I’ve stopped saying “I used an LLM to make this” like it was some sort of punchline. It’s just how I work and the LLM did some of the work, but I knew what I wanted and how I wanted it built before I started.

Listened to that (excellent) interview in the car. That was one slippery fish. Well done for keeping it honest 👏

Spent Ignite either backstage or running around between expert areas and labs. Spent time with amazing colleagues. Covered about 60km 🚶🏻 got asked hundreds of questions for which I had answers for a few and saw this weird frog thing in SF Bay

Worked on several demos and bug bashing for Ignite. Helped with some of the tech showcased at the keynote in front of 20,000 people. Spent the week on campus in rehearsals and running around solving problems. Found a room with unlimited chocolate.

If Microsoft announces the acquisition of Chipotle, we know the root cause

Yes but the health insurance premiums… 💰

@bn1n.bsky.social is Tokyo Python meeting that week?

Clipchamp is good enough for the basics

I read a paper on how “you are a “ makes little to no difference. Although that didn’t include Claude in the benchmarks.

The main reason I get it to write tests is to validate the weirdness in the code it generated and let it iterate without me. I’ll delete a lot of it later otherwise you’re taking on tech debt in these test suites couple to the implementation as you point out