Something that always comes to mind when I see Thiel's name mentioned is a talk he gave--"Competition is for Losers". Pretty much sums up exactly why and how he and other oligarchs are driving free market capitalism off a cliff and into the abyss, taking all of us with it.

I would have vastly preferred to just have been wrong but here we are.

This is a CLASSIC legal problem. Copyright law was written with different scenarios in mind, and applying old rules to a novel situation can be extremely complex and have different results than if the same people/interests/morals who created the old law designed new rules for the new situation.

The interesting fair use analysis starts with that training. Is it analogous to a human reading and understanding a book? When Claude the generates responses to questions, how is this different from that a human? If it's OK for humans, why not LLMs? (I'm identifying questions, not implying answers)

Ripping an mp3 and playing it for others are two separate actions--format shifting is only about the first. Nothing like playing an mp3 for others happens until Claude responds to a user. Format shifting is about what happens between buying a paper book and LLM training.

FWIW, that finding on format shifting is not surprising. Think of it as similar to an individual making mp3s from an album they already own.

It's the actual ingestion into an LLM that's an interesting fair use issue. Consider w/ the four fair-use factors: en.wikipedia.org/wiki/Fair_us...

We have reached the point in our dystopia of stupid that there will soon be a black market for vaccines 🤣🤬

That'd be great, but problem is also only partly generational in the first place. Pew says Gen X were his biggest supporters, but younger voters shifted that way too. Without persuasion, the cult will be able to retain power over at least their own party for the foreseeable future. 😒

I worry even that is optimistic.

IMHO, the only way the Trump cult re-learns that unrestrained ignorance, cruelty, and greed is a bad strategy for life is if the consequences are severe. Conversely, if the damage they could cause can be mitigated, few will be willing to change their outlook.

Where is Spock when you need him? 🤣

The brown one looks like it could be an extremely disappointing brownie 🤣

And further that, since this is already much of how humans obtain our own "intelligence", once LLMs have the capacity for deliberative modification of their own mind state, based on their own experiences and interaction with the world, this line starts to blur.

My intuition (such as it is) is that LLMs represent an amalgamated gestalt of human thought present in their training data, so the question isn't whether or not the "lament" exists (it does) but to whom it should be attributed--the people who produced the training data or the LLM.

Excellent work, as always

I’ve never understood the hate. Looks beautiful to me, especially when combined with some green plants.