
Recursive launches Flint, a laptop coding model it says was largely built by AI
THE SO WHAT
If Recursive can credibly ship a laptop-scale coding model largely trained and iterated by its own AI system, the bottleneck in model development shifts from human research bandwidth to capital and eval quality. For operators, the question becomes how quickly you can stand up closed-loop systems that design, test, and refine your own domain models without adding headcount.
READ THE SOURCE
MORE FROM THE WIRE
Walmart rolls out AI agent shopping features while Amazon tries to curb the tech
Commerce rails are quietly being redesigned for machine customers — Walmart opening APIs to third-party agents while Amazon resists creates a real A/B test of agent-native retail. If you run a marketplace or brand, you now have to decide whether to optimize for human UX, agent UX, or both.
Mark Zuckerberg says Muse has drawn 10M+ users since its launch and 25%+ of users are making purchases through the app, with an average purchase size of $200+
An AI agent app driving 10M users with 25%+ converting to $200+ purchases means personal assistants are already functioning as high-intent commerce front doors. If you own a consumer brand or marketplace, you need a Muse/agent distribution and attribution plan, not just search and social.
Applied AIMeta says Muse has over 10M users, with shopping a top activity
Muse’s early usage skewing toward shopping and personal productivity confirms that AI agents are monetizing fastest where intent and transaction are tightly coupled. Treat these assistants less like chatbots and more like new performance channels — with their own SEO, merchandising, and funnel analytics.
Applied AIMeta expects its 2026 capex to be $135B-$145B, vs. $130B-$145B previously, and forecasts Q4 revenue of $73B-$76B, vs. $73.6B est.; META jumps 3%+ after hours
Meta nudging 2026 capex up to $135B–$145B while the stock trades up tells you public markets are now underwriting hyperscale AI and infra spend as core, not speculative. For smaller players, this raises the bar on owning foundational compute — partnership and specialization become the default strategy, not a compromise.