No-name university graduate and former dishwasher.
Self taught coder.
Philosophy and history enthusiast.
- > Models are tasked with running a simulated vending machine business over a year and scored on their bank account balance at the end.
The article being discussed here is about how AI couldn't run a real world vending machine. There was no issue in the components that would be in a standard simulation.
- 1. niceness and genius are orthogonal
2. your characterization of the article sounds uncharitable
3. my point isn't exactly that this is necessarily the smartest place
- It's entirely possible that this is the smartest place on the internet, but also often dumb. In fact, it seems likely. More of an indictment of the rest of the places on the internet.
- Perhaps, yet it’s a $1.6T company nonetheless.
- But it needs to be convoluted. The problem with the simpler version is the word happy needs to be translated both culturally and more literally.
- Sure, but the headline makes you think this incident caused the uncertain future. It’s definitely clickbait
- I’m confused. Why are you alleging a link between the first few bullet points? I'm not saying there's no link, but it certainly doesn't seem obvious without some substantiation.
- I'm not sure if you meant to reply to this comment. If you did, I'd love to know what you mean :)
- Looks really healthy to me. It’s unhealthy when a partner can’t recognize that they actually were at fault and try to change, but instead needs every fight to resolve with “we were both wrong”.
- Curious what HN thinks about a spaced repetition social network.
You could mark items in the feed to space repeat for yourself. This would also function as a “retweet”, which would align incentives such that content that gets promoted is actually durably useful or interesting. The posts people make would repeat to themselves too, so the source content should be good.
- What’s the point of this? One only can control one’s self. Who cares if the other person was in the wrong too.
- This is what came to mind for me. Certainly there are free versions of everything that used to cost money a decade or two ago. The paid software now is much more advanced.
- Hmm, the way I read “bandit” here was to be implying illegality not necessarily immorality.
- Yes, of course this is right. However, I do think LLMs suffer even more than people from linear extrapolation.
- No, because this is self correcting behavior. If the comments are annoying, people will downvote them. In the rare case this is appropriate, they should be allowed. Guidelines are for things that will naturally be upvoted but shouldn’t be.
- 15 years ago: https://news.ycombinator.com/front?day=2010-12-09
> Evernote makes $800,000 per Month > Dropbox for Teams > Mark Zuckerberg Agrees to Give Away Fortune
- Today's front page is not a clean 10 year extrapolation from this. That's where AI is wrong. The future is weird and zig zags, it's not so linear as the Gemini generated page.
- > Low margins aren't an indicator of anything. They are a component of financial return in addition to capital. One does not make sense without the other. In high frequency trading, they are making 1/100000th of a percent on a trade, that is a very high return business if you can do this millions of times a day. Similarly, if I run a housebuilder then I need a 20% margin because I am going to be turning over my inventory across multiple years.
This is wrong because quarterly reporting normalizes for expenses in same period
Hell yeah, sold, let’s go…
> We're developing a responsible access framework that makes models available to researchers for scholarly purposes while preventing misuse.
Oh. By “imagine you could interview…” they didn’t mean me.