News for hackers. Stuff that matters.

A comment on A guidance language for controlling LLMs. Parent comment

Re:A guidance language for controlling LLMs by chaxor on Wednesday May 17, 2023 @12:02AM (#35971104)
It's rigor applied where we don't need it, and ignores where we do (mathematical proofs and NN theory, architecture, hyper parameters, training schemes, etc).

I have a somewhat irrational hatred towards almost all of the prompt oriented stuff being thrown about recently. There are a few (very few) input related training schemes that are interesting, but quite a bit of the "proompt-physicians" are just heralding the idea of essentially 'concise and effective communication' as 'I'm a ML expert now' ... which is annoying.

[ Parent | on HN ]
Re:A guidance language for controlling LLMs by Der_Einzige on Wednesday May 17, 2023 @12:11AM (#35971160)
Why would you dislike actual prompt engineering? This isn't some grifter trying to claim they're an expert because they wrote a cool prompt, this is a full fledged structured templating system for LLMs from an excellent author whose done a ton of other ML work.

I think you should attack actual grifters instead of an excellent project.

[ Parent | on HN ]