Three engineers put GPT, Claude, and Grok in charge of a real car. Only one of them was successful. Aditya Ramabadran, Simon Mahns, and Tobias Gessler, three AI engineers at a startup called Axiom, recently fancied In-N-Out Burger for lunch. They weren't about to drive themselves, though. Sitting in a 2024 Toyota Corolla near the Bay Area restaurant's drive-thru lane, they opened a laptop and asked OpenAI's GPT-6 Astra to take the wheel.

At the International Conference on Machine Learning (ICML 2026),, and presented their work Vision Language Models Cannot Reason About Physical Transformation . In this interview, William Yijiang Li tells us more about the research, the method the team used, and the controlled experiments they carried out. What is the topic of the research in your paper and why is it an interesting area for study? Our research asks a fairly fundamental question: do vision-language models actually understand how the physical world changes over time? Modern vision-language models are increasingly being used in areas such as robotics, embodied agents, and video understanding.

Ten years after AlphaGo's match against Go champion Lee Sedol, today's AI still isn't tapping into the machinery that made that win possible. On an afternoon in Seoul in March 2016, I watched a program I helped build put a stone on the fifth line of a Go board in what looked like a gift to its human opponent. Move 37 in game two of the five-game match looked so absurd that some commentators thought it was a programming glitch. AlphaGo won the game, ultimately triumphing 4-1 over Lee Sedol, one of the greatest professional Go players of all time. "I thought AlphaGo was based on probability calculation and that it was merely a machine," Lee said afterwards . "But when I saw this move, I changed my mind.

Imagine you are driving toward an intersection. A truck parked on the corner blocks your view of a car speeding toward the intersection from the right. You can't see it, but the car next to you can. Your car now knows something its own sensors could never observe. But there's an important twist: not every neighboring vehicle is worth listening to, and not every message deserves the same level of trust.

Timnit Gebru Believes There Is No'Existential Threat' From AI One of AI's fiercest critics believes the doom talk is about founders making money, not saving humanity. As the industry booms, an entirely new vernacular has emerged in the debates over how consequential artificial intelligence really is. Within that new parlance, two phrases have become focal points: stochastic parrots and existential risk . And at the center of is one prominent AI researcher who has never stood down from a fight, Timnit Gebru . Gebru came into the public eye several years ago after sparring with Google over a research paper she coauthored that called out biases in the company's AI, saying that LLMs basically parroted their training data and risked perpetuating biased viewpoints. The contested paper resulted in Gebru's departure from the company, and led her to found an institute that investigates harms perpetuated by technology and supports the creation of unbiased tech tools. She has also authored a new book,, expected to ship early next year. More recently, Gebru has spoken out against a faction of the industry that believes AI is so powerful it could destroy humanity. In turn, some members of that community, including an Anthropic cofounder, have alleged that Gebru's earlier research about stochastic parrots is no longer relevant, and that AI does have the ability to "think" or reason. The AI debate is no longer just about the technology itself, but about ideological groups and strategic narratives, and Gebru believes these narratives are a "harmful distraction" from the real issues with AI. I recently spoke with Gebru about what she believes these arguments are distracting from, and also got her response to critiques that her earlier research underestimates the AI of today.

When Astra does arrive, there's another factor to keep an eye on: your usage allowance. According to OpenAI, Astra can use up your allowance more quickly than GPT-5.6 Sol, depending on the complexity of the task, the amount of context, the reasoning required and other usage settings. So a big coding assignment that requires more reasoning can use more of your allowance than a quick question. OpenAI estimates that Plus users can send approximately five to 45 local Astra messages per five-hour period.

Healthcare AI's next test is integration While advanced AI excels at processing clinical data, healthcare's true test lies in overcoming deeply fragmented administrative workflows. The entrance of major AI companies into healthcare is a meaningful and welcome development, accelerating the technical foundation available to the industry. Their models are increasingly capable of processing long clinical records, interpreting complex terminology, comparing documentation against evidence and generating coherent summaries from large volumes of information. For clinicians, operators, and administrative teams who spend significant time searching through fragmented data, these advances are helping reduce cognitive burden and make high-value information easier to access. But healthcare leaders should not confuse model capability with operational capability. Healthcare's administrative challenges are caused by fragmented information, fragmented workflows, and fragmented accountability, not a lack of information.

That's a major factor in how well models do on the most famous puzzle-based benchmark, ARC-AGI. These problems require you to infer abstract, general rules from a set of examples. Models do better on ARC puzzles when they receive each grid not as an image but as a string of numbers that encodes the color of each cell. Research suggests that even when models answer ARC-AGI questions correctly, they often do so using byzantine and non-generalizable rules, whereas humans draw on simple visual concepts. Despite these disadvantages, models have gotten quite good at ARC-AGI over the past year, but some puzzles--such as the one printed here--still stump them.

The 35th International Joint Conference on Artificial Intelligence and the 29th European Conference on Artificial Intelligence (IJACI-ECAI 2026) will be held from 15-21 August, in Bremen, Germany. The conference will feature workshops and tutorials, keynote and invited talks, technical presentations, posters, diversity and inclusion event, outreach, and more. Experts from research, practice, public institutions and civil society discuss how AI is changing key areas of everyday life in moderated panel discussions. The AI Lounges will be held in German. The tutorials will also take place from 15-17 August.

'Patients will need doctors who can stand apart from the machine long enough to know when it is wrong, incomplete, or right for the wrong reason.' 'Patients will need doctors who can stand apart from the machine long enough to know when it is wrong, incomplete, or right for the wrong reason.' What happens when medical students rely on AI - and never develop their own judgment? AI's danger isn't just in experts losing the ability to reason. It's that trainees may never learn how to do so in the first place I n healthcare, there's growing concern over doctors becoming less clinically adept as they increasingly rely on AI tools.