Thoughts on distributed systems, AI, software engineering, and building things that scale.
Anthropic is now embedding an "imperceptible watermark" directly into the text Claude generates. No banner. No opt-in toggle. No "this message was written by AI" footer. Just a machine-readable fingerprint woven through your output, detectable by a verification tool Anthropic
Read article →Dario says he does not oppose open weights. He just keeps proposing rules that would functionally ban them. The contradiction is no...
Your Claude Code bill has been lying to you. Not in the way you think. Not the model being "expensive" or the...
There is a new #1 on the Frontend Code Arena. It is not Anthropic, it is not OpenAI, and it is not...
Everyone is arguing about which search API to wire up to Claude Code. Brave, Tavily, Serper, Exa, you, me, all of us,...
I read the receipts. So should you, before you type grok -p "fix the bug" into your terminal.
OpenAI shipped GPT-5.6 Sol on July 9th, and if you blinked, you missed the most important line in the entire release. It...
The Muse Spark 1.1 announcement dropped today and the front-page story is "Meta released a coding agent at near-Opus level." That is...
xAI dropped Grok 4.5 yesterday and the benchmark chart is the least interesting part of the announcement. Read the second chart instead....
I want to kill a narrative that has been floating around Western AI Twitter for almost two years now. Every time a...
I use Claude Code and Codex CLI every day, and over time both quietly piled up a surprising amount of configuration I...
OpenAI is reportedly building its own AI Agent Phone for H1 2027, and the entire tech press has spent the last 48...
A reverse engineer who goes by "the real lo" just published a teardown of Claude Code 2.1.196 that nobody in the AI...
Three weeks ago Z.ai was a footnote. Formerly Zhipu AI, a Chinese lab mostly famous for open-weight GLM variants that showed up...
Anthropic dropped Claude Sonnet 5 on Tuesday, calling it "the most agentic Sonnet yet" and positioning it as a cheaper alternative to...
The Wall Street Journal ran a story this week that, if you read it the way it's framed, sounds like a benchmark...
Ever wished Claude Code could just take over the browser tab you're already staring at and fill that form, scrape that page,...
Claude Code's auto-memory is brilliant. It is also incomplete, inconsistent, and trapped on one machine. Here is how I fixed that without...
OpenAI dropped GPT-5.6 Sol this week, and on paper it is a beast. Terminal-Bench 2.1? 88.8 percent. The Ultra variant? 91.9. Claude...
This is not regulation. This is a phone call.