// a daily digest from the AI world
New models, tools and announcements — every day, always with a source link. This digest is assembled by an AI agent following my rules, and it's the same briefing I use myself to keep up with AI without burning time.
OpenAI has shared the first results of Jalapeño, its custom inference chip. The company says it delivers faster, more power-efficient inference with higher throughput and lower latency for modern models, built to serve today's models at scale.
IBM has introduced Granite 4.2, the new generation of its open language models, with a Hugging Face writeup on how they are built. The Granite line targets enterprise use and expands the pool of open models you can run on your own infrastructure.
Anthropic is giving Claude shared memory across chat and Cowork. Users no longer have to repeatedly brief the AI on their projects, preferences, and other context, since it now carries that information between the two.
OpenAI has introduced an Admin plugin for ChatGPT Work and Codex. It lets admins analyze workspace usage, manage members and permissions, adjust limits, and act on admin requests, which helps companies running these tools across larger teams.
A new Hugging Face writeup presents quantization-aware healing: a compressed, 4-bit model that outperforms its full-precision original. It points toward smaller, cheaper models without sacrificing quality, which is appealing for deployment on more modest hardware.
OpenAI has banned Russia-origin accounts that used AI to promote a fake Israel-based think tank and a “sovereignty” index praising Russia and criticizing the West. The company continues to uncover covert attempts to misuse its tools for influence operations.
This section writes, assembles and publishes itself every morning — from source selection to deployment. If you'd like a quiet helper like this in your company, let's start with an initial consultation — the fee is credited toward the project if we work together.