The level of details these techniques are giving now... this is very impressive.

Interesting analysis around the current situation around web scraping and intellectual property. This moved to being mostly dealt with using contract law which makes it a terrible minefield. Lots of hypocrisy all around too which doesn't help. GPT and the likes will likely be the next area where cases will rise.

Now this could turn out to be interesting. To be confirmed when this get closer to production (if it does), especially on the power consumption side. This will be the important factor to make this viable I think.

Good explanations around the (deserved) complaints against Zoom and their not that new user license.

Excellent piece from an excellent artist. I really thought this through and I think he's going in the right direction.

Very comprehensive (didn't read it all yet) guide about self-supervised learning. It'll likely become good reference material.

It smells a bit like hypocrisy isn't it? On one hand they claim it can make developers more productive on the other they think they shouldn't use it.

Oh the bad feedback loop this introduces... this clearly poison the well of AI training when it goes through such platforms.

Looks like an interesting tool to run LLMs on your own hardware.

Maybe it's time to make so called "reinforcement learning from human feedback" actually humane? It's not the first account along those lines in the industry.

Interesting research turning to genetic algorithms to optimize bytecode handler dispatchers.

This is looking like a bad move. Clearly the fault of western countries though which let things unfold ambiguously regarding copyright... Now Japan is weakening copyright for everyone.

Very good interview. She really point out the main issues. Quite a lot of the current debate is poisoned by simplistic extrapolations based on sci-fi. This distracts everyone from the very real and present problems.

Looks like a promising way to reduce the training cost of large language models.

OK, this is a pre-print so to take with a truckload of salt. If further nice results get built up on this it could turn out interesting though. This is a much more intellectually satisfying approach than the current arm race of "let's throw bigger models at the problem". This has the potentially of reducing the computational complexity of those models, this is definitely welcome in term of energy and hardware requirements. Let's wait and see...

This was only a matter of time. It'll be interesting to see how this will unfold. Potentially it could turn into lawsuit cases being built up, it could also mean content producers get a cut down the line... of course could be both. Since FOSS code also ends up in training those models I'm even wondering if that could lead to money going back to the authors. We'll see where that goes.

This is impressive results. Clearly much less artifacts than on previous such models.

This is important. We need truly open generator models. This can't be left in the hands of a few with only API access, especially since they lack basic transparency.

Clearly aims to demonstrate the superiority of their specialized hardware for training. That said it's nice to have proper open models available (architecture, training data, weights... it's all in the open).

Now, this starts to become interesting. This is a first example of trying to plug symbolic and sub-symbolic approaches together in the wild. This highlights some limitations of this particular (quite a bit rough) approach, we'll see how far that can go before another finer approach is needed.