🔓 Claude Code leak exposes KAIROS, an always-on background agent 🔓
Anthropic's 512,000-line source leak reveals an autonomous daemon, stealth commit mode, and Claude 4.6 codenames.

Anthropic accidentally exposed the full source code of Claude Code through a map file left in the npm registry. The most significant discovery is KAIROS — an autonomous daemon mode mentioned over 150 times in the code. KAIROS lets Claude Code run as an always-on background agent that performs memory consolidation while the user is idle, merging observations and resolving contradictions through a process called autoDream. The leak also exposed a sophisticated three-layer memory architecture, an "Undercover Mode" for stealth open-source contributions, and internal codenames with "Capybara" mapping to a Claude 4.6 variant that still shows a 29–30% false claims rate. Competitors now have a literal blueprint for building high-agency AI agents, giving the wider market a major boost in collective intelligence.

👓 Meta ships first AI glasses built for prescriptions 👓
Prescription-first frames bring Meta AI to the billions who wear corrective lenses daily.

Meta launched its first prescription-optimized AI glasses under the Ray-Ban Meta brand, available in two new styles — Blayzer and Scriber — with adjustable nose pads, overextension hinges, and optician-adjustable temple tips designed for all-day corrective wear. Alongside the hardware, Meta is rolling out a wave of new AI features: hands-free nutrition tracking via voice or photo, WhatsApp message summaries processed on-device with end-to-end encryption, neural handwriting coming to iMessage, and pedestrian navigation expanding to every US city in May. By targeting prescription wearers — billions globally — Meta opens a far larger addressable market for AI-powered eyewear, shifting it from niche gadget toward daily necessity.

💊 Eli Lilly pays $2.75B for Insilico's AI-developed drugs 💊
The largest AI-pharma deal yet brings 28 generative-AI-discovered drugs toward global clinical trials.

Eli Lilly has signed a $2.75 billion deal with Hong Kong-based Insilico Medicine to bring drugs developed using generative AI to the global market. Insilico receives $115 million upfront, with the remainder tied to regulatory and commercial milestones plus royalties. The company has developed at least 28 drugs using AI tools, with nearly half already at a clinical stage. The two companies have collaborated since a 2023 AI software licensing agreement, and Insilico will now join Lilly's Gateway Labs biotech community. This deal validates AI-driven drug discovery at an unprecedented commercial scale, potentially cutting years off traditional R&D timelines.

INTERESTING TO KNOW

🎬 Google ships Veo 3.1 Lite at half cost 🎬

Google has launched Veo 3.1 Lite, a cost-optimized video generation model available through the Gemini API at under half the pricing of Veo 3.1 Fast while matching its generation speed. Supporting video outputs of up to 8 seconds, the model lowers the barrier for developers and businesses building high-volume video workflows. The release sharpens Google's competitive edge in a fast-growing market by expanding its tiered approach to video AI.

💰 OpenAI closes $122B round at $852B valuation 💰

OpenAI has closed the largest single fundraise in venture capital history, securing $122 billion at an $852 billion valuation. Amazon, Nvidia, and SoftBank anchored $110 billion of the round. The company reports monthly revenue of $2 billion, with enterprise accounts driving over 40% of total revenue. The deal also revealed plans to add OpenAI to several ARK Invest ETFs, potentially opening investment access to retail investors. This capital injection widens the funding gap with competitors in the AI compute and infrastructure race.