15) Tokenisation is NP-Complete
Philip Whittington et al
aclanthology.org/2025.acl-lon...

There is a precedent though from before the GenAI hype with algorithmically generated music that was released into the public domain in order to fight back against big labels copyrighting beats
www.vice.com/en/article/m...

I think it’s in their T&C, don’t have the time right now to look, but given the massive copyright infringement that OpenAI is accused of they should just not use this as an argument

It can be done but not directly bsky.app/profile/petr...

Got it. I haven’t tried it locally but I’ve heard that it needs permissions for running arbitrary code though it was on social media so I’m not sure about this either 😄

Any source for this claim?

I guess Github Copilot is worth a try especially because it does autocomplete

And shameless self promotion, a feed I made that is based on hashtags and wordpairs with English language filtering and content moderation in order to check out what’s being discussed on the NLP side of bsky - NLP Newest: bsky.app/profile/did:... (5/5)

Feed that is based on users list: NLP feed ranked: bsky.app/profile/did:... (4/5)

The next 2 feeds can be useful and you can see who posts and interacts there in order to make your list

Paper feed filtered by links to ACL anthology - NLP Papers Test: bsky.app/profile/did:... (3/5)

So, first thing I’ll recommend will be an LLM feed since you are mostly interested in this domain: bsky.app/profile/did:... (2/5)

Welcome to the NLP side of bsky 🤖 If you’re interested in LLMs you may want to check out the entire domain of Natural Language Processing so I’ll just add a few replies with my recommendations. These will be feeds but some of them are based on user lists just how you wanted (1/5)