FaceDeer

I’m not making any statements about what other people may think about it. The question was why this guy is doing this thing, and I expect it’s because he finds it fun.

I’ve seen plenty of weird bots on Reddit over the years that had no point to their existence other than, presumably, having been fun to code. This is probably one such.

Right, and this is presumably something he finds fun. You were asking why, I was explaining why.

No, it’s not the same. I was using basketball as an analogy. Someone who doesn’t enjoy basketball wouldn’t “get it”, just as you’re not “getting” the fun that can come from building and playing around with AI bots. Different people find different things to be fun.

/r/SubSimGPT2Interactive/ does this.

I actually wandered away from the SubredditSimulator successor subreddits because even with GPT2 they were “too good”, they lost their charm. Back when SubredditSimulator was still active it was using simple Markov chain based text generators and they produced the most wonderfully bonkers nonsense, that was hilarious. Modern AIs just sound like regular people, and I get that everywhere already.

What I am failing to understand is: why?

People do things for fun sometimes. You could ask this about almost anything that people do that isn’t directly and immediately related to survival. Why do people play basketball? It’s just pointlessly bouncing a ball around in a room, following arbitrary rules that only serve to make the apparent goal of getting it through the hoop harder.

It’s a bit technical, unfortunately, at least the way that I do it. I installed the Whisper and wrote a Python script that uses it to create transcripts of audio recordings, storing them in JSON files. Then when that’s done I fire up KoboldCPP with the Command-R model, with a context set large enough to hold roughly two hours of transcript - I rarely have anything longer than that to process. I’ve got an RTX 4090 with plenty of RAM and VRAM so I can handle a fairly large model. Then I run a second Python script that uses KoboldCPP’s API to process the transcripts in various ways - getting a paragraph-long summary, getting a list of action items, and so forth.

I wouldn’t be surprised if there aren’t better tools out there now specifically for this kind of task, this is just what I cobbled together over the years as these things came out. So might be worth looking around for something like that.

For the glory of the Empire, of course.

Set a local large language model loose on transcripts of some work meetings to generate summaries and lists of action points. For the glory of the empire!

It’s -25C where I am right now. Bring on 2034.

Heaven forbid an open protocol see adoption.

Except it’s not denying service, so it’s just a D.

Thanks, it was very useful to know that.

Sometimes seems like the people who talk about Musk the most are the people who would like to not hear people talking about Musk.