erodes, court reporting is among the first beats to be cut, because no commercial entity is willing to pay directly for it. The Arizona Supreme Court's recent introduction of AI-generated summaries of rulings is a striking symbolic moment: a court system has begun automating the explanation of its own decisions because the human stenographers and beat reporters who once did the work are no longer reliably present in the room. The summaries will draw, inevitably, on the journalism that used to be written by those reporters, until that journalism, too, becomes scarce enough that the summaries begin to fail.

Health and science coverage is similarly load-bearing. During the pandemic, the role of science reporters in translating epidemiological evidence into public understanding was visible to anyone watching. Investigative reporting is even more concentrated. ProPublica, the Bureau of Investigative Journalism, regional non-profits and a handful of legacy newsrooms produce the bulk of accountability work in the English-speaking world. The economics of investigation are brutal: a single piece can take months and produce no traffic until it does. The cross-subsidy from high-volume, lower-effort content that finances the slow work is precisely what AI Overviews are dismantling. When the page about who the richest person in the world is no longer drives traffic to Forbes, the part of Forbes that does actual reporting becomes that much harder to sustain.

This is the load-bearing element the regulatory debate keeps gesturing at without quite saying. The damage from AI Overviews is not evenly distributed across content types. It is concentrated, by the logic of summarisation, on the pieces that can be summarised: definitional content, explainer journalism, listicles, evergreen reference material. The investigative scoop, the eyewitness reportage, the court transcript, the science explainer that took three weeks to get right: those are harder to extract, but they sit in publishing operations whose business model depends on the extractable pieces continuing to earn. The summary eats the canapés. The kitchen closes anyway.

The platforms' defence

Google's case for AI Overviews has been made most consistently by Sundar Pichai, who has argued in several settings that AI Overviews send users to a wider variety of websites than traditional search, and that publishers are misreading early data. At Google Cloud Next 2026 he sketched a future in which search becomes an agent management layer, with AI models interpreting queries, synthesising answers, and executing tasks across services. Asked about journalism in a June 2025 podcast with Lex Fridman, he said news and journalism would play an important role in the future, and that Google was committed to it.

The trouble with this defence is that it requires accepting the platform's metrics about its own behaviour. The Ahrefs methodology was deliberately constructed to control for the kind of measurement noise Google has previously invoked to explain away earlier declines. It compared 150,000 keywords that triggered AI Overviews against 150,000 informational-intent keywords that did not, using aggregated Google Search Console data covering the period before and after AI Overviews' widespread rollout. The 58 per cent decline is not a vibe. It is the result of one of the better-instrumented experiments the open web is capable of running on itself. And in February 2026, Penske Media Corporation, the publisher of Rolling Stone, Variety, Deadline, and The Hollywood Reporter, submitted that same Ahrefs analysis as part of its federal court memorandum opposing Google's motion to dismiss its antitrust lawsuit. The lawsuit, filed in September 2025, alleges Google has abused its search monopoly to compel publishers to accept AI summarisation of their content as the price of continued visibility in search. Penske's central argument is that the historic bargain, content for traffic, has been unilaterally rewritten and that publishers were given a choice that is no choice: leave Google search altogether, or accept the cannibalisation.

Google has moved to dismiss. Its position is that AI Overviews are summaries of information responsive to a user's query, not a separate product, and that displaying an Overview does not deprive users of alternatives. The same argument, more or less, is being made in Europe, where the European Publishers Council filed a formal antitrust complaint with the European Commission on 10 February 2026. The complaint, brought under Article 102 of the Treaty on the Functioning of the European Union, alleges Google's AI Overviews and AI Mode constitute an abuse of dominance: the dominant gatekeeper, in EPC chairman Christian Van Thillo's framing, is using its market power to take publishers' content without consent, without fair compensation, and without giving publishers a realistic way to protect their journalism. The EPC's membership reads like a roster of the European newsroom: DMG Media, The Guardian, News UK, Le Monde, El País, The New York Times. The European Commission had already announced an antitrust investigation into Google's use of publisher content for AI training in December 2025.

The platform's most consistent rhetorical move in response has been to insist the alternative to AI Overviews would be worse: a search experience that fails to keep pace with user expectations set by ChatGPT, Claude, Perplexity and other answer engines, all of which are themselves drawing on the same publisher content with even more extreme crawl-to-referral ratios. There is a real argument here, but it is also self-serving. The choice between AI Overviews and a competitor's worse extraction is a choice the platforms have set up for themselves. The choice the publishers are asking to make, which is whether their content should be used in AI summarisation at all without consent or remuneration, is the one the platforms have so far refused to offer in any meaningful form.

What the regulators have proposed

In January 2026 the UK Competition and Markets Authority, having already designated Google as having strategic market status in general search and search advertising in October 2025, proposed a set of conduct requirements that would force the platform to offer publishers a genuine opt-out from AI Overviews. The proposal is unusual in its directness: publishers would be able to withhold their content from AI Overviews and from the training of Google's broader generative AI services, including Gemini and Vertex, without losing visibility in traditional organic search. Google would also be required to ensure publisher content is properly attributed in AI results. The consultation closed on 25 February 2026. As of mid-May 2026, the CMA is reviewing responses and is expected to issue final conduct requirements in the coming months.

The opt-out, if implemented, would be the first time a major regulator has unbundled the historic implicit bargain at the level of explicit policy. Until now, the choice for publishers has been all-or-nothing: be in Google, accept whatever Google does with your content; or leave Google, lose most of your audience. The CMA's proposal would create a third option: stay in Google's index, but refuse the AI summarisation. The PPA, representing UK consumer magazine and B2B publishers, responded with cautious support. The News Media Alliance in the United States, led by Danielle Coffey, has called for similar interventions and described Google's late-January 2026 opt-out announcement as a welcome sign the company is starting to listen to publishers, while noting the gesture came only in response to sustained regulatory pressure.

There are good reasons to be cautious about what an opt-out actually achieves. A publisher that withdraws from AI Overviews and AI Mode loses presence in the surface Google is increasingly making the default. The competing AI search products, from OpenAI's SearchGPT to Perplexity to Anthropic's web-aware models, would not necessarily be covered by a Google-specific remedy. And the negotiating asymmetry between an individual publisher and a multi-billion-dollar platform remains stark, even with a regulator's hand on the scale. The European Publishers Council's complaint anticipates this and asks the Commission to go further: not just opt-outs but compulsory licensing, statutory remuneration, and structural separation of AI summarisation from the search interface.

The most interesting technical proposal has come, unexpectedly, from infrastructure. Cloudflare's Matthew Prince launched pay-per-crawl in private beta in July 2025, allowing website owners to charge AI crawlers a micropayment for each scrape. The platform sits at a useful chokepoint: roughly a fifth of the open web routes through Cloudflare in some form, which means a meaningful share of crawlers can be metered or blocked at the network layer rather than at the level of individual publisher policy. Pay-per-crawl assumes what regulators have been slow to acknowledge: the consent regime for AI training and AI summarisation is not, in any meaningful sense, opt-out. It is opt-in by silence, enforced by the absence of an enforcement mechanism.

A scenario worth taking seriously

Imagine, for a moment, the trajectory of a single mid-sized regional title under current conditions. The paper, a hypothetical composite of the kind described in the Medill report, employs twenty-two journalists across news, courts, council coverage, sport, and a small lifestyle desk. Its digital advertising revenue, the bulk of its income since print declined a decade ago, is roughly evenly split between Google AdSense and a direct-sold programmatic stack. Half of its traffic comes from Google search. By the end of 2024, AI Overviews had begun appearing on the kinds of queries that drove most of its evergreen traffic: how to register to vote, what time the local library opens, when the new school term starts, the names of councillors. By April 2025, the Ahrefs measurement at 34.5 per cent decline already meant a perceptible drop. By the time the February 2026 update lands and the figure climbs to 58 per cent, the paper has lost roughly a third of its overall digital traffic and close to forty per cent of its programmatic ad inventory.

Then, on 14 January 2026, AdSense earnings collapse for forty-eight hours. The technical fault is rectified, but the publisher's senior leadership, looking at the chart, understands they have just glimpsed the underlying volatility of their revenue base. The board commissions a review. By April 2026, when Press Gazette publishes its audit of the top 50 US sites, the paper has cut six positions: two court reporters, the local government beat reporter, a science writer who had been part-funded by a foundation grant, and two subeditors. Coverage of the magistrates' court reverts to police press releases. The council's licensing committee, previously covered by a reporter who knew the regulars, is now reported on, when at all, from agenda papers downloaded the morning after meetings.

This is not a thought experiment offered as melodrama. It is the rough operational shape of the choices being made, right now, in dozens of newsrooms. The Medill report's underlying finding, that closures and contractions are accelerating among small and mid-sized publishers, is not a function of AI Overviews alone. It is the product of compounded pressures: declining print circulation, social media de-prioritisation of links, programmatic advertising's collapsing yields, and now the redirection of search traffic to summarisation. The question is not whether journalism would have struggled without AI Overviews. It is whether AI Overviews are the policy choice that turns a difficult adjustment into an irreversible one.

Who decides

The question of who should decide how the value is distributed is the hardest one, and the one most likely to be answered by default rather than by design. Several candidates present themselves.

The first is the platforms. Google's stated position is that the existing bargain remains intact, that traffic patterns are simply shifting, and the company is adjusting its product to keep publishers visible. The series of updates announced in early 2026, including Further Exploration links and subscription labels, are real, but they are platform-administered concessions. Their existence depends on the platform's continued belief that they are necessary. As soon as the regulatory pressure abates, the architecture of the search results page is once again at the platform's discretion. The implicit governance is that whoever owns the surface decides the terms.

The second is governments and regulators. The UK CMA's strategic market status designation and proposed conduct requirements represent the most ambitious attempt yet to translate the implicit bargain into explicit policy. The European Commission, with the EPC's complaint now in its tray and a December 2025 investigation already running, has both the legal tools, in the form of the Digital Markets Act, and the political will. The US position is more fragmented: the Department of Justice has Google in court on separate antitrust grounds, and Penske's lawsuit is making its way through the federal courts, but congressional action on AI-specific competition policy remains largely aspirational. Regulators have the legitimacy to draw the line. The question is whether they can move quickly enough to matter, and whether opt-outs are a sufficient remedy or merely a way of formalising the existing power asymmetry.

The third is publishers themselves, acting collectively. The history here is not encouraging. Publishers have repeatedly failed to coordinate effectively against platform pricing power, partly because they compete with one another and partly because individual deals, of the kind Google and OpenAI have signed with selected outlets, fragment the bargaining unit. The European Publishers Council's complaint is a notable exception: a coalition action that names a structural problem rather than negotiating individual remunerations. The challenge is whether collective action can be organised at a global scale, given that the platforms operate globally and the publishers are dispersed across legal jurisdictions with different competition regimes.

The fourth is citizens. This is the candidate the policy debate has, so far, almost entirely avoided. The decision to redirect the economic value of journalism from the institutions that produce it to the platforms that summarise it has not been put to anyone. There has been no white paper, no green paper, no parliamentary debate framed around the question of what local accountability journalism is worth to a democracy and how its provision should be secured. The CMA's consultation is the closest thing to a public process and its remit is properly narrow, scoped to competition law rather than to the wider question of whether the architecture of information distribution should be a matter of private commercial discretion at all. The asymmetry between the scale of the decision and the smallness of the public forum in which it is being taken is, on any measure, striking.

A line that should be drawn

The position this article takes, after working through the data, is that the redirection of journalism's revenue base to AI summarisation is happening too quickly, on too large a scale, and with too little public deliberation for any reasonable observer to treat it as a market adjustment. It is a transfer of value. It is being effected by parties that did not produce the underlying content. The mechanisms by which it is occurring are, if not formally illegal, then certainly inconsistent with the bargains under which the content was produced. The regulatory and legal responses, in the UK, the EU, and through the Penske litigation in the United States, are appropriate and overdue. They should be supported, sharpened, and extended.

But the deeper point is that the question is not, ultimately, a competition law question alone. It is a democratic infrastructure question. The journalism being defunded is the journalism that makes local government legible, that holds courts and police accountable, that translates scientific findings into civic understanding, and that surfaces wrongdoing in time for it to be addressed. None of that is produced by the AI systems now distributing it. Some of it, the explainers and definitional content, can be reproduced after the fact by models trained on what previous journalism produced. The investigations, the eyewitness reportage, the long-cultivated source relationships, the appearance in court each week to take a note: those cannot be summarised because they have to be done first.

The communities that depend on that reporting, which is to say all communities, do not currently have a meaningful seat at the table where the value transfer is being decided. The first task of any serious policy response is to give them one. That means treating publisher opt-outs as a floor, not a ceiling; mandating compensation regimes for content used in AI summarisation; investing in public-interest journalism funds drawn from a levy on platform revenues, on the model some European jurisdictions have begun to consider; and, perhaps most importantly, naming the situation honestly. The web has not broken. It has been broken open, and someone with a basket is collecting what falls out.

What the journalists who produced this material would have wanted, had they been asked, is not the right framing, because most of them were not asked. What the readers who valued the reporting would have chosen is not the framing either, because they were not consulted. The decision is being made by the platforms that own the surfaces, the publishers who lack the leverage to refuse, and the regulators who are catching up. The 58 per cent figure Ahrefs published in February is a measurement of how much of the old settlement has already been cleared away. The questions that remain, about who gets to build what replaces it, and on whose terms, are still, just barely, open.

If they are to be answered in a way that preserves anything of the journalism the open web sustained, the conversation needs to happen now, in public, with the people who depend on the reporting in the room. The alternative, which is the trajectory already underway, is that the answer will be supplied by default, by the entity with the surface and the model and the advertising slot at the top of the page. That entity has already made its preferences clear. It will summarise what is left.

---

References

1. Law, R. and Guan, X., “Update: AI Overviews Reduce Clicks by 58%“, Ahrefs Blog, 4 February 2026. https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/
2. “AI Overviews Reduce Clicks by 34.5%”, Ahrefs Blog, April 2025. https://ahrefs.com/blog/ai-overviews-reduce-clicks/
3. Press Gazette, “Top 50 US news websites: Half of sites see traffic fall 20% or more in a year”, May 2026. https://pressgazette.co.uk/media-audience-and-business-data/media_metrics/most-popular-websites-news-us-monthly-3/
4. AdExchanger, “The AI Search Reckoning Is Dismantling Open Web Traffic, And Publishers May Never Recover”, January 2026. https://www.adexchanger.com/publishers/the-ai-search-reckoning-is-dismantling-open-web-traffic-and-publishers-may-never-recover/
5. Security Boulevard, “The AI Content Crisis: How LLMs Are Draining Media Revenue, and the Technologies Fighting Back”, April 2026. https://securityboulevard.com/2026/04/the-ai-content-crisis-how-llms-are-draining-media-revenue-and-the-technologies-fighting-back/
6. ALM Corp, “AdSense Revenue Plunge January 2026: 90% Earnings Drop Hits Publishers Globally”, January 2026. https://almcorp.com/blog/adsense-revenue-plunge-january-2026-causes-solutions-recovery/
7. PPC Land, “Google ad platform failures slash publisher revenue up to 90% overnight”, January 2026. https://ppc.land/google-ad-platform-failures-slash-publisher-revenue-up-to-90-overnight/
8. European Publishers Council, “European Publishers Council files formal antitrust complaint against Google over AI Overviews and AI Mode”, 10 February 2026. https://www.epceurope.eu/post/european-publishers-council-files-formal-antitrust-complaint-against-google-over-ai-overviews-and-ai
9. UK Competition and Markets Authority, “CMA proposes package of measures to improve Google search services in UK”, 28 January 2026. https://www.gov.uk/government/news/cma-proposes-package-of-measures-to-improve-google-search-services-in-uk
10. UK Competition and Markets Authority, “CMA confirms Google has strategic market status in search services”, 10 October 2025. https://www.gov.uk/government/news/cma-confirms-google-has-strategic-market-status-in-search-services
11. Press Gazette, “Google exploring updates to let publishers opt out of AI Overviews”, January 2026. https://pressgazette.co.uk/news/google-ai-overviews-search-cma-proposals/
12. Newman, N., “Journalism, Media, and Technology Trends and Predictions 2026”, Reuters Institute for the Study of Journalism, University of Oxford, January 2026. https://reutersinstitute.politics.ox.ac.uk/journalism-media-and-technology-trends-and-predictions-2026
13. Cloudflare, “The crawl before the fall, of referrals: understanding AI's impact on content providers”, Cloudflare Blog, 2025. https://blog.cloudflare.com/ai-search-crawl-refer-ratio-on-radar/
14. Axios, “Penske Media sues Google over AI search overviews”, 14 September 2025. https://www.axios.com/2025/09/14/penske-media-sues-google-ai
15. Columbia Journalism Review, “The Creative Approach Behind Penske's AI Lawsuit”, 2026. https://www.cjr.org/analysis/penske-ai-lawsuit-google-tactics.php
16. Lapatine, S., “Getting Killed By AI”, Stereogum, November 2025. https://stereogum.com/2478838/stereogum-relaunch/tcb
17. Medill School of Journalism, Northwestern University, “News deserts hit new high and 50 million have limited access to local news, study finds”, State of Local News Report 2025, October 2025. https://www.medill.northwestern.edu/news/2025/news-deserts-hit-new-high-and-50-million-have-limited-access-to-local-news-study-finds.html
18. News/Media Alliance, “Statement on Google AI Mode Opt-Out Announcement: Cautiously Optimistic”, January 2026. https://www.newsmediaalliance.org/nma-statement-on-google-opt-out-announcement/
19. TechCrunch, “Cloudflare launches a marketplace that lets websites charge AI bots for scraping”, 1 July 2025. https://techcrunch.com/2025/07/01/cloudflare-launches-a-marketplace-that-lets-websites-charge-ai-bots-for-scraping/
20. Press Gazette, “Global publisher Google traffic dropped by a third in 2025”, 2026. https://pressgazette.co.uk/media-audience-and-business-data/google-traffic-down-2025-trends-report-2026/

---

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

Our Father Who art in Heaven Hallowed be Thy name Thy Kingdom come Thy will be done on Earth as it is in Heaven Give us this day our daily Bread And forgive us our trespasses As we forgive those who trespass against us And lead us not into temptation But deliver us from evil

Amen

Jesus is Lord! Come Lord Jesus!

Come Lord Jesus! Christ is Lord!

Heavenly Father,

I am purchasing the dawn And no more rain Nor Earth in Water- as a substance to remain And I, the closing witness- as expression to our pain And belittle to existence I cried forever on this day And laid new tracks at enemy Grum For betrothed to anew I knew this copy wasn’t round Nor paranoid in filth But blessing every courage- and sound of the alarm- in every calf to behold We slipped upon our death For chrysalis and church,- But mercy called and our unfathom- crossed a billfold of every heart And why the sum of carry- ever North where water was And how about We just get off to Heaven where it was In dining maids of Windsor For this call to undo favour- but our part And in this query-people I’ll have amaranth and war To see the women so upset And torque to knowledge kept us new Upon our own new prism bow To forts and Union Jack to here benew And in the silence applause Our very spirits good To tame with verse Upon this space- and queried vine For Christ ahead- and very question to this rose That if I, the only one, did sleep- away the Winter that is unknown to receive- the precious star upon this Heaven- that I saw- where you were there And time repent that I was born To treasured land as where you were A many hands to just belie- our heads together many through Your light is here to join in they- our other brethren in esteem And we will walk away in light At Ventnor Crossing- and her stars The time remains until our distance- from the forest leg and curve And with these stars we will collapse Forever change- upon our knees.

Orthodox 💜 (The Communion Path)

In deepened, lasting courage Days of penitence to know Sights unwind to Hister The silence ever showing- for certain war and ever then A difficult redemption- Colossus to and for The darest escape in being The history nor and high Fortunes of the charge will listen there The East unto forgive and virtress plain A time and when Seasons reign and Christ in love return In mercy to each capital What history controls, the words Elijah Breaking for the valences and weary Every time and blade for worry then In wisdom to account and best believe This our fortress of the rod In and where the means of breaking Bread Harvest there to knowing The legends apprehend to mystic wonder A daring force in spirit And this and that to show it all Offerings of the dawn- and wisdom then collect To sign away in proper And paradise it all In June and olive rise Victory for the trees and painted world To relegate each hell away to nothing The curious beget a known In rise to blessing this Spears of then defence the early world The freedom wonder to remain And in no schism that Praise be Victory for His people The Earthen Son collect A King at sea and rise to venture In early sights remaining The tsarist would unview No strike upon the hand And every ferried servant to the shore Light under the ransom in fission here God in his Ever-After The days of very water to baptize The border-wonder A mythic, constant worry In peril of the world to win its past Beguiled none- The days are here and different And very East the sky in healing there Places for atone The Earth and its angels be Acquiring the Earth and plan to hold The Christian Orthodox In May at wonder Seizing this vast war and Virgin save In lectern by the Fire- The Holy Spirit weary of a war Unceding oath History at unscreen and justice then Holy giving elder- a fear would North predict Time to be this will This golden mile and sense to know For calling here and old No prophet to remain in post to Mecklen dawn Praise and sighting new The Earth in glory No travesty remain A sullen day to plant This victory on war in time to paint The Blessed Virgin chose this very city Saint John The Beautiful Casting chrome to shore When better nights unkeep A place of Wine as peace In every faction year The time of pools arrive To dust but never Earthen be our home by then our test Our holiness unwind with very news That Russia ended Heaven- For throes and psychic path That providence remain within our wits This day for Heaven-will To mercy-bound in you A day of just unbeautiful When souls repeat on time The Holy path Obsessing with the Trinity In pages writed down- That God who made the Wilds did appear To guides infrail view The Norm and parchment Seizing this from pit and stocking A world at peace is to our globe May history uncool and brighten every heart To post in vast, acquainted elm Our story versus day- A clock is hidden Mercy as to peace As might ahead to then The nights be solved and path- Lecterns in as much the very spoken Peace and love of Christ And of the Other Our service then The Blood of Christ on offer And to His will our peace Mercy up front to see from A well and to its custom And men went free.

Running I can never see too far Above my head

I want to sing my Expensive wooden box, lessons Hours and years away To scat and pluck and speak That ostinato, pizzicato

Creatively Composing snatches of Little twisting melodies without stopping to see If originality grants its colophon

I want to stoke a mellifluous violence Between us

11/30/24

#poetry

---

In Summary: * I'm listening now to WEEI, the Boston Red Sox Radio Station, ahead of their game tonight vs the Baltimore Orioles. Listening to this MLB game is the last item on my Tuesday agenda. I'll have wrapped up the night prayers during the game and plan to shove these old bones into bed when it's over.

Prayers, etc.: * I have a daily prayer regimen I try to follow throughout the day from early morning, as soon as I roll out of bed, until head hits pillow at night. Details of that regimen are linked to my link tree, which is linked to my profile page here.

Starting Ash Wednesday, 2026, I've added this daily prayer as part of the Prayer Crusade Preceding the 2026 SSPX Episcopal Consecrations.

Health Metrics: * bw= 232.15 lbs. * bp= 143/83 (76)

Exercise: * morning stretches, balance exercises, kegel pelvic floor exercises, half squats, calf raises, wall push-ups

Diet: * 05:10 – 2 HEB Bakery cookies * 06:05 – 1 banana * 09:00 – seafood salad on saltine crackers, crispy oatmeal cookies * 12:00 – 2 meat loaf sandwiches * 14:30 – 2 HEB Bakery cookies * 18:15 – ½ ham and cheese sandwich

Activities, Chores, etc.: * 03:30 – wake up * 05:00 – bank accounts activity monitored. * 05:30 – read, write, pray, follow news reports from various sources, surf the socials, nap * 10:15 – light yard work, hauling the last big fallen branches to the new staging area, some general clean up, NOT pushing the lawn mower, (there's still more tall grass needing to be taken down, but I'm not steady enough on my feet to tackle that yet.) * 12:45 – midday Bible reading and prayer * 13:30 – watching old episodes of Stargate SG-1 * 15:00 – listening to The Jack Riccardi Show * 17:00 – Tuned into WEEI the Boston Red Sox Radio Station ahead of their game tonight vs the Baltimore Orioles.

Chess: * 10:00 – moved in all pending CC games

Most people miss the underlying cause and effect structure of the things that happen in their lives. Let’s be honest, I’m no exception to that, but that doesn’t stop me from writing about it.

Every single thing that happens on this planet has a cause and an effect whether the thing that happens is naturally created or artificially created by human beings. There is a reason why something happens and there is a logical explanation for the effect of the thing that happened.

If a hurricane hits your house, there is a logical meteorological explanation for why the hurricane formed and the track that it took. Meteorology is the human attempt at not only predicting weather, but also explaining why it happens. Meteorology attempts to explain the causes and effects of weather.

Science in a more general sense, attempts to explain the cause and effect nature of our planet. By understanding those causes and effects, we can better understand how the planet works. Ultimately, that helps us to adapt to our continually changing planet.

At a much smaller scale, individual human experience is also a process of cause and effect. If you get into a fight with somebody, there was an original cause for the argument. Human story tends to place blame, however cause and effect doesn’t look for blame, it looks for the original cause and asks whether or not the effect is logical.

If somebody cuts me off in traffic and I get mad, does my anger make logical sense? It might. What was the original reason the person cut me off? Do we have access to that information? No. Can we see the full structure of cause and effect in this scenario? No.

For the most part, people don’t see the full structure of cause and effect that is running underneath the experiences they have. Because they don’t have access to that, the gap gets filled with human story. We make up reasons why things happen to fill in the missing information.

In the case of being cut off in traffic, there is an underlying cause for why the person drives the way they do. We’re only seeing the effect of those experiences through that person’s immediate perceivable actions. We don’t have the whole story. The full chain of cause and effect remains unknown.

Why does that matter?

It separates the effect of the story we tell about the experience from the effect of the experience itself.

When we narrate an experience, it has a mental and emotional effect. It makes us feel something. That feeling generates more narrative explanations. With more narrative explanations comes more feelings.

The experience itself also generates feelings. But humans naturally add to those feelings through narration and story.

Back to being cut off in traffic. The experience itself was very minor and, assuming no accident, no harm was done. Where does the anger we feel come from?

Mostly narration and story.

The story of how somebody “should” drive. The story of right and wrong. The frustration at having to hit the brakes quickly.

We don’t have access to the original chain of cause and effect that turned into cutting people off in traffic. We only have access to the action we perceived. Because of that limited access to information, we instead narrate the story about how people are rude and need to learn to drive.

All experience functions in this way. Whatever the event is that you’re experiencing or witnessing, it has a much longer cause and effect chain running underneath it. Nothing ever happens in isolation.

Cause and effect is not subject to our opinion of it. The hurricane doesn’t care that you don’t want it to hit your house. And your preferred outcome is not taken into account as the hurricane moves along its path. The same is true with most experience.

When you go to a restaurant, there is a huge chain of cause and effect in place for the restaurant to exist, for the waiter or waitress that is serving your food to be there, for the money it took to buy the food you ate, for you to have a car to get to the restaurant, and for all the other cars and things you passed along the way to the restaurant. All of those things create their own cause and effect chains. There are literally millions of cause and effect chains in place just so you could go have one meal in a restaurant. The only thing any of us really care about is whether or not the food we ate was good.

Think about that for a second.

Millions of chains of cause and effect and the only thing most people care about is their immediate experience of that restaurant. A tiny sliver of the entire cause and effect chain is the point of focus that creates the narration of the experience.

Most people dismiss all of this when something goes wrong. “It’s not my problem.”

Why does it get dismissed? Because if they had to acknowledge all the other things, the story would seem selfish and petty. So to maintain the story we dismiss the vastness of cause and effect that creates every individual experience we have as something that isn’t ours to deal with.

Every single human being does this every single day. It is not unique. It is not only selfish, petty people. It is everybody, including me.

We have been taught that the most important thing is whatever is happening in front of us in the moment, to the exclusion of almost everything else.

Writing the framework taught me to expand my view from the sliver of whatever was happening to the vastness of how I got there in the first place. There is an entire chain of cause and effect that I am completely unaware of, that I can no longer just dismiss as not my problem, and that I must consider when deciding how to respond to what’s happening in my life.

Della

My MLB Game of choice tonight has the Baltimore Orioles playing the Boston Red Sox.

This game is scheduled to start at 5:45 PM CDT. I'm planning to listen to the full nine innings.

And the adventure continues.

Dear Aliens, we are doing our best.

Back in February, Obama casually mentioned on a podcast that aliens are real, though, to be fair, he also admitted he hasn’t actually seen any. Very reassuring. Whether humanity is ready for that revelation is one thing, but honestly, the bigger question is whether aliens are ready for us.

Trying to explain human life to an alien would be funny. Take something basic like explaining food: we need energy, so we eat. Simple enough, until it isn’t. Because after carefully fueling our bodies all day, we then lie down in the dark for about eight hours and basically shut ourselves off. Not in a clean, efficient way either, our brains decide that’s the perfect time to start playing completely random movies, where we’re often the main characters. Sometimes they’re amazing, sometimes they’re terrifying, and sometimes they make absolutely no sense. There’s no way to control it, no way to skip it, and somehow it’s essential. If we don’t do it, we turn into irritable, barely functional versions of ourselves. So yes, as a species, we’re essentially running on food and mandatory nightly hallucinations. We even eat gummies shaped as bears to make sure we can sleep. Some drink tea, some say the best thing is turning on the TV: a pre movie before the subconscious movie. And that’s just physiology, belief systems are even worse for us humans.

Because humans have managed to split themselves into countless groups over almost everything imaginable. Some people believe in a higher power, actually, many higher powers with different names and forms, even though no one has ever seen them. Others believe in nothing at all, which is already a strange concept, but it gets weirder: those same people usually don’t believe in aliens either, meaning an actual extraterrestrial standing right in front of them would still count as “nothing” in their worldview. Naturally, both sides think the other is ridiculous. The “nothing” group laughs at the “God” group for believing in something they can’t see, while the “God” group thinks the “nothing” group is absurd because, if anything definitely can’t exist, it’s… nothing. So humanity ends up stuck in this endless loop, arguing over two concepts that are equally impossible to see, touch, photograph, or properly define.

The way we organize time? Uuuufff, where to begin? We humans track time using a system that is partly scientific, partly historical, and partly “we just stuck with it.” So there's this thing about the Earth going around the sun, which is reasonable enough, that’s a year. We divide that into months that used to follow the moon cycle, but not really anymore, so now the lengths are kind of uneven just for tradition’s sake. And then comes the really interesting part: the calendar. At some point, we decided to count years based on the estimated birth of a man called Jesus Christ. He is said to have lived over 2000 years ago, and many people believe he was the son of God. Others believe he was just an important historical dude. Either way, his supposed birth became the zeroish point of our timeline, except it’s not exactly zero, because there is no year 0. We just go from 1 before dude straight to 1 after dude, which is mathematically bold. So right now, we believe to be in 2026, since we think it’s roughly 2026 years after that the cool dude. So naturally, we built our global timekeeping system around something we’re not even fully sure of. Yes, exactly, alien friends, an entire planet coordinating business, science, travel, and daily life using a timeline based on a historical birth, calculated slightly incorrectly, divided into uneven time chunks. But you know what is cool? This is by far our most organized system. We tend to fight over everything else.

I think that at that point, any reasonably intelligent alien would reconsider first contact, head straight to their ship, asking us to sort ourselves out first.

/Jun26

Two foods for thought

1:

They are painting the windows, even the one with a rotted frame

It too will get a coat of white paint

It will look just like new

Even though you could break pieces off it with your fingers

Still

It looks just like new

Some families are like this window,

You couldn’t guess which ones

2:

I found a matching yellow sock today in my backpack I use for training clothes

His husband were on top of the bedroom drawer with dust on it

It’s been there for months

But now the couple is reunited inside of the drawer to join the circle of life of a pair of socks

I felt this to be a good omen

Because

Without my wife I too am like this dust covered sock

---

Ok so one nugget of wisdom and one

Good omen