Introducing Claude Opus 4.6. Our smartest model got an upgrade. Opus 4.6 plans more carefully, sustains agentic tasks for longer, operates reliably in massive codebases, and catches its own mistakes. It’s also our first Opus-class model with 1M token context in beta.

GPT-5.3-Codex is here! *Best coding performance (57% SWE-Bench Pro, 76% TerminalBench 2.0, 64% OSWorld). *Mid-task steerability and live updates during tasks. *Faster! Less than half the tokens of 5.2-Codex for same tasks, and >25% faster per token! *Good computer use.

Yo OpenAI vs Anthropic is literally this generation Windows vs Mac

What a day!

NASA's An Introduction to Tensors for Students of Physics and Engineering by Joseph C. Kolecki PDF: grc.nasa.gov/WWW/k-12/Numbe…

GPT-5.3-Codex-Spark is now in research preview. You can just build things—faster.

Teams treat AI like "just another feature," only to get blindsided by the bill! Yes, this happened to me. :) AI isn't priced like software licenses. It's priced like a utility. Every prompt, agent step, and retrieval call is metered. Tokens are basically your new kilowatt-hours.

so many startups just died

You’re in an ML Engineer interview at Google. Interviewer: We need to train an LLM across 1,000 GPUs. How would you make sure all GPUs share what they learn? You: Use a central parameter server to aggregate and redistribute the weights. Interview over. Here’s what you missed:

Probably a hot take, but James Bond seems like an HR nightmare.

We're launching Claude Community Ambassadors. Ambassadors host meetups, bring builders together, and partner with the Anthropic team. The program is open to any background, anywhere in the world.

New on the Anthropic Engineering Blog: How we use a multi-agent harness to push Claude further in frontend design and long-running autonomous software engineering. Read more: anthropic.com/engineering/ha…

NASA's Artemis II mission nasa.gov/blogs/missions…

Why is the Strait of Hormuz a geological wonder? nationalgeographic.com/science/articl…

Introducing Claude Managed Agents: everything you need to build and deploy agents at scale. It pairs an agent harness tuned for performance with production infrastructure, so you can go from prototype to launch in days. Now in public beta on the Claude Platform.

x.com/i/article/2042…

The issue is not just “I manage my agents.” The issue is “we manage agents horizontally across a team, siloed context, domain assumptions, and delivery artifacts.” As we transition to Agent Bosses, how do we still work as a team. Curious about everyone's thoughts on this shift.

Just because LLMs can spit out words doesn't mean they're doing engineering. We have problems to solve and people to serve, and you should be learning the LLM new tool.

I am starting a case and movement to add Solar Power Canopy to all parking spaces in the U.S. as a Public Good.

"U.S. Debt Tops 100% of GDP." My bad. Nevermind. 86 this idea. 😁