RIP IN PEPPERONI MATHEMATICIANS
https://intelligence.org/2016/09/12/new-paper-logical-induction/
>>8341931
Okay I'm fucking stupid. Somebody explain this shit to me in a way a normal person can understand, and why does it blow mathematicians out?
>>8341960
Somebody finally went and made a long-fabled automatic theorem prover that actually works. Now we can use supercomputers to automatically discover new mathematical concepts at a far faster rate than before. There will be all sorts of crazy new discoveries that will end up revolutionizing the applied sciences.
>>8341994
Really? That's pretty fucking cool Anon. At what rate would mathematical concepts be discovered? And what is the estimated time before any new concept discovered be applied?
>>8341931
Oh shit, the singularity is here
>>8341994
in b4 significantly shorter and more intuitive proof of ABC conjecture
okay, I just skimmed, but it seems like this is just theoretical, or they don't have it working efficiently yet.
If that's the case (please prove me wrong) then nothing to see here. The potential of this machine is cool though
yeah, it says :The logical induction algorithm that we provide is theoretical rather than practical.
As if we didnĀ“t expect nohing less from probabilistic metamathematical approaches...
However still very cool and interesting, and still seems like a good object that someone will definitely use in some point later.
>>8342171
>but it seems like this is just theoretical
They say inefficient.
That means it doesn't work and possibly it never will.
Now they finally need to get one of thosed formalized math database/wiki things with auto proof checker working. Then let the ai have access and do there thang
>>8341994
>that actually works
Don't get too excited. It's still in the theoretical stage. From what I can tell, they have so far used machine learning algorithms applied to mathematical theorems to guess which theorems are more or less likely to be true. This is kinda neat, but not exactly earth-shattering. They've pretty much simulated, to an extent, a mathematician's intuition; which, while important, is not enough in itself to actually be useful. For instance, most mathematicians believe the Riemann hypothesis to be true, but this is not sufficient. If their algorithm can actually prove theorems, then we can get excited.
>>8342223
What actually goes into a proof of a mathematical hypothesis? Why do I picture it as some sort of number wizardry?
>>8342225
an equation that proves the hypothesis