but we already know it's not. self training leads to model collapse

so, half an hour's worth?

Or maybe "how hard is it to find a file you saved 1 minute ago?"

why dmes spacex need an ide?

and it doesn't support NTLM auth afaik! Which is an extra bonus, as I don't remember there being a way to disable it in the windows version

Are the tests testing?

Mostly yes — they were written by Claude/Gemini/ChatGPT, but they've already caught several regressions introduced by the models during later refactors; there’s a fascinating loop in letting one iteration of an LLM audit the next via a static test suite.

WDYM mostly? Did you scrutinize the tests even harder than the code?

crypto mining firms have specialized asics for mining. They are incapable of doing anything else.

I have never had it automate any tedium for me. Because if I didn't write it, I need to scrutinize all of it, but without the benefit of a prebuilt mental model. Reading through tons of similar looking code (because it's supposed to automate boilerplate) looking for subtle mistakes is mind numbingly boring. It's like looking at a wall full of periods in font size 8, and trying to find the one comma.

_Writing_ tons of similar looking code is equally error prone.

Only a complete psycho doesn't copy-paste code when doing something thats 90% repetitive.

Ironically our AI PR review has caught a bunch of those during unrelated commits :D

what i really hate is the "remind me later" buttons. I want to say a plain "no" but the app won't let me. It promises not to respect my decision right there in the popup itself!

GPUs in your average home PC has a longer lifespan. Datacenters run them at full load for very long periods of time. Some datacenters literally burn through hundreds of GPUs a day.