And then you use AI to summarize those 50 pages :)

Could it be also a toggle to skip/not show any AI-generated content? And all child branches?

That might take me another decade.

I'm joking, but we've always resisted partitioning HN. Here a bunch of past explanations about that: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

I do sort of like the idea (suggested by mthurman) that we let users prompt HN to be the kind of HN they want. That could be the ultimate dump of long-requested features (dark mode! tags! blocklists!)

When I opened the link the title was adjusted and showed 25%..

Thanks for sharing! Good old README.md in the repo would not hurt :)

There is "MESSAGE IN A BOTTLE.txt".

The first game here is to understand what the code does without a high-level README.

I was thinking the "MESSAGE IN A BOTTLE.txt" is rather special than a proper README :)

Long live the new Flash! :)

> One thing I do know for sure. LLMs would have been impossible.

Maybe they could, as ASICs in some laboratories :)

Honestly, I think we could’ve pulled off a lot earlier if GPU development had invested in GPGPU earlier.

I can see it now… the a national lab can run ImageNet, but it takes so many nodes with unobtanium 3dfx stuff that you have to wait 24 hours for a run to be scheduled and completed.

Are there any nice TTS models? It could be handy to have them as well.

Check out "FUTO" keyboard Best on device STT I've ever used

sure, I'll take a look at it. Thank you for the recommendation

I don't really do TTS right now, but i allow for STT using whisper. TTS is part of the backlog and I should be adding it soon

I wish they published it in a more organized manner.. At the moment all published episodes are mixed with other content of their channel.

The cyberpunk we deserved :)