this is just wrong: these vulns were found by defenders using Daybreak rather than attackers. It's just evidence Glasswing didn't go far enough, possibly due to capacity constraints. It's also very likely that March Mythos was not as good as Sol/Astra of September

Incredible things are happening in San Fransisco

The most insightful thing I read about technology eleven years ago was @benthompson’s Aggregation Theory. Amazon blocking Muse is another perfect example. Customers want one agent that knows them and can get things done. The platforms being aggregated want the opposite. They want to own the customer relationship, not become interchangeable suppliers behind someone else’s interface + protect their ad business. Imagine groceries: Your agent can compare Amazon, Instacart, DoorDash, and Uber Eats, then route every order to the best option. Amazon can block that access because it is a gigantic company. But smaller players have every incentive to offer agents a clean API and a seamless experience. If Instacart embraces agents while Amazon blocks them, the agent will increasingly route demand to Instacart. Customers will be happy, and Amazon will eventually face the dilemma: keep protecting the relationship, or match the better value proposition. This will be the defining tension of the agent economy: aggregators realizing they are now being aggregated. Amazon will push its own assistant, of course. But a vertical Amazon agent will struggle against a horizontal agent that already knows my email, calendar, preferences, memories, and entire life context. Many companies will want to become the aggregation layer. Very few actually can.

In 12 weeks, we built a research facility that is run entirely by AI. AI designs, executes, and observes experiments end-to-end across biology, chemistry, and materials science. We’re introducing SciUniverse: a benchmark that measures AI’s ability to do real-world scientific research.

I don’t post much, but this felt worth unfolding :) For the past year, I’ve helped shape the core software experience for iPhone Duo, especially multitasking. A lot of it came down to tiny decisions, the kind that, when they’re right, just disappear into how the product feels to use. Very grateful for the team and everything I learned along the way. So happy to finally see it out in the world!

BREAKING: Anthropic threat intelligence team published a report on everyone misusing Claude "We believe we have a responsibility to disclose misuse of our services" It’s insane, everybody should read it: