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.