I bet she could do it. Maybe not in a sentence, but at least the Kakeya problem is easy to understand, so maybe these other things would be explainable by an expert. I would like to see them try at least!
fwiw, I feel the same way about biology "Lysing action of the (1,2)b-carotene receptive encephalopathy pathway" type shit.
I'm pretty much hardline anti-AI and even I would say this is too far. If I read documentation or ask my wife to review something, those people did not write the final product. Perhaps it would be mentioned in a citation, like this person has.
I think most people know this and are fine with it. YC owns the site and advertises their stuff on it sometimes. The site itself is not trying to milk you for every penny or trying to exploit you.
Naw, limiting LLMs for an event that's specifically about learning, growing, and collaborating makes a lot of sense. If it ends up dying it won't be because of their stance on LLMs for a conference.
AI can find useful exploits but the highly publicized ones are among a sea of false positives and the successes I've read were found by people who were already experts. I can 100% see a public bug bounty program being inundated with garbage even if there are diamonds in the rough.
Is this a troll comment? I don't see where the author used AI to generate the code and if you don't see the point of experimenting with technology, you're on the wrong website.
Too late for me. I was on Spotify since 2013 and switched to Qobuz due to AI, bad recs, and dislike for the company. Qobuz puts much more effort into manual curation so I still find awesome weird music and have encountered 0 AI. Mainly due to not relying on recommendation algos anymore. I'm sure there is still AI in there. Only issue I've encountered is an annoying playback bug when switching from wifi to data.
You would also have a whole team of consultants, advisors, lawyers, and VP+ people specializing in each area telling you what the problems and possibilities are if you actually had that job. They're not operating in a vaccuum.
Of course they would. A criminal is just a person. And with such an extraordinary percentage of the US population in prison, you can expect the full spectrum of ability, intelligence, passion, compassion, and everything else. Our prison system is an extreme tragedy that most people are numb to because it's been that way forever.
I think this is true for many. Hence paid services that will keep track of subscriptions for you. I saw somewhere a proposal to force subscription services to refund you if you didn't interact with it at all during the payment period. That seems reasonable. If it's only profitable through people's inattention, it's a leech and shouldn't exist.
I agree. When my mother died I got access to her emails, diaries etc. I read some and as you would expect there are a whole range of emotions and opinions in there, many of which I did not care to engage with. So I asked my wife to read some and she said said she thought it was worth keeping so we do. I will not read it, but perhaps someone else will get some value from it someday. It's no effort to keep (no boxes or terabytes of data).
It's a funny way of imposing a very large fine. Make the service only available during predefined "visitation hours", prevent updated learning except from resources available in the prison, restrict speech and actions according to prison rules etc.
Maybe for literate programming, we can switch from common, ambiguous human languages like English and Spanish to [Lojban](https://en.wikipedia.org/wiki/Lojban)! That way our human language will be unambiguous which will translate to machine code much better. We'll call this the de facto "language for programming". Improvements and other variants may pop up in the future as new needs arise. All that is old is new again.
This is exactly what turned me off. It supports so many paradigms that every line of code I wrote I had to sit and think if I was doing it the "right" way and it was miserable.
Whenever I see defences of AI "art" people very often reduce the arguments to these analogies of using tools, but it's ineffective. Whether you use MS Paint, Photoshop, pencil, watercolor etc. That all requires skill, practice, and is this great intersection of intent and ability. It's authentic. Generating media with AI requires no skill, no intent, and very minimal labor. It is an approximation of the words you typed in and reduces you to a commissioner. You created nothing. You commissioned a work from a machine and are claiming creative authorship.
Except a decent number of those examples have legally binding oaths not to screw you. And if they are found to have screwed you, are barred from their profession and have to pay you a lot of money. Maybe that should be more common.
I know nothing about theorem provers. Is the idea that you can prove a large number of simpler statements that build on each other until it all implies the top level theorem you're proving (like how IRL math works)? That way you don't have to understand the gargantuan computer program all at once?
It just seems like it would be as hard to verify the accuracy of the code written to prove a complex theorem like FLT as a manuscript written in English. But if you can rely on smaller statements that build on each other, it would make more sense.
To me, using AI to convert an idea or paper into working code is outsourcing the only enjoyable part of programming to a machine. Do we not appreciate problem solving anymore? Wild times.