684 points prettyblocks | 1 comments | 21 Jan 25 19:39 UTC | HN request time: 0.621s | source

I mean anything in the 0.5B-3B range that's available on Ollama (for example). Have you built any cool tooling that uses these models as part of your work flow?

Show context

flippyhead ◴[21 Jan 25 22:12 UTC] No.42785739[source]▶

>>42784365 (OP) #

I have a tiny device that listens to conversations between two people or more and constantly tries to declare a "winner"

replies(14): >>42785781 #>>42785791 #>>42785949 #>>42785970 #>>42785979 #>>42786455 #>>42786672 #>>42787108 #>>42788174 #>>42788937 #>>42789840 #>>42791711 #>>42807514 #>>42890452 #

1. eddd-ddde ◴[21 Jan 25 23:41 UTC] No.42786672[source]▶

>>42785739 #

I love that there's not even a vague idea of the winner "metric" in your explanation. Like it's just, _the_ winner.

↑

Ask HN: Is anyone doing anything cool with tiny language models?