In July 2025, a developer's Claude Code instance hit a recursion loop and burned through 1.67 billion tokens in 5 hours, generating an estimated $16,000 to $50,000 in API charges before anyone noticed. The agent did not crash. It did not throw an error. It just kept calling tools, getting confused, calling more tools, and silently accumulating cost. Old software crashes. LLM agents spend. This is
You're in another app and there's a timer counting down at the top of your phone. You lock the screen and the same timer is sitting there. You swipe down to the Notification Center and it's there too, still ticking. It looks like a notification, but a notification can't tick. That's a Live Activity. It looks like three different surfaces (Dynamic Island, lock-screen banner, Notification Center ent
The Reality Check ClickHouse just dropped a study that every executive should read: LLMs are great at some things, but basing your infrastructure on them? Too much, too soon. They tested five leading models (Claude Sonnet 4, GPT-o3, GPT-4.1, Gemini 2.5 Pro, and the newly released GPT-5) against real observability scenarios. The verdict? We're nowhere near the autonomous operations future Silicon
The 800 Million Weekly ChatGPT Users Who Are Just Getting Started Here's something that should excite everyone: ChatGPT just hit 800 million weekly active users. That's one in ten humans on Earth. Adoption faster than the world wide web. 18 billion messages every single week. And the really wild part: we haven't even scratched the surface of what's possible. OpenAI's latest research shows that ~
I finished an English series on the way I think ordinary people can start using AI for real work. The point is not to become an AI expert first. The point is to have one place where you can say what you want, give the tool access to the right folder, and check the result. Anything important still needs a human pause: publishing, deleting, paying, or authorizing. My preferred starting point is simp
The Model Context Protocol (MCP) has become the default standard for connecting AI agents to external tools and APIs. Governed by the Linux Foundation since early 2025 and adopted by OpenAI, Anthropic, Microsoft, and Vercel, MCP is the USB-C port of the AI ecosystem — one protocol that lets any LLM application talk to any tool server. But there's a gap between reading the spec and building somethi
Background I did some research online and found a nice course that teach how to build LLM from scratch. The course is shared public online and all the assignment resources are here: https://cs336.stanford.edu/. In the following series, I will put the summary and notes starting from lession 1. Tokenization is at the very beginning of the LLM. There were many different tokenization algorithm, suc
If you have spent any real time with Claude Code, you have probably noticed the same problem I did. You write the same instructions in the prompt every other day. "Use four-space indentation here." "Always run the linter after edits." "Format commit messages this way." After the third or fourth repeat, it stops feeling like a prompt and starts feeling like missing config. Skills are how Claude Cod