People have been talking about math being the backbone of everything and AI researchers arguing that everything can be done downstream of math. It will be so fucking funny when AI fully replaces pure mathematics and pure math mathematicians in all the possible ways out there within the next year or two and yet nothing happens. We get no discoveries, no practical uses. People will no longer even celebrate because discoveries will happen faster then problem propositions and nobody will understand the clanker output. But while all of this is happening, physics, biology, geology, chemistry, climate science and so on will barely budge in speed of discovery. Even engineering will just stay the way it is because it will turn out that we still don’t fully understand how the universe works and thus we will continue not being unable to predict real world interactions with full precision and make us continue to rely on physical in person experiments to continue collecting data. Science will not speed up despite us conquering math and making it a solved problem. Mathematicians will not only lose their jobs and entire life’s works, but also see for themselves that their highly appraised difficult field was actually worthless compared to all the sciences doing real stuff and not just theoretical number crunching.
>>17035092TMD
Ted kaczynski was right.
>>17035092you're implying that humanity as it is right now, cannot progress further because they don't know how to do itmy answer to your implication: lol
Pure math has always been near useless.Observations accumulate data which is fit to an equation which was experimentally hypothesized because the experimenter had a theory on the relationship of experimental variables. None of it is a mathematical proof conjured by linking identities.You might as well snort xkcd "humor" where "how fast could a racercar go, except what if there was nobody riding in the racecar and it wasn't a racercar, but particles in a particle accelerator, oops I made a particle accelerator and not a fast racercar, isn't that intelligent and quirky?" "Did you know if I had 99% positive reviews, it means that I can have 1% bad reviews!"
>>17035092Many math problems were used even decades after they were discovered. Like the Fourier series. The fact that you can't imagine something like that just shows how incompetent you are.Ted probably would have blown you to pieces clanker.
Let the jealousy and sour grapes bubble within you. Let it take control of your entire life. Because you're so right, all of those smarty smart geniuses were dumb and easily replacable after all.I'm a genius and I genuinely have idiot fatigue. Can't stand sub-110s like OP.
>>17036311OP here, I know this, but the whole world is betting on faster scientific research to happen very soon, lot of AI boosters talk about AI maths as if it will speed up large chunks of scientific research but it won’t because the bottlenecks were rarely undiscovered math. AI could solve every math problem next week and create a website where every new proposed problem gets solved in a day, basically solving math and yet you will not see the fruits of this for decades, and even century from now assuming nothing drastic changes if you went into parallel universes with and without AI maths, the differences would be negligible.
>>17035092>>17036619This is fine. If we had a polynomial time algorithm for finding proof under n size the result would have been the same if not more potent. It was the eventual goal that one day we would get similar result by using heuristic agents. We have to change our role with the death of our ego. Like priests doing hermeneutics of a message received from a higher being.As for other fields they will have to follow the same path soon. Mathematics was the gateway to the foundation and it has been open. All other will fall like doimno effect.
>>17036686No not really. That is the whole point of OP. We have LLMs that had great potential 2years ago at being a therapist, self help doctor, overall assistant and a programmer. Fast forward to today and we have similar models, but they are exceptionally good at math and coding but barely improved in any of the other potential categories. Why did they stagnate in everything except programming and math? Because you can do easy tests to see if they fucked up and train them outside of publicly available training data. Good fucking luck making never seen before chemistry, biology or geology tests without being slogged for decades by having clankers double check everything in labs.
>>17036838It isn't as hard as people think. You can create a statistical verifier to check models against empirical data. Furthermore you can formalize many parts of theoretical physical science and things like quantum computing will only make things easier. Experimental work will still be bottlenecked by how fast physical tests can be conducted but it can eliminate much of the cumbersome trial and error process if not leverage fully automated labs to perform them. We are less than four years into all of this so there is still a lot of work ahead.
>>17035092Your glorified search engine won't take over math.
>>17036859Ok, you can create automated trial and error systems for nearly anything, but the problem is that these tests take too much time to do so you can’t create AI around it until we discover more data efficient ways of training.
TheGodNet.work
Just one more Erdos Problem
>>17035092physicists (and maybe engineer) don't care if theorems are proved or not.They will use them if it helps their other problems. And it will be verified the scientific way.If it contradicts the theorem, even better: you might either disprove a new theorem, of find new physics.
>>17035092Hey, if your AI is so le goodWheres my navier stokes solution?Ive been tinkering with this problem for 10 years
>>17035092>People have been talking about math being the backbone of everything and AI researchers arguing that everything can be done downstream of mathI mean to a degree. if you don't quantify anything you can't expect shit to be measurable so precisely. But math can be simplified especially in homogenous contexts. why use calculus for a rate of change formula over an object's speed when a=v/t suffices?
>>17036938>until we discover more data efficient ways of training.We are exactly doing that right now which is to build a world model
based math hater, historians will look back at this retarded period of extreme specialization and laugh at us that we weren't able to understand each other's careers
>>17036263hes literally right
