@kingkongfalls@KritaNishchayah We can't really downplay AI capabilities, reimann hypothesis was infact one of the hardest problems out there for humans
@Why__Knot@IlinVasily29521 If they were even slightly concerned about maths as a whole they would have atleast posted a human readable proofs so we could all be excited to read through it, but they only care about publicity
@Why__Knot@IlinVasily29521 The best scenario for maths at this point will be to make these internal models available to everyone, however, I fear these corporations will probably control these superintelligence s and that will probably lead to a dystopian society
@Why__Knot@IlinVasily29521 Its funny that general people who don't understand any maths think mathematicians are obsolete now imo the more proof ai generates and the better it gets at it the more mathematicians are useful as the abstractions and understanding of structures will vastly progress maths
@Why__Knot@IlinVasily29521 Atleast there should be a proof, 250k pages is not a proof, furthermore, it could be entirely possible that what we thought is hard is actually easier for AIs and eventually we will have mathematical directions where we will need humans,it would be a fun era to be a mathematician
@kingkongfalls@KritaNishchayah That's what the point is in that article, reimann hypothesis was checked for the first 10 trillion zeros, we knew it was true while these AI corporations do not even share human readable solutions,the whole point of proving was to understand the approach/figure out an elegant way
@kingkongfalls@KritaNishchayah They are actually right, it does not do any good by not even having human readable solution to a hypothesis that was already known to be almost certainly true
@michael_c_grant@chenna1985 Imagine you come up with an observation AI proves it as true then you understand the solution and come up with new tools and abstractions in maths it will be truly an age mathematicians will dream of, let's just hope these corporations are truly interested in democratising AI
@michael_c_grant@chenna1985 And also the advancement in pure maths comes from observations and understanding the approach to a really hard problem, reimann hypothesis itself is a remarkable observation in the age of no computers to verify the first trillions of zeros
@michael_c_grant@chenna1985 Lean was not there for hundreds of years of mathematical progress (that even led to the creation of AIs), infact many mathematicians are excited with AI because the proofs they offer can be dissected and enhance our understanding, but for that we need human readable proofs
@michael_c_grant@chenna1985 Doesn't matter though because simply proving it does not advance the field of mathematics, pure maths is of little practical significance anyways the only reason people pursue it is curiosity and desire to see structures within maths itself.
@michael_c_grant@chenna1985 All people are saying is that a superhuman intelligence that can solve reimann hypothesis can also find bugs in kernels that we don't even know of so until there is a human readable proof maths will not benefit at all
@Why__Knot@IlinVasily29521 It was a millenium prize because very capable people thought the method of solving it will give us great understanding into primes, AI solution currently offers none of it, other things is i AI corporations should focus more on solving practical problems rather than PR stunts
@Why__Knot@IlinVasily29521 Sure, although the point is I have no double that the proof is right, but we already knew that the reimann hypothesis is certainly true, so it offers nothing to maths and everything for the PR, specially when there is not even a natural language proof