News for hackers. Stuff that matters.
Re:The Curse of Recursion: Training on Generated Data Makes ... by chongli on Saturday December 07, 2024 @10:24AM (#42350365)
Think about this: The sum total of the human-generated knowledge was derived in a similar manner, with each generation learning from the one before and expanding the pool of knowledge incrementally.

Not true. No amount of such iteration gets you from buffalo cave paintings to particle accelerators.

Humans generate knowledge by acting in the world, not by dwelling on our thoughts. The empiricists won a very long time ago.

Re:The Curse of Recursion: Training on Generated Data Makes ... by brookst on Saturday December 07, 2024 @12:11PM (#42351097)
It’s not binary. Humans generate plenty of knowledge from pure abstract thought.
Re:The Curse of Recursion: Training on Generated Data Makes ... by kerkeslager on Saturday December 07, 2024 @01:21PM (#42351636)
Do they?

When I pursued creative writing in my teens and early 20s, it became clear to me that originality is extremely difficult. I am not entirely sure I have ever had an original thought--every idea I've put to paper thinking it was original, I later realized was a recombination of ideas I had come across somewhere else. The only exceptions I've found were places where I had a fairly unusual experience which I was able to interpret and relate, i.e. a unique interaction with the world.

Perhaps more importantly, LLMs do not contain any mechanism which even attempts to perform pure abstract thought, so even if we accept the questionable assumption that humans can generate ideas ex nihilo, that doesn't mean that LLMs can.

Re:The Curse of Recursion: Training on Generated Data Makes ... by jiggawatts on Saturday December 07, 2024 @04:24PM (#42352959)
Unless your argument is that all creative writing is inspired by God, or some similar "external" source, then clearly a closed system such as "humanity" alone is capable of generating new creative works.
Re:The Curse of Recursion: Training on Generated Data Makes ... by kerkeslager on Saturday December 07, 2024 @07:44PM (#42354077)
Did you even read the post you're responding to?
Re:The Curse of Recursion: Training on Generated Data Makes ... by jiggawatts on Saturday December 07, 2024 @02:35PM (#42352237)
You’re right, we obtained the knowledge externally. It was aliens! I knew it!
Re:The Curse of Recursion: Training on Generated Data Makes ... by chongli on Sunday December 08, 2024 @03:16AM (#42355789)
Externally, yes, we obtain knowledge from the world around us. We’re not brains in vats conjuring knowledge from the void of our isolated minds.