Comment by xianshou | Hacker News Reader

Comment by xianshou | Hacker News Reader

Parent Full thread

I would agree that summarization is a killer use case, but would actually argue for three:

1. Condense (summarize, synthesize, extract, etc.)

2. Expand (explain, copilot, suggest, etc.)

3. Transform (translation, style transfer, switch modality between text/image/audio, etc.)

All of the major consumer use cases of LLMs at this point boil down to one of these three core operations. To the extent that LLMs allow you to perform "magic", that magic is comprised by a chain of {condense, expand, transform} operations, each of which could be performed in half an hour by a reasonably informed college graduate with access to Google.

Even if chain length maxes out at 3 or 4 for production-grade applications, that solves a frightening number of problems.

I'm an early engineer at a startup that has nothing to do with AI. I use LLMs for 2 and 3 all the time. A lot of times we'll release a new feature and I'll write my own fairly technical summary of it, then I'll ask an LLM to either target it for a specific channel (e.g. create a version of the post targeting execs vs creating one that's focused on users) or condense it down if we're posting a summary onto a social media channel. In all cases I check the LLM's work, but I rarely have to do much.

When I led teams at a big tech, I generally had a human whose job it was to do this kind of copy refinement/editing when my teams released new features. I find the LLM to be just as good, though obviously our reach is much smaller and the consequences of a mistake are much milder given our size. This is the kind of productivity boost I think of that LLMs offer; pre-LLM we would probably contract out a copy editor for this kind of work but now we do it all in-house and don't have to deal with working with a contractor's hours.

Chatting seems very popular too.

I don't understand why this got downvoted, but there's absolutely a huge unserved(or, rather, anti-served) and somewhat-looked-down-upon market here, and it's actually seemingly more women then men when i ran some numbers.

If LLMs can be a cheap and instant therapist (even if they are just somebody who listens) then they could do a massive amount of good in this world.

Reply on news.ycombinator.com