Skip to main content
19 events
when toggle format what by license comment
Jan 24, 2023 at 13:20 comment added fedja @Frax at the same time considering Hitler the best candidate as the next USA president If it does not know who Hitler was, it should be smart enough just to say "I don't know who that is, so I cannot tell whether it is a good candidate or not"; if it does know who that was and still considers him a good candidate, then, apparently, its logic led to that conclusion and it should show that logic. Hitler, BTW, has a huge advantage over many other politicians: he is already dead. So in some elections I would certainly vote for him over both main candidates :-)
Jan 24, 2023 at 12:07 comment added Frax @fedja Sad part is that it's necessary, as it may have prevented further optimization of chatGPB in general if it were capable of solving very difficult math problems while at the same time considering Hitler the best candidate as the next USA president...
Jan 24, 2023 at 0:24 comment added fedja Wolfram Alpha is incapable of proving or disproving anything. Mainly it is just a symbolic manipulator, so unless it can reduce it to $e\sqrt 5$ or something else like that, it has no opinion about the irrationality of the number. On the other hand, chatGPT would be capable of doing such stuff by now if its creators hadn't wasted time on teaching it who Columbus was and how to appear nice and virtuous and answer clear and well-posed questions by long preaching about good and evil when the conversation makes a dangerous turn :-).
Jan 23, 2023 at 16:08 history edited YCor
edited tags
Jan 23, 2023 at 16:00 comment added Frax I changed the title and added one more example to generalize the point of my question.
Jan 23, 2023 at 15:59 history edited Frax CC BY-SA 4.0
Changed title and added one more example to generalize my question
Jan 22, 2023 at 21:21 history edited Frax CC BY-SA 4.0
Correction.
Jan 22, 2023 at 18:19 history edited Frax CC BY-SA 4.0
Inverted order of some values for a more readable experience
Jan 22, 2023 at 18:14 history edited Frax CC BY-SA 4.0
Error in the proof: corrected.
Jan 22, 2023 at 15:23 comment added Carl-Fredrik Nyberg Brodda @CarloBeenakker Well, ChatGPT has given me several elementary proofs of Fermat’s last theorem, so I’d take anything it says about maths with a massive pinch of salt.
Jan 22, 2023 at 15:12 history edited Frax CC BY-SA 4.0
added 5 characters in body
Jan 22, 2023 at 15:04 comment added Frax @GeraldEdgar isn't that a proof about the standard Bessel function of the first kind? Does it still work here (Modified one)? Genuine question
Jan 22, 2023 at 15:00 comment added Gerald Edgar Proof here ... mathoverflow.net/a/103161/454
Jan 22, 2023 at 14:57 history edited Frax CC BY-SA 4.0
added 264 characters in body
Jan 22, 2023 at 14:41 history edited Frax CC BY-SA 4.0
[Edit removed during grace period]
Jan 22, 2023 at 13:24 comment added Carlo Beenakker chatGPT doesn't know either... "It is not known whether the value of I_0(2) (Modified Bessel function of the first kind) is irrational or not. While some special values of the Bessel functions are known to be irrational, there is no proof that all values are irrational, and it is not known if there are any general methods for determining whether a particular value is irrational."
Jan 22, 2023 at 12:33 comment added username Yes, your proof is (to my eyes) perfectly natural and correct. Is unknown an error ? It could be unknown to the person who wrote it. There are other mistakes of that type in Wolfram Alpha (I know of one with respect to the loosest sphere packing, which has been shown two decades ago to be arbitrarily close to zero,, and yet is still announced as 0,055%).
S Jan 22, 2023 at 11:32 review First questions
Jan 22, 2023 at 13:11
S Jan 22, 2023 at 11:32 history asked Frax CC BY-SA 4.0