AI

The IMF spring meetings were supposed to be about trade policy. Instead, every closed-door session circled back to one word: Mythos. Anthropic’s latest AI model has found thousands of high-severity vulnerabilities across every major operating system and browser on the planet. Not theoretical weaknesses. Exploitable ones. And it chains them together autonomously. Finance ministers are terrified. The global banking system just became the softest target on earth. What Mythos Actually Does Mythos doesn’t just scan for bugs. It reasons about them. It discovers a buffer overflow in one system, a privilege escalation in another, and a logic flaw in a third — then stitches them into a working attack chain without human guidance. ...

A new drug takes 10 to 15 years and roughly $2.6 billion to go from lab bench to pharmacy shelf. Most candidates fail. The process is brutal, expensive, and maddeningly slow — especially if you’re one of the millions of people waiting for a treatment that doesn’t exist yet. On April 16, 2026, OpenAI decided it wanted to fix that. The company launched GPT-Rosalind, its first AI model built specifically for life sciences research — and in doing so fired a shot directly at Google DeepMind’s long-standing dominance in AI-powered biology. Named after Rosalind Franklin, the scientist whose X-ray crystallography work was essential to discovering DNA’s double helix, this isn’t just another ChatGPT update. It’s OpenAI’s declaration that the future of drug discovery runs through AI. ...

The European Commission just slapped Meta with a regulatory uppercut. On April 15, Brussels formally rejected Meta’s attempt to charge rival AI companies for access to WhatsApp’s 2 billion users — and threatened to force the tech giant to restore free access for competitors like Microsoft’s Copilot, OpenAI, and Perplexity. This isn’t just a legal skirmish. It’s the fight that decides whether you get to choose your AI assistant, or whether the company that owns your messaging app chooses for you. ...

The Treasury Secretary and the Fed Chair don’t pull bank CEOs into surprise meetings over product launches. They do it over threats. And last week, the threat was an AI model. Anthropic’s Claude Mythos Preview — announced April 7 — can autonomously discover and exploit zero-day vulnerabilities across every major operating system and web browser. Three days later, Treasury Secretary Scott Bessent and Fed Chair Jerome Powell had executives from Citi, Bank of America, Wells Fargo, Goldman Sachs, and Morgan Stanley in a room. ...

Stanford’s Institute for Human-Centered AI just dropped the ninth edition of its AI Index Report — 400+ pages of data on where AI actually stands. Not where the hype says it is. Not where the doomers think it’s headed. Where it measurably is. The short version: AI is more capable, more adopted, and more expensive than ever. It’s also less transparent, more environmentally destructive, and outrunning every guardrail we’ve built. Here are the numbers that matter. ...

Here’s the uncomfortable truth about enterprise AI in 2026: the technology works. It works really, really well. Just not for you. That’s the takeaway from PwC’s massive new AI Performance Study, which surveyed 1,217 senior executives across 25 sectors. The headline stat is brutal: 74% of all AI-generated economic value is being captured by just 20% of organizations. The other 80%? They’re splitting the scraps. The Canyon Nobody Talks About Winner-take-most dynamics aren’t new in tech. We saw it with cloud, with mobile, with the internet itself. But AI is compressing a decade’s worth of stratification into two or three years. ...

Every few months someone claims AI is going to “accelerate science.” It usually means a marginally better protein fold. This week is different. This week, an AI system took a quantum algorithm that its own human authors had nearly thrown in the trash, rewrote it, and handed back a result that made the people who secure most of the internet publicly move up their doomsday clock by six years. If you read one AI story this week, make it this one. ...

Choosing a college major has always been stressful. In 2026, it’s become existential gambling. A sweeping new survey from Gallup and the Lumina Foundation finds that 47 percent of currently enrolled college students have seriously considered changing their major because of AI’s impact on the job market. And 16 percent have already done it. One in six college students in America has changed the trajectory of their education — not because they discovered a new passion or failed organic chemistry, but because a technology that barely existed in its current form four years ago has made them question whether their degree will be worth anything at graduation. ...

Sometimes the biggest AI announcements aren’t announcements at all. They’re accidents. On March 26, a misconfigured content management system at Anthropic — the $60 billion company behind Claude — spilled nearly 3,000 unpublished assets into a publicly searchable data cache. Among the wreckage: a draft blog post describing Claude Mythos, which Anthropic has since confirmed is “by far the most powerful AI model we’ve ever developed.” This wasn’t a controlled product launch. It was a human error that gave the world an unfiltered look at what’s next in AI. And it’s equal parts thrilling and terrifying. ...

Cancer is terrifying for many reasons, but here’s the one that haunts oncologists: they often can’t tell which tumors will stay put and which will spread. By the time metastasis is detected — cancer cells colonizing distant organs — the window for effective intervention has often closed. A new AI tool from the University of Geneva is changing that equation. Called MangroveGS (Mangrove Gene Signatures), it predicts whether a cancer is likely to metastasize with nearly 80% accuracy — and it works across multiple cancer types. The research, published this week in Cell Reports, could reshape how doctors decide who needs aggressive treatment and who can be safely monitored. ...