It's worth noting that this is a gzip bomb (acts just like a normal compressed webpage), not a classical zip file that uses nested zips to knock out antiviruses.
Other companies have done this (see Qwen Coder). It doesn't scale past a few disciplines like math and code though, and using mixtures of experts give you most of the same benefits.
I'm a bit skeptical of this until it's proven that they're getting the right answers in the right ways. It could be that base models are just more random and when given 200 guesses out of 1000 possible answers tend to distribute them more evenly, bringing up the pass@k number.
Good correction - while the SDK used has supported changing the API through environment variables for a long time, Codex only recently added Chat Completions support recently.
I wonder if using thinking models would work better here. They generally have less variance and consider more options, which could achieve the same goal.
Those points are only true if you're comparing to a React app that uses bloated frameworks and inefficient list manipulation (as other commenters pointed out, `list.push(...items)` causes problems)
Ideally web browsers would use better defaults so you wouldn't have to do this. Firefox is good with this in my opinion: scrollbars are very subtle, don't affect content width, and adapt to the `color-scheme`.