Half a century worth of reporters’ notes might be some valuable training data.

> The AI company asked Judge Sidney H. Stein of the US District Court for the Southern District of New York to step in and compel the Times to produce reporters’ notes, interview memos, and other materials for each of the roughly 10 million contested articles the publication alleges were illegally plugged into the company’s AI models. OpenAI said it needs the material to suss out the copyrightability of the articles. The Times quickly fired back, calling the request absurd.

Can any lawyer on here defend OpenAI's request? Or is the article not characterizing it well in the quote?

> Encyclopedias are copyrightable. Phone books are not.

It depends on the jurisdiction. The US Supreme Court ruled that phone books are not copyrightable in the 1991 case Feist Publications, Inc., v. Rural Telephone Service Co.. However, that is not the law in the UK, which generally follows the 1900 House of Lords decision Walter v Lane that found that mere "sweat of the brow" is enough to establish copyright – that case upheld a publisher's copyright on a book of speeches by politicians, purely on the grounds of the human effort involved in transcribing them.

Furthermore, under its 1996 Database Directive, the EU introduced the sui generis database right, which is a legally distinct form of intellectual property from copyright, but with many of the same features, protecting mere aggregations of information, including phone directories. The UK has retained this after Brexit. However, EU directives give member states discretion over the precise legal mechanism of their implementation, and the UK used that discretion to make database rights a subset of copyright – so, while in EU law they are a technically distinct type of IP from copyright, under UK law they are an application of copyright. EU law only requires database rights to have a term of 15 years.

Do not be surprised if in the next couple of years the EU comes out with a "AI Model Weights Directive" establishing a "sui generis AI model weights right". And I'm sure US Congress will be interested in following suit. I expect OpenAI / Meta / Google / Microsoft / etc will be lobbying for them to do so.

Encyclopedias may be collections of facts, but the writing is generally creative. Phone books are literally just facts. AI models are literally just facts.

I will address both points:

1. No one is training on users' bank details, but if you're training on the whole Internet, it's hard to be sure if you've filtered out all PII, or even who is in there.

2. This isn't happening because no one has time for more time-wasting lawsuits.

??

Doesn't that mean if they used data created by, (or even the data of), anyone in the EU, that they would want to not release that model in the EU?

This sounds like "if an EU citizen created, or has data referenced, in any piece of the data you trained from then..."

Which, I mean, I can kind of see why US and Chinese companies prefer to just not release their models in the EU. How could a company ever make a guarantee satisfying those requirements? It would take a massive filtering effort.

If that meaning is "intuitive", why was it not used before the Open Source Initiative introduced their definition? The competing uses are the ones co-opting an existing phrase.

You have to realize that as an individual, you have no power anyways

It doesn't matter if an individual personally has access to ML models, because government and/or huge corporations will ensure that individuals cannot use them for anything that would threaten government or corporate interests

This unfettered explosion of ML growth is disempowering all of us. Those with power are not using these tools to augment us, they are hoping to replace us.

Can you name some of these extremely powerful resources? I’m fine without access to AI hallucinations and poorly made images with six fingers.

For decode steps it depends on the number of inputs you run at a time. If your batch size is 1 then it runs in line with active params, then as you get to like batch size 8 it runs in line with all params, then as you increase to 128ish it runs like the active params again.

For the context encode it’s always close to as fast as a model with a similar number of active params.

For running on your own the issue is going to be fitting all the params on your gpu. If you’re loading off disk anyways this will be faster but if this forces you to put stuff on disk it will be much slower.

An H100 supports 80 GB of memory. so at FP8 that would allow 3 of the 16+1 models per GPU (assuming around 26B per model), requiring 9 H100s, that usually would not fit one chassis I guess.

Once you have something with 192 GB it gets interesting. You could probably have 7 at FP8 per GPU. At FP16 it probably only would fit 3 per card, requiring 9 again.

I'd say for the current memory layout of cards they missed a little bit the sweet spot. With slightly smaller models or one expert less one should be able to run it on 8 H100s at FP8 or 2 B100s at FP8 or even on 4 B100s at FP16 if I calculated correctly.

How many cores do your CPUs have? Are you using the 64 core EPYC 9334 mentioned in the linked page? Do that many cores provide a speedup versus having fewer cores?

Interesting, it seems to have the “correct” answers to all those issues. I wonder if this is mostly just a US based model as a base.