Notices by Tanguy ⧓ Herrmann (dolanor@hachyderm.io)

@thomasfuchs what appals me is that some FLOSS people go full in with LLM.

I can see the: we have all the knowledge of the world in a chat.
It's like a wikipedia fever dream.
And there is the technical achievement, which, errors in the models apart and error of anthropomizing the tech apart, is still incredible.

BUT, we have to depend on giant assholes for this. And SOMEHOW, they can't see that it is extremely binding, as binding as hardware companies that were selling proprietary software in the days of yore that created the FLOSS movement.
No frontier model is sharing their whole pipeline, from stealing the data from the whole world, to the training algorithm, to the data they amassed from other places to the weight and all.

But let's go, let's use this technology because with it, we can create more FLOSS.

REALLY?

@dalias @cstross @cavyherd @Richard_Littler I'm interested in the answer for that, too.

@phiofx @orsinium @avlap2 I don't know, it seems pretty explicit
Opt in by default, and a lot of pros and cons are written in his blog post.

@aral well, X isn't so far from Z neither, so I'm sure he can get a scale price reduction with the Russian uniforms as well.