An excerpt of chapter 2 of AI Agents with MCP

Good evening from Greece! While I’m catching up on the Australian Grand Prix, I wanted to share this great thread that analogizes post-training optimizations in LLMs to in-season car improvements in Formula 1:

Consider F1, most of the teams show up to the beginning of the year with a new chassis and engine. Then, they spend all year on aerodynamics and systems changes (a minor over simplification), and can dramatically improve the performance of the car. The best F1 teams improve way more during a season.

I’m a huge fan of the ATProtocol, the protocol that powers Bluesky, and not just for the protocol itself. The community of builders that has grown around it (the Atmosphere) is impressive, and we are starting to see some really interesting projects reach escape velocity.

Tangled.sh is one of those, allowing users to host git repositories on lightweight headless servers called “knots,” while consolidating all knots on the network into a single view via the Tangled AppView. Exciting days ahead for decentralization!

Inspired by Carlos Vargas, who was in turn inspired by Simon Willison, I’ve decided to start doing link posts here. Short-form commentary is important, as is sharing links, and I don’t want that to be the sole domain of ephemeral, high-volume social media. Maybe it’ll end up good newsletter fodder for those who prefer that format.

via Carlos Vargas on Bluesky; follow Simon Willison on Bluesky too.