Talk like Paul Dirac – fewer words is better
This is the entire contents of my home dir AGENTS.md file now.
Does it help reduce the slop you have to read for you the same way it feels like it does it for me?
This is the entire contents of my home dir AGENTS.md file now.
Does it help reduce the slop you have to read for you the same way it feels like it does it for me?
To me adding 'tldr' at beginning or end of the prompt whenever I need is enough. Sometimes I want more explanatory responses.
I actually take this advice in real life. I don’t talk unless really needed. There are exceptions like to my kid, but that’s the general principle to non family/friend members.
Thanks. And maybe I am wrong, but I feel that Paul Dirac should be better represented in the weights than Attempto Controlled English, which sounds niche to me - or at least - I hear it for the first time.
Easy for anyone to try out themselves, and my post is only half a joke - it actually works. For me, at least :)
You want people to talk to you the way you listen. So does everyone else, and most of them do not listen the way you do.
this is an instruction for LLM, not for people
LLMs are what you get when you put people through a stack of filters.
Dirac would never make that less vs fewer mistake.
Agreed, thanks! This kind of makes it obvious that I am not a native speaker, but kind of also proves its was not written by an LLM :)
Edit: fixed now: less words -> fewer words
FYI it's a somewhat regional or (in the UK) class stratified grammar distinction.
While nobody would say "fewer sand", "less" is in practice often used by native English speakers for both count and non-count nouns so it's not really wrong so much as it is revealing.
Would Dirac get a job these days?
The reason i believe is good is because with very few tokens you encode two properties, for which Paul Dirac is most famous for.
* a great scientist (which hopefully encourages researching and verifying answers)
* a person who doesn't like to use many words at all
The later hopefully compensates the apparent "output-prose-greediness" most models seem to have. No idea if that "token greedines" (or slop) is a result from the companies business models or inherent for the models and their training :/