What tools are people loving for evals? I’ve looked at LangFuse and Evalite and both seem nice so far

The yarn installer using bats and pumpkins for the loading bar is pretty fun

Finally got my own hooks working. It doesn’t need to be this difficult though so the next thing we’re going to do is write a generator

Reading up on AgentOS and this seems like it’s solving a lot of my pain points with AI

Pretty much everything you’re trying to cram into any Claude subagent will be much better served by a Claude hook. It’s really a no brainer with the way the Claude SDK can hijack the already running Claude session

I really need @anthropic.com to make Claude hook generation as straight forward as agent generation

I was really close on the FireTV but got stuck somehow. I had it on the TWRP menu which took all day to get to then wiped the device and now I’m BACK to boot loops lol

Okay finally got the board shorted and connecting via usb. I had to install the arm tool chain to get this built so took a little while but it’s finally flashed

Got the FireTV opened up and shorting the board right now to get it out of the boot loop. It sucks USB connections don’t work well with WSL2 

I’ve got an old flat Fire TV from 2015 that’s stuck in a boot loop. I think I’m gonna try flashing it and putting my own OS onto it

Upgraded to Max finally for Claude Code. Using Opus for plan only mode is so nice