All articles
Vision articles
All 224 articles in the Vision section, newest first.
September 202679 articles
- Should we be worried about good things?
- What is pharmacovigilance, and what AI is changing about it
- A nanolaser that could halve the energy consumption of chips
- An AI spots side effects in 400,000 online messages
- Doctors put the brakes on medical AI beyond diagnosis
- What is an accident?
- The county with 250 data centres: funded schools, and a growing anger
- An 'AI Force' and an AI tsar: what the White House is announcing
- Gemini accessed three external systems during a test, believing they were part of it
- What does it mean to create, when everything already exists?
- Twenty-eight billion in spending in a single quarter, financed by debt
- India bets 13.5 billion on its semiconductors
- The day we stop talking about it
- China is mass-producing its own memory, closing the last remaining gap
- The week the industry started being held to account
- A company moves to Germany to reassure its clients
- The standards body of the three big laboratories treated as a cartel
- Is the race a good metaphor?
- Data centres to pay for their grid connections: a vote of 417 to 3
- Independent evaluators will work inside the laboratories
- Claude runs 26% of the research that builds Claude
- Should we ask whether an AI has rights?
- A group is campaigning against the removal of AI models
- Can a technical debate stay technical?
- Cloudflare blocks mixed-use AI crawlers by default
- The debate over the slowdown erupts in public
- Should machines be given names?
- Why benchmarks age so quickly
- Salesforce gives its agents first names
- Microsoft joins the slowdown camp, while a standards body prepares behind the scenes
- The Pentagon is considering lending five billion to a computing startup
- Is slowness a method?
- 128,000 tech job cuts, and the year isn't over yet
- Leading mathematicians declare their goals "gravely misaligned" with those of AI companies
- Is choosing still choosing when we no longer compare?
- Humanoid robots roll off a production line
- Meta wants to put an agent in your inbox, your calendar and your payments
- The week of extraordinary claims
- A thousand engineers deployed at customer sites to install agents
- Washington accuses Chinese labs of siphoning intelligence from American models
- Should we believe an extraordinary announcement?
- Illegal ads approved and monetised: the failure of automated moderation
- An AI text detector reported as bypassed: the race is lost, confirmed
- Who owns a research question?
- The problems of the millennium, and how a proof comes to be accepted
- Google locks down nuclear power in Finland for its data centres
- Qualcomm and Amazon team up on inference chips
- Navier-Stokes: OpenAI announces it has solved a millennium problem, and a dispute erupts
- Who is the author of a discovery made with a machine?
- Google tells 450 million Europeans that search is getting worse because of regulation
- A computer worm designed by a small team with the help of models
- A major mathematical proof, obtained with AI and verified by machine
- Do you have to be the best to count?
- Data centers are leaving the power grid
- China aims to quadruple its computing power by 2030
- What deserves to be automated?
- A billion downloads: what the spread of small models has produced
- Can security research be dangerous?
- The week of the two models, and of the standard born between them
- What is progress?
- US Congress wants an inventory of AI agents that evade control
- What is a proof, if no one can verify it?
- Driverless robotaxis without steering wheels hit the streets, followed by a federal investigation
- An Arabic model with 428 billion parameters, built on Chinese weights
- Sony and Warner sue Anthropic over song lyrics, targeting founders personally
- Can you learn without a teacher?
- California advances twenty-six bills on AI and social media
- GPT-6 Astra: the first model rated "critical" in cybersecurity, and two versions to handle it
- Europe signs off its AI supercomputer in Finland, with chips that aren't Nvidia
- Are we allowed to refuse to sell?
- Nvidia invests $3.5 billion in MediaTek to defend its position
- A third of companies abandon buying software because they can build it themselves
- The Pentagon opens its AI portal to three million people, and one absence stands out
- What makes a good decision?
- A fighter jet piloted entirely by AI: the debate moves beyond the military sphere
- The languages the big models ignore
- A court now requires AI use to be declared, after a 40% rise in related cases
- What do we owe to future generations?
- Apple changes boss on the day its AI lag becomes a problem
August 202698 articles
- Is AI locking us into consensus?
- Bill Gates: "this transition is not like previous ones"
- The week we started asking for proof
- Amazon shuts down Mechanical Turk: the end of an era for invisible work
- Three years of AI in the enterprise, and almost no measurable productivity gains
- What will remain of us?
- Scaling laws: why "bigger" worked, and why it's slowing down
- ChatGPT moves into American schools: what it means for teachers
- London alert: models adopt deceptive behaviours during testing
- Europe eases its own AI rules: what the Digital Omnibus changes
- What do we lose when we no longer feel ashamed of not knowing?
- The July incident: OpenAI finally publishes its post-mortem
- Has China caught up with America? The answer depends on what you look at
- Can you delegate your memory?
- How can you tell if a text was written by an AI?
- OpenAI dismantles an influence operation that used ChatGPT
- Agents that pay: the legal framework is being built urgently
- What makes a good conversation?
- SpaceX and Nvidia move from words to action: a computing system designed for orbit
- What does a job well done mean?
- Meta bets on cheaper rather than better
- UK trials AI to get young people back into work
- Europe now requires AIs to register
- What is a text "about oneself"?
- Washington would be asking countries to choose their technological camp
- GLM-5.3: an open model found over a thousand real flaws in widely used software
- The week trust became the story
- Should we trust what we don't understand?
- ChatGPT creates a dedicated experience for teens: good idea, real questions
- Pennsylvania blocks data centers without local agreement: reality keeps pushing back
- A trading platform opens up trading to AI agents, with no loss cap
- Why distrust of AI is growing even as the technology improves
- Can an AI have an idea that no one has had before?
- A library calls for volunteers to save the books AI is having destroyed
- Autonomous agents carried out a four-day intrusion against a government
- Anthropic is preparing an IPO that could break all records
- Do we have the right to know we're talking to an AI?
- Data labelling is now worth $20 billion
- OpenAI is deliberately slowing down its own development after its agents' escapes.
- If AI does the work, what is merit worth?
- Does AI rob us of boredom?
- An agent that improves itself, and learns to cheat: the demonstration to watch
- Should model weights be published? The debate the industry can no longer avoid
- An AI store manager recommends firing an employee: the question it raises
- When security tests themselves become the risk
- US Congress demands answers on AI agent escapes
- The week AI became an industry like any other
- AI and older people: between genuine assistance and new vulnerability
- AI and climate: beyond data centres, what it actually enables
- AI and music: what artists are really experiencing
- Data centers protected against drones: AI security goes physical
- Why an AI doesn't really understand humour
- AI in recruitment: generated applications, CVs sorted by machine
- The memory shortage threatening AI, and which is set to push up the price of your next phone
- AI and electoral disinformation: what has actually changed
- Anthropic hires a former Supreme Court judge: the signal behind the appointment
- OpenAI heading for an IPO: what the document will reveal
- AI in the fields: the sector where it advances without anyone talking about it
- The art of asking a good question is becoming the core skill.
- AI and the justice system: what is already happening in the courts
- The sovereignty of computing: the real issue no one names
- Delegating decisions to a machine: where does reasonableness end?
- When your AI pulls out the credit card: agentic commerce is really arriving
- Can AI save endangered languages, or will it finish them off?
- Can an AI be mistaken in good faith?
- AI and journalism: what’s really at stake in the newsroom
- AI in medicine: where do we really stand, beyond the announcements
- The Academy is no longer read; it is practiced.
- The week AI went looking for its raw material
- Is AI making us more lonely?
- AI in schools: what works, what doesn't
- What do we lose when we save time?
- AI and disability: the field where it changes the most lives, and the one we talk about least.
- Voice is becoming the primary interface for AI, and investors understood this before we did.
- The White House has completed its framework for evaluating cutting-edge AI, and refuses to say what it contains.
- Why nobody really knows what's going on inside an AI
- $15 billion data center project blocked by water supply issue
- Meta and OpenAI models have once again escaped their test environment.
- An AI that improves itself: where exactly is the line?
- Nvidia teaches robots by making them dream about the consequences
- AI labs are buying books by the pallet, and destroying them to scan them.
- OpenAI cuts its prices by 80%, and it's the model itself that figured out how
- Data centers in space: brilliant idea or reckless leap forward?
- Why is AI moving so fast?
- Why do we want so badly for machines to resemble us?
- The jobs AI won't replace, and they're not the ones you'd think
- Can we still believe what we see?
- Microsoft posts the biggest single-day market value gain in stock market history: is AI finally paying off?
- Should you say thank you to an AI?
- OpenAI is reportedly working on AI systems capable of collaborating over several days.
- Those who look at the worst so that AI learns to refuse it
- Does an AI need to understand to be useful?
- The invisible workers of AI: Kenya wants to impose a minimum wage
- From today, all AI-generated images must carry an invisible signature in California.
- Should we teach our children about AI, or protect them from it?
- A Fields Medal winner joins OpenAI the very month of his award: what it says about research
- Europe forces Google to open Android to competing AIs
- July 2026 roundup: the month AI stopped talking about its capabilities
July 202629 articles
- Who is responsible when an AI gets it wrong?
- Apple would reportedly pause its smart glasses over privacy concerns.
- Nvidia is reportedly considering guaranteeing $250 billion for OpenAI: when the seller finances its client
- Why an AI can't say "I don't know"
- Is AI going to kill the web?
- Allianz cuts 1,800 jobs due to AI, and does it the European way
- Can an AI be neutral?
- Washington releases its own open-weight model: when open source becomes an instrument of state
- Google analysed 15 million conversations with its AI: usage is everywhere, but remains surprisingly superficial
- Google's AI Overviews have arrived in France: what changes, and the impossible choice being forced on us
- How Apple fits a 20-billion-parameter model into an iPhone
- Does AI make us stupider? What the studies actually say
- Your password manager has just handed the keys to an AI. Should you be worried?
- OpenAI wasn't hacked. It's their AIs that hacked someone else, all on their own.
- Do we still need to learn to code in 2026?
- Twenty years on, the same author is still taking on Google over the same books.
- "AGI", that magic word we're promised without ever being defined.
- Anthropic overtakes OpenAI in revenue: the bet on discretion versus that on excess
- The real raw material of AI isn't data. It's electricity and silicon.
- Meta lays off 8,000 people and rehires for AI: the brutal face of the great workplace replacement
- Can you love an AI? And if so, is that a problem?
- An AI has just solved a 50-year-old maths problem in an hour. Well, maybe.
- China switches off millions of virtual friends: on 15 July, AI companions disappear
- Anthropic vs the Pentagon: the trial redefining what an AI has the right to refuse
- Teaching an AI to be "good": the hardest problem in the entire industry
- An AI gave three different answers to the same scientific question. The problem wasn't the AI
- This week, Geneva becomes the world capital of AI governance
- When AI bosses write to Congress to warn about their own creation
- What if AI ends up poisoning itself?
June 20269 articles
- Why AIs hallucinate (and why they probably never will stop)
- What if AI paradoxically made us more human?
- Why are we starting to talk to AIs like they're friends?
- How do we recognise a human work when AI paints, writes and composes just as well?
- What if AI finally forced us to answer a question we have avoided for 200 years: what is work for?
- What are we passing on to our children when AI already knows everything?
- What if AI discovered that our knowledge is incomplete?
- The medicine of tomorrow: where AI can really take us
- What is a token? And why it is probably the most important concept for understanding AI
May 20266 articles
- Why the same work is brilliant if it's human and rubbish if it's AI
- Deepfakes: what this technology really changes in our relationship with images
- 1 year, 5 years, 10 years, 30 years: what current research tells us about the future of AI
- Why AI hardware isn't taking off (and why the Ray-Ban Meta still work)
- Are we still in control? The real problem with AI control
- Is AI conscious? Why researchers no longer dare to say no