// daily-digest · thu 25 jun 2026 · morning refresh · last 48h focus
Nine items on Anthropic’s Senate letter accusing Alibaba of AI model distillation, OpenAI’s first custom chip, a continuing Google-to-Anthropic talent wave, and Ben Thompson’s firsthand vibe-coding account.
— refreshed for you, in 9 items.
Anthropic’s Senate letter accuses Alibaba of running the largest known campaign to extract its AI model capabilities through distillation.
OpenAI and Broadcom unveiled Jalapeño, a custom LLM inference chip designed to accelerate AI deployment, targeting availability by end of 2026.
Gemini contributors Jonas Adler and Alexander Pritzel are planning to join Anthropic, extending the recent Google AI talent wave.
Thompson documents building a real personal-use app through vibe coding and draws ten observations about the practice and its implications.
Hoffman argues both OpenAI and Anthropic can coexist, characterizes xAI as struggling, and says SpaceX is not an AI company.
Willison describes using Claude Code to port a 0.2B-parameter image inpainting model to run entirely in the browser.
Singh describes leaving Google to build an interview prep startup and argues deep expertise persists alongside AI coding tools.
The Rails Foundation announced 20 speakers for Rails World 2026 in Austin covering AI integration, framework updates, and deployment.
NASA’s SWOT satellite data shows large tsunamis scatter into multiple trailing waves during ocean transit, contradicting decades of assumptions.