David Wells

If you‘re saying „we just need to train it more“ - we can‘t. We‘ve reached the stage of enshittification. The web has been flooded with incestuous data. And that‘s after it crawled all of your work illegally.

What's happening: Elon Musk confirmed that X will delete all heuristics from its recommendation system within six weeks. In other words, the familiar logic of likes, replies, and reposts that shaped Twitter for years is about to disappear. In its place, Grok, the platform's in-house Al model, will read and watch more than one hundred million posts and videos every day to decide what users see. At the same time, X is testing a new interface that keeps users inside the app even when they open external links.

Welp

The dead internet theory is feeling pretty smug right now

The breaking down of anti-racism norms *in elite circles* in British politics is starkly at odds with the direction of public opinion on race.

* BANK OF ENGLAND PROBES DATA CENTER LENDING AMID AI BUBBLE FEAR

@bloomberg.com 🇬🇧

Meta Layoffs Included Employees Who Monitored Risks to User Privacy

"But buried amid the A.I. division layoffs was a different set of cuts. The company laid off more than 100 people in its risk review organization, according to three people familiar with the move and internal memos viewed by The Times."

Largest study of its kind shows AI assistants misrepresent news content 45% of the time – regardless of language or territory An intensive international study was coordinated by the European Broadcasting Union (EBU) and led by the BBC

Yet again, we can't afford to let LLMs become a source of epistemic grounding for society.

Big tech is paying for Trump's White House ballroom | TechCrunch Trump's $250 million White House ballroom will be funded in part by big tech companies like Amazon, Apple, Google, Meta, and Microsoft.

Trump's $250 million White House ballroom will be funded in part by big tech companies like Amazon, Apple, Google, Meta, and Microsoft.

The AI Shift: where are all the job losses? A macroeconomic change isn’t clear yet but payroll data shows some types of work are already being displaced

'It’s worth raising an eyebrow when companies announcing lay-offs link them in vague terms to becoming “AI ready”, but without any corroborating detail. Let’s face it, it sounds a lot more dynamic than just saying the business isn’t doing very well.' 👏 @sarahoconnorft.ft.com

I personally am significantly more worried about the effects of LLMs and AI video slop on children’s brain development than I am about even social media. I am very worried that our school systems and laws are pushing kids toward AI artificial “friends” while banning social apps.

Katie Lam’s immigration panic is built on lies As a Tory proposes the harshest migration policy since Idi Amin’s, data shows the public doesn't understand the numbers - thanks to politicians too venal or timid to tell the truth

“Labour is losing more voters to the Greens and the Lib Dems than it is losing to Reform, not least because many are appalled at Labour’s cowardice on immigration—at seeing the government essentially concede the argument to Reform and the newly extremist Conservatives without even trying to fight.”

Graphic depicts select capital flows between OpenAI, Microsoft, Nvidia, CoreWeave, AMD, and Oracle.

Another depiction if the circularity of capital flows between OpenAI, Microsoft, Nvidia, CoreWeave, AMD, and Oracle.

"Almost half of everything AI assistants tell you is wrong" and "Oxford university gives AI assistants to all its students" are two headlines we are forced to accept as coexisting in this appalling timeline.

Some interesting (if for me unsurprising) examples in the full report.

"Largest study of its kind shows AI assistants misrepresent news content 45% of the time"

Link: www.bbc.co.uk/mediacentre/...

Report: www.bbc.co.uk/aboutthebbc/...

We're going to use military force to ethnically cleanse a nation of at least 10% of it's population: sensible politics of the everyman

We're going to tax rich people more so everyone except rich people aren't totally fucked the whole time: fantasy politics of the metropolitan elite

Bubble-talk is breaking out everywhere But more optimistic investors continue to bank on the cavalry arriving if things get really dicey

“All of a sudden, the warnings are coming from all sides.” - @katie0martin.ft.com here the market
Bubble-talk breaking out everywhere www.ft.com/content/ce8b...

"Viewpoint diversity" about whether human activity is warming the planet, or whether vaccines work, is like "viewpoint diversity" about whether the sky is blue or whether the Earth is round. It falls outside the range of good faith scientific and academic discourse.
#ScienceUnderSiege

Here Lam explicitly sets out her proposal - which is official Conservative Party policy - to deport long-standing legal permanent residents who have *ever* claimed any benefit, including the state pension or child benefit (even if the child is British), or who earn less than £39K.

This is deranged and dangerous from the Conservatives, and similarly from Reform. If they could enact it, it would rip British society apart, destroy the economy & NHS, & make the UK an international pariah. Every decent politician & political journalist/commentator needs to push back hard.

Also that if you’re an immigrant who stands to earn well above the average wage in the UK, you probably have a fair few options that are much better than moving to the UK. Delusional

‘We asked Israel to investigate the deaths of our journalists. We’re still waiting’ Reuters editor-in-chief Alessandra Galloni on the targeting of journalists in Gaza

More journalists have been killed in Gaza than in any other conflict in history. The EIC at Reuters, Alessandra Galloni, joins us to reflect on the last two years of covering the war on Gaza. www.prospectmagazine.co.uk/podcasts/med...

This hits a lot of the notes I've been shouting from the rooftops: the frictionlessness of both producing and consuming slop, and the value of labor as friction that lets us slow down and think. Rather than "unlocking creativity" it's rendering our collective creativity down into a big vat of slop.

Anthropic is spending more on its Amazon Web Services contract than *all of the revenue it's generating*, according to Ed's source who's familiar with the deals. Pretty wild snapshot of the #2 AI company here.

Honestly what is the point of a Labour government that won’t take a stand against mass revocation deportation of people with permanent settlement rights?

If you won’t stand against anything you stand for nothing.

Tory MP criticised after demanding legally settled families be deported Katie Lam said move would make UK ‘culturally coherent’ and that a large number of people ‘need to go home’

A Conservative MP tipped as a future party leader has been condemned for saying large numbers of legally settled families must be deported, in order to ensure the UK is mostly “culturally coherent”.

www.theguardian.com/politics/202...

It’s just as clear that this approach has long stopped serving the company well. From the Scarlett Johansson voice mode controversy to GPT-4o sycophancy to the more recent backlash against GPT-5, OpenAI’s rushed release schedule now routinely puts the company on the defensive. But with Altman now openly acknowledging the connection between revenue pressures and product releases, there is every reason to believe that policy concerns will continue to be eclipsed by the urgency of shipping.

OpenAI's flailing over content policy for Sora is just the latest example of the company rushing out a product at the expense of its own reputation. As AI tools get more powerful, that's reason for concern www.platformer.news/openai-sora-...

LLMs get stupider the more junk web content they consume. "As LLMs scale and ingest ever-larger corpora of web data, careful curation and quality control will be essential to prevent cumulative harms": www.arxiv.org/pdf/2510.13928

This remains the funniest way to hear about an internet outage, though.

AI-generated ‘poverty porn’ fake images being used by aid agencies Exclusive: Pictures depicting the most vulnerable and poorest people are being used in social media campaigns in the sector, driven by concerns over consent and cost AI-generated images of extreme poverty, children and sexual violence survivors are flooding stock photo sites and increasingly being used by leading health NGOs, according to global health professionals who have voiced concern over a new era of “poverty porn”. “All over the place, people are using it,” said Noah Arnold, who works at Fairpicture, a Swiss-based organisation focused on promoting ethical imagery in global development. “Some are actively using AI imagery, and others, we know that they’re experimenting at least.” Continue reading...

AI-generated ‘poverty porn’ fake images being used by aid agencies

The fact that something like this is being published - *published* - on a major newspaper tells you everything you need to know about how certain media are actively contributing to the collapse of societal norms