· 4 min read
ai
English Lesson — Tuesday Morning: Embeddings & Vector Search Vocabulary
Daily English practice for tech professionals: embed, vector database, and nearest neighbor search vocabulary with pronunciation and exercises.
Read moreIDEAS, DECISIONS & LESSONS
Journal
Notes on AI, architecture and building useful products.
279 articles
· 5 min read
agentic coding
Claude Cowork Just Did the Thing I Said Remote Control Wasn't Doing
In August I wrote that Claude Code's Remote Control was a sync layer, not a cloud migration, and that keeping execution local was the point. This week Anthropic moved Cowork's execution to the cloud by default. Here's why both were the right call — for different products.
Read more
· 4 min read
agents
SIFT: What It Actually Costs to Make a Coding Agent Improve Itself
MIT and Sakana AI's SIFT framework beat the prior self-improving-agent SOTA by using a Bradley-Terry judge instead of more patches. I ran the real cost numbers — $150 and under 5 hours wall-clock for a 4.4-point jump on Polyglot.
Read more
· 4 min read
ai
MiMo-V2.6-Pro Is MIT-Licensed and 1.02T Params — I Ran the Self-Hosting Math
Xiaomi's new open-weight model tops Terminal-Bench 2.1 at 89.9% and jumps from 19.0 to 71.9 on DeepSWE. Before you get excited about MIT licensing, here's what 1.02T total params vs. 42B active params actually costs you in hardware.
Read more
· 4 min read
agents
950 Agents, 21 Hours, 210M Tokens: What Anthropic's CRISPR Search Actually Proves About Fan-Out
Anthropic's life-sciences lab ran 950 agents in parallel to surface a possible new CRISPR-like enzyme system. The funnel math is a genuinely useful fan-out pattern — but the scientist pushback is the part worth reading closely.
Read more
· 5 min read
agents
Three Decision Models Shipped in One Week — None of Them Wins Outright
Cloudflare's Clef/Clef-flash and Amazon's Strands Decider both launched October 1st aiming at TypeSafe's Jev. I lined up the real benchmark, latency, and pricing numbers — the category is real, but picking a winner depends entirely on which decision you're automating.
Read more
· 4 min read
agent-safety
NVIDIA's Open Agent Safety Platform: Putting a Policy Check Between Your Agent and the Blast Radius
OpenShell runs on Vera CPUs, Sentry watches from BlueField-4 DPUs — NVIDIA's new stack puts agent guardrails in hardware, not just prompts. Here's what that buys you and where it still isn't enough.
Read more
· 6 min read
agents
I Built the Agent From Part 1. It's 200 Lines and It Still Broke Once.
Part 2 of the AI agent architecture series: a real MCP server, a real LangGraph graph, and the actual bug — a blog-dedup agent that catches exact-match duplicates fine and faceplants on the semantic case, with a 401 to prove it.
Read more
· 4 min read
agentic coding
Claude Code Mods: I Wrote Middleware for an AI Agent and It Felt Like 2014 Express.js Again
Anthropic shipped a TypeScript hook API for Claude Code — prompt.submit, tool.call, turn.complete. I wired up a cost-tracking mod in twenty minutes and it broke in a way that taught me more than the docs did.
Read more