X changes its terms to bar training of AI models using its content

That way he can continue to steal from others and lock competitors out whilst being comfortable knowing that no laws will be enacted to prevent it.

It means that assuming training AI models is fair use (if it wasn't AI companies including xAI would be in trouble), they can't really stop you.

But now, essentially, they are telling you that they can block your account or IP address if you do. Which I believe they can for basically any reason anyways.

https://github.com/google-deepmind/pg19

That gives us a model that's 100% open and reproducible with low, legal risk. It would also be a nice test of how much AI's generalize from or repeat behavior in their pretraining data.

Then, a new model using that, The Stack, and FreeLaw's stuff (by paying them to open source it). No Github Issues or anything with questionable licenses or terms of service violations. That could be the next baseline for lawful models with coding ability, too. Research in coding AI's might use it.

You must not, and must not allow those acting on your behalf to:

...use the Data APIs to encourage or promote illegal activity or violation of third party rights (including using User Content to train a machine learning or AI model without the express permission of rightsholders in the applicable User Content);

> The concept of copyright first developed in England. In reaction to the printing of "scandalous books and pamphlets", the English Parliament passed the Licensing of the Press Act 1662,[16] which required all intended publications to be registered with the government-approved Stationers' Company, giving the Stationers the right to regulate what material could be printed.[20]

> The Statute of Anne, enacted in 1710 in England and Scotland, provided the first legislation to protect copyrights (but not authors' rights)

It’s not been tested in court though

Without search engines what the point in posting it on open net if nobody can find.

Even if this is done, the case of starving artist v. megacorp will probably go to whoever wields the most money and lawyers. To add insult to injury, the artist’s opponent is fueled by their ill-gotten gains.

> We need regulations to limit the power grabs. Train all you like, but don't dare try to constrain to your walled gardens.

No, no one should train, period.

Then maybe the recording companies will start defending artist rights.

Because not sure what all the other industry bodies are doing.

GPT-3 was trained on approximately 300 billion tokens. An small sized technical textbook might contain something like... 130,000 tokens? (1 token ~= 0.75 words, ~100k words in the book).

Thus, say you wrote a textbook on quantum mechanics that was included in the training corpus. A naive computation of the fraction of your textbook's contribution to the total number of training tokens would be 300B/130K = 0.0000004333333333, or 0.000043%.

If our hypothetical AI company here reported, say $500M in yearly profit, if all of that was distributed 100% based on our naive training token ratio (notice I say naive because it isn't as simple to say that every training token contributes equally to the final weights of a model. That is part of the magic.) then $500M * 0.000043% = $215.

You could imagine a simpler world where it was required by law that any such profitable company redistribute, say, %20 (taking the 'anti-VAT' idea) back to the copyright holders / originators of the training tokens. So, our fictitious QM textbook author would receive a check in the mail for $43 for that year of $500M in revenue. Not great, but not zero.

Since then, training corpuses are much, much larger, and most people's contributions would be much smaller. Someone who writes witty tweets? Maybe 1/100th the length of our above example in am model with now 100x the training corpus.

So fractions of a penny for your tweets. Maybe that is fitting after all...

Rather, we are back to that age of state- (now corporate-) backed privateering.

Any country that seriously implemented this would just end up being completely dominated by the autonomous robot soldiers of another country that didn't, because it effectively bans the development of embodied AGI (which can learn live from seeing/reading something, like a human can).