Daily Digest — 2026-01-12
Must Read
[AI] Apple picks Google's Gemini to power Siri — Major partnership shift. Apple outsourcing AI to Google signals they're behind on foundational models.
[AI] The next two years of software engineering — Addy Osmani on how AI changes the craft. 238 points on HN.
[AI] Simon Willison's LLMs in 2025 year-end review — Comprehensive 26-section roundup. Covers agents, Claude Code, Chinese models, OpenAI losing their lead. Reposted by Karpathy.
[AI] Google launches Universal Commerce Protocol for AI agents — Infrastructure for AI agents to transact. This is the rails for agent economy.
Worth Reading
[AI] Karpathy on LLM mental model — "The correct way to think about LLMs is that you are not optimizing for a single specific model but for a family of models controlled by a single dial — the compute you wish to spend."
- https://x.com/karpathy (Jan 7)
[AI] Nvidia and Eli Lilly invest $1B in AI drug laboratory — AI + pharma convergence accelerating.
[AI] Anthropic launches Claude for Healthcare — Vertical AI play in healthcare.
[AI] OpenAI Codex agent updates — gpt-5.2-codex shown, skill invocation with $. "A ton of new things coming in Jan." Sam Altman repost.
- https://x.com/thsottiaux (Jan 2)
[AI] FUSE is All You Need — Giving agents access to anything via filesystems. 188 points on HN.
[STARTUPS] Alphabet reaches $4 trillion market valuation — First company to hit this milestone.
Monitor
[AI] Chinese AI leaders warn of widening US gap — Geopolitical AI dynamics.
[AI] Google removes AI Overviews for medical queries — AI liability in health context.
From X/Twitter
@karpathy
- Jan 7: nanochat miniseries v1 — LLMs as family of models on compute dial
- Jan 6: "The majority of the ruff ruff is people who look at the current point and people who look at the current slope"
@sama
- Reposted ChatGPT Health launch (Fidji Simo)
- Reposted Codex agent updates (gpt-5.2-codex)
@naval
- Reposted Mastering Zcash guide — crypto/privacy focused
Technical Deep Dives
- Reproducing DeepSeek's MHC: When Residual Connections Explode — https://taylorkolasinski.com
- LLVM: The Bad Parts — https://npopov.com
- Zen-C: Write like a high-level language, run like C — https://github.com/z-libs
Sources: Hacker News, Techmeme, X/Twitter