My thoughts exactly, but they also suck up more than the commons, according to our intellectual property laws. I was glad when the courts decided that LLM products were not copyright-able works. A commons for the models also means much less redundant training infrastructure, i.e. fewer data centers, and ability to reign in market externalities, like burning fossil fuels to power them.

I am not here to defend that institution at all.

The pope’s AI encyclical was good reading and reminds us to center human dignity, compassion, etc. Lots of ethics to explore with regard to AI and related cultural phenomena.

Close the lid, buddy. You gave your only good ideas to a stranger and they made you obsolete.

Thanks! I suspected as much, but lack adequate expertise to comment.

That is cheap enough to be an expendable decoy, I expect.

They could add tremendous focal points for debate / discussion, but I would rather have them own only their topics than own the whole social graph. I would not trust any org to own the social graph, so yes Fedi. I would really like govs to secure my digital rights via zero knowledge proofs of things like citizenship, age, etc… Give me a way to reveal only what a particular service needs to know.

This one helped me too, especially on the special user/group mappings for lsio containers. That’s a trick.