TL;DR: On Tuesday, Anthropic released Claude Fable 5, a version of its vulnerability-hunting Mythos model it says is safe for public use. According to the company, Fable 5 is state-of-the-art on nearly every benchmark it tested. But it’s still a leashed version of Mythos: Ask anything in areas like cybersecurity and biology, and you’ll be rerouted to the older Claude Opus 4.8 model.

What happened: The model is now free on Pro, Max, Team, and seat-based Enterprise plans through June 22, after which it will require usage credits. Fable 5 is the first Mythos-class model Anthropic has made generally available, with the highest scores across some coding and finance benchmarks, according to the company.

Initial tests: Early reviews indicate that Fable 5 is a monster that’s exhausting to feed. After spending five-plus hours testing it and running up a $110 token bill in a single day (that was fortunately covered as part of his subscription), developer Simon Willison called it “something of a beast. It’s slow, expensive and has been quite happily churning through everything I’ve thrown at it so far.”

Wharton professor Ethan Mollick, who had early access, was impressed to see it spin up its own swarm of agents to research and check one another’s work, but griped that the guardrails sending his queries to Opus 4.8 tripped “way too often.” (Anthropic conceded the safeguards are tuned conservatively and will “sometimes catch harmless requests.”)

Timely flex: All eyes are on how Anthropic justifies its $965 billion valuation ahead of its planned IPO. Fable 5 isn't cheap—it costs $10 per million input tokens and $50 per million output tokens (double that of Opus 4.8), but that price tag could make it a tidy new earner for Anthropic.

“For us, it’s really around what we call ‘race to the top,’ being able to provide this technology in a valuable fashion,” Dianne Penn, Anthropic’s head of product management for research, told CNBC. “And at the same time providing the right safety guardrails so that it can do asymmetrically more benefits than harm,”

Mythos 5, which has the same underlying model as Fable 5 with some safeguards lifted, launches first for a small group of cyberdefenders and infrastructure providers through Project Glasswing, Anthropic said, before a wider rollout.

Bottom line: Anthropic spent two months calling Mythos too dangerous to ship. Now, a guardrailed version is here, and the company is betting its most-feared model can go mainstream without blowing up—and make it a lot of money in the process. —LC

• For the first time, a vaccine whose core component was designed entirely by AI has passed a human clinical trial. A University of Cambridge team fed genetic data from a range of coronaviruses into an AI, which engineered a "super-antigen" targeting the parts of the virus family least likely to mutate. The vaccine covers existing Covid variants as well as animal strains with pandemic potential. In a trial of 39 people, the shot generated immune responses across multiple coronavirus types; researchers described the immune effect as modest but promising. A larger follow-on study involving around 200 people is planned, and the team is applying the same method to the flu and Ebola.
• Human testing has now begun on a gene therapy designed to reverse cellular aging. A longevity startup co-founded by a Harvard professor announced this week that it dosed its first patient with ER-100—a modified virus that instructs aging retinal cells to reset their epigenetic clocks. The initial target is glaucoma patients whose optic-nerve cells are slowly dying. While the near-term goal is restored vision, the company's longer-term bet is that the same partial reprogramming approach could eventually apply to tissues throughout the body. For now, the trial is primarily a safety test; whether the cells actually respond comes next. —SM

A Redditor who says they spent years in government surveillance—and claims to make upwards of £300,000 a year (over $400,000)—has apparently been hiding in the bushes outside Grand Theft Auto VI developer Rockstar's office in Edinburgh, Scotland. They say they’re monitoring oxygen levels, acoustic output, and every car in the lot to predict when the game's third trailer will drop (apparently, it’s “imminent”).

Their theory is that elevated O2 concentrations mean executives are working weekends, decibel spikes signal heightened activity, and a Rolls-Royce Cullinan in the lot means notable visitors are on-site. This is either an elaborate troll (likely, given they have posted a total of five times ever, and all the posts are about this) or the most inspired misuse of a government surveillance background in recent history. The Redditor says they're doing it because they were bored. They also claim their gear has been confiscated—presumably by Rockstar's security—but that they remain in the field.

Others in the GTA community have called the behavior unnecessary and dangerously obsessive, though a few Redditors helpfully suggested laser listening devices. Rockstar has not yet commented. The whole thing has drawn comparisons to the Pentagon pizza theory—the idea that late-night pizza delivery spikes indicate when the government is about to make a big move. We got government-level surveillance of a video game developer before GTA VI. —SM

Chaos Brewing Meter: /5