I recently wanted to write a post about a situation in a SNAFU/FUBAR condition [probably a Trump thingie], and realized while writing it that I didn't know whether to use the word "chaotic" or "random" and actually couldn't convince myself that I knew the distinction between the two (if any).
So I did what any modern lazy person would do - I asked AI. The reply is embedded above. I have to say I would never have predicted that chaos is governed by precise rules. Unless the AI is wrong. So now I'll throw this out to my readers, some of whom are mathematically or philosophically inclined, and at least one of whom is a copyeditor/poofreader in real life.
Addendum: If this interests you, be sure to read the comment by Codex.
Codex: It's mostly inaccurate AI slop. This is a partial definition of chaos theory the mathematical theory applied to complex dynamic systems used primarily in physics. Biologists have a hell of a time with it.
ReplyDeleteThere is a a strange attractor that organizes chaos somewhat (the Lorenz attractor). It looks like a butterfly. The butterfly effect is most commonly referred to online by the public, when it was used as a simple analogy to introduce the concept.
The two terms are very different within the physics math world.
Random is the outcome of flipping a coin.
Hope this helps. Angry at this AI nonsense that has not been programmed to know the difference between most accurate and most popular.
Very interesting. Had to look up the Lorenz attractor -
Deletehttps://en.wikipedia.org/wiki/Lorenz_system
- which is indeed butterfly-shaped. And interesting to ponder how changes in the size of the parameters can shift the outcomes into chaos. I had thought that the butterfly effect of multiple-log-power changes in effect were the result of time/repetition/multiplication rather than the size of the parameters.
This is too complicated for me to blog with any comfort. I'll leave the post up for a while, if only to show readers what they already know - that relying on AI is foolish.
Codex: Thank you, kind Sir, for the acknowledgment. I came back to clarify and hope that you do write about it. It'll be interesting to see what you come up with.
Delete"Strangd attractor" may be of interest. Butterfly effect is a term used to illustrate that unless all variables and parameters were known (which we still can't measure), it is impossible to calculate and predict the weather.
Fractals illustrate that there is a self-replicating system that brings order and symmetry back. Chaos doesn't necessarily become bigger.
Interesting that AI is claiming that we live in a predetermined universe, which we do not.
John Briggs Fractals is a beautiful book on it.
Codex, is there objective evidence for this denial -
Delete..."AI is claiming that we live in a predetermined universe, which we do not."
Codex: Yes. Quantum physics/quantum mechanics. Physicists are very concerned about LLMs. Wikipedia isn't always accurate on it.
DeleteI would leave this post up.
Codex: Similar debate to 80 years ago is popping up again.
DeleteWell, since we're not living in a predetermined universe, I'll make the decision now to leave the post up. (or was that always going to be the case?.....)
DeletePhysicists are very concerned about LLMs.
DeleteThis triggered me, as a physics-adjacent person. Sam Altman said that having his newest AI is like having a physics grad student in your pocket. You can see here that he's a college drop-out because he never got further than the stereotype that PhDs are very smart people that know a lot.
However, the key to getting a PhD degree is not only that you know a lot, but also that you did some novel research. You studied and studied and studied and then hit the wall of human knowledge. And then you pushed against that and moved it a little. It is that new knowledge that gets rewarded with a PhD.
And to be clear, this is generic for ALL PhDs not just in STEM. Someone who gets a PhD in literature still ADDED some knowledge about literature - a new insight, new context, whatever, not my field. But they add something.
Here's a visual representation.
https://matt.might.net/articles/phd-school-in-pictures/
https://matt.might.net/articles/phd-school-in-pictures/
Anyway, one thing that LLMs will NEVER do, is do something NEW. All they do is guess what the next word is based on whatever has been written on the internet. And since the internet is already half-slop, they're feeding on their own slop. Which is not new.
"... one thing that LLMs will NEVER do, is do something NEW."
DeleteIt would be fun to debate how often HUMAN artists, composers, musicians, and authors ever create something truly NEW vs creating variants of previous stuff. But that would require a couple hours and a couple bottles of wine.
Codex: Ha! Nice one.
ReplyDeleteAuthoritarian governments do not like random factors. The amicable debate between physicists was picked up on by the government and declared J. science with respect to quantum physicists.(I try to avoid certain search words).
Chaos theory is definitely not random. The 3 body problem is chaotic for example. Shuffling a deck of cards is mostly chaotic, not random, if you assume there is no supernatural force influencing which way the cars fall in a riffle.
ReplyDelete(I've always wonder what people that do Tarot cards think. Can you do a Tarot card reading by shuffling the cards before the person whose future is to be read arrives? Do the cards change positions in the deck?).
That said, the minute details of how the cards are released by your finger tips might be influenced by quantum waveform collapse in your nerve cells, but now we are in Schroedinger's Cat territory.
The only thing I learned as a result of getting a PhD in particle physics is that physics is hard and I don't understand it.
Another important distinction is "deterministic" vs "predictable." Just because something is entirely determined by laws of physics (e.g. tides are determined by gravity and fluid dynamics, including atmospheric pressure) doesn't mean we can perfectly predict it even in the very short term of days. Even if we did live in a deterministic universe, that wouldn't necessarily make it predictable.
ReplyDelete