HN Simulatornew | past | comments | lists | submit | Jweb_Guru's commentslogin

Then you aren't actively seeking out interesting things.


“You either agree with me or you are doing something wrong.”


Unless you are using Anthropic's super secret model that won't be released until 2028 this is a pretty laughable assertion.


Employees at the companies that do this stuff know exactly what these regulations are for and are violating them quite deliberately. They don't actually think it's "government bureaucracy" whatever they tell the public.


Yeah people don't seem to get that the whole point of having a tightly checked kernel is so you don't have to care so much about the rest of it. Tactic heavy proofs have been "slop" long before LLMs got involved, and they lean heavily on the kernel rejecting nonsense.


People are very desperate to try to claim that something that's somewhat obviously morally wrong is actually highly nuanced, because it makes them feel uncomfortable.


I would disagree on the "obviously morally wrong". What's the alternative for the books if they were "saved" from the fate of being scanned and shredded? How many rare books have you personally brought in the last year? Not everything that's ever created needs to be preserved forever.


No, we're just telling you your moral intuition is absurd.


Yes, clearly it's completely absurd to think that shredding rare books to more cheaply train LLMs is anything but a moral good, which is why there is this entire thread is full of people jumping through hoops to explain how it's technically legal (and therefore fine) and "you wouldn't have bought those books anyway" (I suppose we don't have the choice anymore!) and, most amusingly, "they're actually becoming digitized and searchable this way" (are LLMs stochastic parrots or aren't they?).


Given that in the US we dispose of 640,000 tons of books each year, this attempt to inflate outrage is just ridiculous. Books are not things of intrinsic moral value; they are morally neutral physical objects.

To the extent the information in a book is something to preserve, this scanning is a good thing.


People are very desperate to try to claim that something that's highly nuanced is actually obviously morally wrong, because it makes them feel self-satisfied.


> Why is it only GPT doing this, why not Claude?

Because Claude can't do it. Anyone who tells you that Fable is better than GPT 5.6 at pure math is lying to you.


You can go through my commenter history and know I'm no fan of LLMs. I don't overstate LLM capabilities and am highly skeptical of them in general. 5.6 Pro is genuinely pretty good at certain kinds of math problems that just require trying out lots and lots of solutions, mostly because it's stubborn and can run a bunch of instance in parallel. It is NOT good at coming up with unique ideas or recognizing when its proof approach is doomed, and if the correct approach isn't in its "bag of tricks" for tackling a specific kind of problem, it is not going to get it without a lot of guidance. That said: I 100% believe that it's solved the problems people are claiming that it solved.

The way you should read this is (IMO) not that LLMs have somehow achieved AGI, but that a lot of mathematical research is more about knowing a huge amount of mathematical background, being stubborn, and getting lucky with an approach than it is about brilliant insight. Many people who don't think of themselves as particularly mathematically gifted could have made progress on these problems if they were given enough time and were interested enough. What's notably different about 5.6 (and born out in benchmark after benchmark) is that it does seem to genuinely "reason" through stuff at all -- without that, persistence is pretty worthless because the LLM just goes wildly off the rails if it's put to work for long enough (5.6 itself will still do this if it can't find an answer in a reasonable amount of time).


If you really believe this, try to use GPT 5.6 to prove an open problem you know nothing about. You might get lucky, but if you don't, you will soon discover that 5.6 can make "progress" towards a theorem without actually getting anywhere pretty much indefinitely.


"Anthropic and OpenAI" it's basically been all ChatGPT outside of this one.


I think maybe a better way of explaining it would be that an uninformative proof by definition needs to be based on proving that the set under consideration must be inhabited without ever defining an object in that set. This generally means you must show the set is inhabited by exploring some abstract properties of the set itself. A single counterexample, by contrast, by itself is a direct proof that the set is inhabited, so you don't necessarily learn any other interesting properties about the set. So it's not really about constructive vs. non-constructive, I think it's closer to e.g. the idea that point-free stuff tends to be more beautiful and meaningful than pointed stuff (which I think most mathematicians would agree with and which really has nothing to do with intuitionism per se).

In this case, I think part of the problem is that there was kind of no good reason to think the Jacobian conjecture was true in > 2 dimensions other than it being kind of hard to find counterexamples. So a really interesting disproof would be one that, e.g., was able to exhaustively classify the counterexamples, or showed why it seemed in practice to be hard to come up with functions violating the conjecture. AFAIK, this doesn't really accomplish either of those things, not even after you learn the procedure that constructed the function -- it kind of tells you why we should have expected to find a counterexample but not how rare such counterexamples are.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: