chemoelectric

chemoelectric, 15 hours ago

I am looking to decorate the bathtub area very tastefully and low-keyed with epoxy sea animals.

chemoelectric, 15 hours ago

I used SigWong epoxy because I have a bunch to use up. I have switched to Yieho epoxy as my choice for general purpose resin that doesn't brag it is "industrial" and made from recycled battleships.

chemoelectric, 22 hours ago

There were some bubbles. Not extremely many. I used Let's Resin "Magic Pour" plaster, which I am trying to use up, because I am not exactly fond of it. I colored it with a mica pigment that seems to turn the stuff pink no matter what color you use! In this case it was "copper".

Better to use squeezes of acrylic paint to color the plaster, I guess. That works just fine.

chemoelectric, 1 day ago

@troed @axbom

Notice how what it does is GENERATE A LOT OF GRAMMAR. :) A true intelligence would think in pictures. But the people who write this software think the way the software does, judging from a few chats I've had.

chemoelectric, 1 day ago

@troed @axbom Cats obviously can figure out the part about being "in" something, and think in pictures, proprioceptions, etc. So these things are absolutely necessary for AI actually to be successful. These facts are too simple for Silicon Valley to comprehend.

chemoelectric, 23 hours ago

@troed @axbom

Somebody once asked one of these what the probability was that six was the last digit of pi. It came up with all sorts of responses, TENDING TO BE VERY, VERY WORDY. That's what I find most interesting. All of them incorrect. I didn't see humans give a correct answer, either.

I solved it quickly, but do a lot of math. The answer is zero. Since there is no last digit, there is zero chance six is it. But this is something I solved with mental pictures and only state with words.

chemoelectric, 23 hours ago (edited 23 hours ago)

@troed @axbom Describing a scene is generating grammar! This is not something a CAT would even do! Would I want THAT to make a decision for me?

chemoelectric, 23 hours ago

@troed @axbom Part of the problem here is that advocates of the various AI go around TELLING other people things, rather than TAKING AN INTEREST in what others have to say.

chemoelectric, 23 hours ago

@troed @axbom If the AI hasn't even reached the level of a cat first, then how can it be plausibly claimed it has reached the level of a human? It is some kind of defective being spewing short stories.

chemoelectric, 23 hours ago

@Aileme @troed @axbom

I include "thinking in proprioceptions" as kind of the same thing. Thinking with body positions and motions, etc

chemoelectric, 23 hours ago

@Aileme @troed @axbom

Perhaps the more important thing is why does anyone waste their time trying to make me content with LLMs. For that really IS what they are doing, when I am saying, no, you are not paying attention to the world outside enough. You are entrapped in the model too much.

I say folk are developing a stunted view of the human brain, made to match the capabilities of LLMs.

Let's keep in mind that brains cannot possibly be digital devices, despite the delusions of many.

chemoelectric, 23 hours ago

@Aileme @troed @axbom

Neurons fire in pulse chains. They are sending neither parallel nor serial digital signals, but rather analog ones.

So whatever is going on is something that a digital computer will have to simulate mathematically, and probably with a lot of what amounts to geometry (since practically all continuous math ends up being geometry).

chemoelectric, 23 hours ago

@Aileme @troed @axbom

What was supposed to impress me about the quotation generated by the LLM above was what? THE ELEGANCE OF ITS ENGLISH. So really, inadvertently, the quotation made MY point.

This is not real smarts, like a cat. This is flinging words around. Verbal puzzles.

chemoelectric, 23 hours ago

@Aileme @troed @axbom

Are you able to eat a piece of food that you see?

chemoelectric, 23 hours ago

@Aileme @troed @axbom I ask because you may be under the misapprehension that by intelligence I mean only that of which you are "consciously aware". This is actually a trivial characteristic of intelligence. It isn't even necessary. It is also disruptible by brain injury, although I think such injuries are rare.

A cat can decide to eat a piece of kibble. It cannot think in sentences at all. It is not possible to live by thinking only in sentences, unless kept alive as a vegetable.

chemoelectric, 22 hours ago

@Aileme @troed @axbom A good book to read (all the AI folk might consider it) is "Zen in the Art of Archery" by Eugen Herrigel.

Herrigel struggles to understand how one can shoot an arrow perfectly accurately without aiming it. He never arrives at an answer.

I reproduced the phenomenon, though using different apparatus than bow and arrow. I struggled but solved it. One aims, but unconsciously. This is how come I know conscious awareness is a trivial characteristic of intelligence.

chemoelectric, 22 hours ago

@Aileme @troed @axbom

I actually did experimental science, in the 1990s.