Hacker News new | ask | show | jobs
by andrekandre 171 days ago 

  > But for most tasks, a fleet of tailor-made smaller models being called on by an agent seems like a solidly-precedented (albeit not singularity-triggering) bet.
not an expert by any means, but wouldn't smaller but highly refined models also output more reproducible results?

intuitively it sounds akin to the unix model...
1 comments
nubg 171 days ago
But then again the main selling point of using LLMs as part of some code that solves a certain business need is that you don't have to finetune a usecase-specific model (like in the mid 2010s), you just prompt engineer a bit and it often magically works.
link
andrekandre 170 days ago 

  > you just prompt engineer a bit and it often magically works
yea, i agree, thats the biggest selling point right now

i just get the feeling reproducibility, performance and cost will start to become more and more important as time goes on... jmo tho

link