← Back to context Comment by cratermoon 8 months ago LLMs can't do math. 9 comments cratermoon Reply 3np 8 months ago This. People need to manage their expectations. ianks 8 months ago LLMs and tempered expectations, like oil and water Spivak 8 months ago We're giving them calculators though, surely Google could provide a limited set of tools given Search already has a fairly sophisticated calculator.I've been having my AI stuff successfully do math since early gp3 days with this method— even before "tool calling." ckemere 8 months ago Why offer a solution then? Seems fairly easy for google to avoid giving the final number? cratermoon 8 months ago That's not how LLMs work. These systems take a stream of tokens, do some linear algebra to find a stream of tokens within the parameters of their vector similarity, and spit out the result. jbs789 8 months ago Expectations can only be managed by someone who has sufficient understanding - in this case Google by not providing the result. scarface_74 8 months ago LLMs can’t do math. But that’s a solved problem. ChatGPT has had a built in Python runtime that can do math for years - at least the paid version. cratermoon 8 months ago Oh boy, python but now it costs $10,000 per formula.Why? scarface_74 8 months ago It’s $20/month for ChatGPT+.
3np 8 months ago This. People need to manage their expectations. ianks 8 months ago LLMs and tempered expectations, like oil and water Spivak 8 months ago We're giving them calculators though, surely Google could provide a limited set of tools given Search already has a fairly sophisticated calculator.I've been having my AI stuff successfully do math since early gp3 days with this method— even before "tool calling." ckemere 8 months ago Why offer a solution then? Seems fairly easy for google to avoid giving the final number? cratermoon 8 months ago That's not how LLMs work. These systems take a stream of tokens, do some linear algebra to find a stream of tokens within the parameters of their vector similarity, and spit out the result. jbs789 8 months ago Expectations can only be managed by someone who has sufficient understanding - in this case Google by not providing the result.
Spivak 8 months ago We're giving them calculators though, surely Google could provide a limited set of tools given Search already has a fairly sophisticated calculator.I've been having my AI stuff successfully do math since early gp3 days with this method— even before "tool calling."
ckemere 8 months ago Why offer a solution then? Seems fairly easy for google to avoid giving the final number? cratermoon 8 months ago That's not how LLMs work. These systems take a stream of tokens, do some linear algebra to find a stream of tokens within the parameters of their vector similarity, and spit out the result.
cratermoon 8 months ago That's not how LLMs work. These systems take a stream of tokens, do some linear algebra to find a stream of tokens within the parameters of their vector similarity, and spit out the result.
jbs789 8 months ago Expectations can only be managed by someone who has sufficient understanding - in this case Google by not providing the result.
scarface_74 8 months ago LLMs can’t do math. But that’s a solved problem. ChatGPT has had a built in Python runtime that can do math for years - at least the paid version. cratermoon 8 months ago Oh boy, python but now it costs $10,000 per formula.Why? scarface_74 8 months ago It’s $20/month for ChatGPT+.
cratermoon 8 months ago Oh boy, python but now it costs $10,000 per formula.Why? scarface_74 8 months ago It’s $20/month for ChatGPT+.
This. People need to manage their expectations.
LLMs and tempered expectations, like oil and water
We're giving them calculators though, surely Google could provide a limited set of tools given Search already has a fairly sophisticated calculator.
I've been having my AI stuff successfully do math since early gp3 days with this method— even before "tool calling."
Why offer a solution then? Seems fairly easy for google to avoid giving the final number?
That's not how LLMs work. These systems take a stream of tokens, do some linear algebra to find a stream of tokens within the parameters of their vector similarity, and spit out the result.
Expectations can only be managed by someone who has sufficient understanding - in this case Google by not providing the result.
LLMs can’t do math. But that’s a solved problem. ChatGPT has had a built in Python runtime that can do math for years - at least the paid version.
Oh boy, python but now it costs $10,000 per formula.
Why?
It’s $20/month for ChatGPT+.