Formative London experience was having a KX train to Glasgow get canceled and having to walk with an ancient irate Scottish lady to hellish Euston station for an alternative route

Which are these?

Worth it tbh

You’ve sold me on Rye

I guess this gets messy bc (a) each model will respond differently to template “harnesses” and (b) so will each eval

I’m sure this is a naive question but what would it look like to do best-effort cost-matched perf? I.e. for a model comparison where one is much better but orders of magnitude more expensive, you can try all sorts of multi-agent setups (parallelization, search, review) while coming well under cost. Some tasks would still fail, but others wouldn’t. I guess ppl show this with reasoning effort but it’s far from the only knob available. 4-Argon Minecraft win related

It’s always faster than I expected, somehow 😅

Hell yeah let’s go @ansonyuu

Beautiful. What’s a house go for there 🤔

This is what i thought London would be

Do one in London plz 😭😭😭

@nickcammarata apparently!

oh wow

Hope you get well soon 🙏

It’s good!!

🫡

👀👀👀

Hell yeah! Congrats 🚀🚀

Look upon my work ye mighty and despaaah

I’m gatherin nuts for da wintah