Nicola Romanò

@karthikganapathy@mathstodon.xyz @svat@mathstodon.xyz Or Eddie Izzard's take on it https://youtu.be/TjC3R6jOtUo?si=wlpODLzANqR3Rq9Y

@skewray@mathstodon.xyz @rperezrosario@mastodon.social or a better way of solving it is teaching people they can, and how to do that? I never realised until now that I could!

@grauhausen @zotero Out of curiosity, what type of issues do you see with author handling? I have never really had a problem.

@grauhausen @zotero Ah... I see, I never noticed that, good to know!

@Kiloku@burnthis.town @eb@social.coop If you use a hammer to eat soup you generally don't end up with good results. If you want to sum numbers use the sum function, not one that predicts the next most probable token. I dislike this type of bashing of LLMs because it's trivial to dismiss (ok they can't do trivial maths, but they can write an entire piece of software for me). There are much more risky outputs that could be used as an example. Funky Excel formulas have always existed...

@reinouts@mastodon.green @knud@mastodon.social @Kiloku@burnthis.town @eb@social.coop It takes effort and money to fine tune things and given this is probably just a move to keep some investors happy they can't be bothered? Agree with you it should be possible (also, I suspect it probably would work with a more carefully crafted prompt)

@grayabbott@mastodon.social @jalefkowit@vmst.io That's very much sector-dependent and it is a terrible attitude that should be challenged. Usually those people yelling at each other are the same who don't even share what they're working on with colleagues in their own Institute. They tend to think their work is so important that everyone else will steal it, despite the fact that 95% of times nobody really cares... collaboration is a much more productive attitude

@PalmAndNeedle@norden.social @Kiloku@burnthis.town @eb@social.coop Because code is a series of words, which is what LLMs generate. They're not designed to do calculations. Indeed, sometimes when asked to calculate something these systems generate code that when executed gives the (mathematically correct, of course) answer. The reality is, very often LLMs generate good, working code. The issue is not that. There are big ethical and environmental issues. Also, because the code works often, but not always, you need to double check it at all times, effectively taking longer then writing it by yourself (eg https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/)

@MylesRyden@social.vivaldi.net @terri@social.afront.org Yet, we don't teach students how to do science well. A lot of the bad practices that are behind "bad science" could be mitigated by better teaching of statistics, data analysis, experimental design etc.

@onterof@sueden.social @PalmAndNeedle@norden.social @Kiloku@burnthis.town @eb@social.coop I suspect some big investors want to see AI in every software they put money in. Doubt the decision came down to a bunch of programmers... Also, hype is a funny thing. I've seen students develop suboptimal, very convoluted solutions to fairly simple problems so they could put some transformer or whatnot in there. They're baffled when they don't get full marks...