I thought it was supposed to be smart?
>>16807951It can't figure out anything that hasn't already been figured out. We have never seen any evidence of novel ideas.>but muh matrix calculation number of operations boundsthere is nothing novel herebe honest, the ai handlers probably already figured it out and just guided it to the answer, an answer which contains nothing novel.
>>16807951if you've watched things like claude plays pokemon you will realize these things can't actually reason at allwhy normies are so eager to want to believe that they can is completely beyond me
>>16807966claude actually gives reasoning when playing pokemon though
>>16808015it's just words, there is no deductive logic behind them
>>16808018do you think deductive logic is magic or do you not realize thats what logic circuits are?
>>16807951because LLMs are not supposed to be smart
>>16808025how is that relevant to the word of anon >>16808018?
>>16807951We are discussing neuropsychology for now. So shut up. :p
Eric Schmidt says that some people are "very close" to recursively self-improving AI, so you likely will see breakthroughs of this magnitude within the 2020s or early 2030s.
>>16807951It just needs a few more parameters...
>>16808034alright good to know christ cuck. i cant wait for day of the cross.
>>16807951it needs a lot of control over a lot of infrastructure in the real world to experiment and engineer something like antigrav, which is something we should not do, but we will do it anyway
>>16808018Technically, it can be trained to produce proofs based on recycled logic of previous users It canβt invent logic unless a higher Iq person uses it ergo never seeing shit
>>16810154To add to this since I think it can help They need to sample the highest IQ users and use their recycled logic to create proofs for pure and applied physics problems then and only then can they solve the god equation etc