5 ms·
"After extensive trial-and-error..." IMO, this is the difference between building deterministic software and non-deterministic software (like an AI agent). It
by mattas 1y ago
"After extensive trial-and-error..."
IMO, this is the difference between building deterministic software and non-deterministic software (like an AI agent). It often boils down to randomly making tweaks and evaluating the outcome of those tweaks.
- s1mplicissimus 1y agoAfaik alchemists had a more reliable method than ... whatever this state of affairs is ^^
- snapcaster 1y agoYou're saying alchemy is better than the scientific method?
- AndrewKemendo 1y agoOtherwise known as science 1:Observation 2:Hypothesis 3:test 4:GOTO:1 This is every thing ever built ever What is the problem exactly?
- wrs 1y agoFor one thing, what you learned can stop working when you switch to a new model, or just a newer version of the “same” model.
- AndrewKemendo 1y agoAll that means is that you verified the null hypothesis which should be that it doesn’t work If you create hypothesis tests that are not written in or specific enough then you’re right you’re not gonna be able to do science Incidentally 99.9% of people I know have no instinct for how to actually do science or have rigor or focus to actually do it in a way that is usable
- neuronic 1y agoThat's because there is no intelligence or understanding involved. They are just trying to brute force a tool for a different purpose into their use case because marketing can't stop overselling AI.