If LLMs are so good at math, why do even the frontier models suck when I try using them for math? Are they machanical turking this shit? Are all the recents proofs just 5 mathematicians with an NDA to build hype before the IPO? If not, why does vibemath suck when I am the one renting the natural language tool?
Please describe your workflow/pipeline.
>>17048868What the fuck are you using? The free google search Ai?Everything's been running wolfram since 2025
>>17048868Don't use their models for math, your work is being stolen, retard. Use local models or nothing.
>>17048868they both hire fleets of mathematicians to guide the models and just bruteforce shit with thousands of agents until something sticks
Because you suck at math.
>>17048868>certified coal burner build
Why do you start sentences with "If", and then proceed tacitly assuming your conditional clause is true?