"OpenAI haven’t shared fundamental information about how the problems are selected or the failure rates of their attempts"
This seems unfair. Do human mathematicians have to announce all problems they tried working on and failed to solve?
Already love your podcast. I’m excited to listen to every episode.
Two things:
- The audio podcast version starts with a long ad and has a long ad in between. I’m constantly skipping.
- The YouTube version doesn’t have ads, but is very hard to find. It’s part of the gigantic Bloomberg Podcasts channel instead of having its own channel. There *is* currently an option to “save” a podcast, but that one is not well-known.
I totally get that things have to make money, and that there’s no way around some advertising.
@hanno_sauer@Benthamsbulldog Reminds me of this type of infographic that was popular a while ago, educating you on how to treat a certain type of person (e.g. introvert)
@ArthurCDent@abuchanlife Imagine how much better AI will be at sorting through all the (non-AI) slop products on Amazon to find the one that’s actually good
@leifweatherby People who like solving chess puzzles don’t look at the solutions to the puzzles, even though a chess engine could easily provide one.
If you want to solve a mathematical problem by yourself, just do it! And don’t look at the solution.
Nobody is taking that away from anyone.
@StephenPiment People now use the term ‘slop’ in a non-dismissive way, just to refer to AI-generated content that may be correct and just hard to parse by humans