There is and always will be […] fancy ass business rules behind it all.
Not if you run your own open-source LLM locally!
Comment on 95% of Companies See ‘Zero Return’ on $30 Billion Generative AI Spend, MIT Report Finds
callouscomic@lemmy.zip 22 hours agoGo learn simple regression analysis. Then you’ll understand why it’s simply a prediction machine. It’s guessing probabilities for what the next character or word is.
Also simply the training of these models has already done the energy damage.
explodicle@sh.itjust.works 9 hours ago
Knock_Knock_Lemmy_In@lemmy.world 13 hours ago
AI is interpolating data. It’s not great at extrapolation. That’s why it struggles with things outside its training set.
fuck_u_spez_in_particular@lemmy.world 10 hours ago
I’d still call it extrapolation, it creates new stuff, based on previous data. Is it novel (like science) and creative? Nah, but it’s new. Otherwise I couldn’t give it simple stuff and let it extend it.
Knock_Knock_Lemmy_In@lemmy.world 4 hours ago
We are using the word extend in different ways.
It’s like statistics. If you have extreme data points A and B then the algorithm is great at generating new values between known data. Ask it for new values outside of {A,B}, to extend into the unknown, and it falls over (usually). True in both traditional statistics and machine learning