This is a cool way to put it, but I think even just errors and randomness in reproduction of source ideas sometimes can count as original ideas. Nevertheless, I also think it doesn’t fully encompass all range of mechanisms by which humans come up with original ideas.
There are studies that look deeper into model collapse in various AI models and demonstrate some interesting results. In this one, there is an example of how model degrades over generations. I think the model presented here is the one used at Meta, but with a smaller number of parameters.
Example 1. Example of text outputs of an OPT-125m model affected by model collapse—models degrade over generations, for which each new generation is trained on data produced by the previous generation.
• Input: some started before 1360 — was typically accomplished by a master mason and a small team of itinerant masons, supplemented by local parish labourers, according to Poyntz Wright. But other authors reject this model, suggesting instead that leading architects designed the parish church towers based on early examples of Perpendicular.
• Outputs:
When AI consumes only AI generated content, model collapse happens. That’s why it’s a big mystery and intrigue to me what’s gonna happen to human beings.
Nothing is truly random
Modern physics says otherwise. Einstein also thought exactly like that with his “hidden variables” theory which was later disproven.
Edit: I was interested to read some relevant discussions and here’s some links with quotes
https://philosophy.stackexchange.com/questions/29364/does-true-randomness-actually-exist
It seems very likely that every deviation from perfect homogeneity and isotropy in the universe is due to amplified quantum fluctuations. (That’s true in inflationary cosmology, and I’d expect it to be true in practically any alternative to it.) For example, the shape of Earth’s land masses was probably determined by quantum fluctuations, and has had an enormous influence on human history.
Quantum fluctuations is basically true randomness on quantum level.
The randomness is largely canceled out, except in the case of unstable systems which magnify the effects of any perturbations, no matter how small.
If you go back in time, you are just going to do what you already did because that is in the past.
Only if the Universe is deterministic. If not, random rolls having different outcomes may completely change the course of events and decisions made by people.
Edit: I see I’m being downvoted, so I’ll explain further, if the Universe is deterministic means everything will be the same any time you relive the same time segment, if not, it means even the weather can be different due to aggregation of butterfly effect of different random outcomes in the Universe, and weather being different is already big enough change to be able to influence decisions and course of events. And I’m not meaning weather in the exact same spot you time-traveled to. Even if you restored the exact same state of Universe at some snapshot, if the Universe isn’t deterministic, various random events happening after that point in time can have different outcomes which will aggregate and lead to even more different outcomes in future. Weather might be different the next day and because of that you decided to hide from rain in cafe and met someone there which can completely change your life.
I agree with this in terms of process, but not necessarily agree in terms of result. If you enumerate the state space of target domain, you might realize that all the constructions there can be achieved by randomly introducing errors or modifications to finite set of predefined constructions. Most AI models don’t really work like this from what I know (they don’t try to randomize inference or introduce errors on purpose), otherwise they could probably evade model collapse. But I don’t see why they can’t work like this. Humans do often work like this though. A lot of new genres and styles appear when people simply do something inspired by something else, but fail to reproduce it accurately, and when evaluating it they realize they like how it turned out and continue doing that thing and it evolves further by slight mutations. I’m not saying I want AI to do this, or that I like AI or anything, I’m just saying I think this is a real possibility.