[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

View post   
File: 13 KB, 275x183, download.jpg [View same] [iqdb] [saucenao] [google]
15337316 No.15337316 [Reply] [Original]

I've been dreaming of being a pure mathematician since I was 7, but AI has come out now that it's time to transition into post-secondary. Is it even worth trying to get into a math program in hopes of landing a researcher position in the future now that AI can do math better than humans? What the fuck should I do

>> No.15337318

AI will not replace pure mathematicians. We've tried building special purpose compilers that can compute new theorems based on the axioms and rules of mathematics since the 70s but it hasn't worked.
>t. knower

>> No.15337322

>>15337318
What if a Large Language Model (LLM) was trained solely on mathematical texts?

>> No.15337325

>>15337322
It fundamentally cannot work. An AI couldn't have conceptualized the idea of a derivative with what was known in mathematics before Newton came around. Pushing the boundaries requires creativity, which AI lacks.

>> No.15337358

>>15337316
Even if we disregard the creative aspects and focus merely on the well-defined problem of proving a theorem, this is NP-complete so unless these AIs are leveraging on some hypothetical breakthrough in SAT-solving technology (protip: they aren't), they won't be able to compete with human mathematicians unless they can first prove that P=NP (protip: they won't).

>> No.15337368
File: 1.15 MB, 1239x1758, mathematics is not worthwhile.png [View same] [iqdb] [saucenao] [google]
15337368

>> No.15337380

>>15337358
AI can produce code, albeit problematic code from text. There's nothing stopping it from proving a theorem if it had the right training materials. Learning how to prove something that's already been proved is something it can easily learn.

>> No.15337388

>>15337316
Drop out, go into politics.

>> No.15337744

>>15337316
yes

"Yes, there is still much more room for math research. While AI can perform some tasks that humans still need, there is still plenty of room for human creativity and rigorous proof in the field of mathematics. The exploration of new mathematical concepts and theories is a never-ending process, and there are always new discoveries waiting to be made. Additionally, while AI can assist with some mathematical computations, it is still up to humans to design algorithms, interpret results, and formulate new hypotheses.

>> No.15337845

>>15337316
same problems here but I have been romantisizing either a PhD in physics or math

>> No.15338098

>>15337316
Pure mathematics in the proof conjuring sense will be one of the last things to be conquered by AI. If you knew anything about either you'd realize that. Do you really expect dressed-up applied statistics coupled with large amounts of data to be able to solve the Riemann hypothesis anytime soon?

>> No.15338309

>>15337316
Research position - where? Academia is now rotten to the core and is not a viable career.

>> No.15338387

of course AI will eventually be doing research level math, only midwits think otherwise. but that doesnt mean mathematicians will be replaced, and it doesnt mean you cannot take advantage of the symbiosis between AI and math.

>> No.15338411

pure mathematician is a stupid idea without AI. you'd have to be one of a million to be able to graduate and also find a job that would pay your bills. it is not realistic. it is like dreaming about making into pro sports and become a world champion.

>> No.15339120

>>15338387
Lol midwits are the ones most excited by AI

>> No.15339169

>>15339120
>Lol midwits are the ones most excited by AI
Only midwits say this.

>> No.15339267

>>15337368
based Uncle

>> No.15339477

>>15337316
Kek most of the economy is tied up in healthcare, production of goods, or transportation of goods. You could make a little less money than most high performing STEM careers with several orders of magnitude less work by being a truck driver or tradie.

>> No.15339595

>>15339120
>filtered by AI
retard

>> No.15339746

>>15337316
AI cannot solve pure mathematics because stats and empiricism is a COPE with an inability at deduction.