A.I.

While I understand the impulse, I believe term limits would have little effect on the situation. The bureaucrats, NGOs and plutocrats who actually pull the string here would have no such term limits. With Congressional term limits, they would probably shear those new sheep coming into Congress even more closely. 

The fundamental problem is not one of term lengths. It is one of an overbearing and out-of-control federal government, which has accrued to itself dictatorial powers. The only hope for the country is to shrink this Leviathan in any way possible. 

I have to admit. The article got to the heart of it when it said "boys will be boys." I wouldn't have done this in high school. But I can see how effortless and easy it would be to do it. And, well...boys will be boys. Especially high school boys.

But now wait until they start locking high schools boys up as "sexual predators" for it. That's the real danger.

And are the girls really shocked or are they flattered because the bodies their faces are pasted onto via AI are solid 10's? I don't know. But all that is in play. Many will doth protest too much.

And some of the boys just need a good ass-kicking from their father. And if a father isn't at home, well, we can surely trust mom to lay down the law. Right? Right?

Tip of the iceberg. If feminism continues to marginalize boys, AI in the form of physical robots are only a decade or so (if that) from being able to replace them. And that may seem an inappropriate statement in the context of nudie AI girls. But consider that girls are now being hyper-sexualized (many or most dressing like hookers), and yet the boys are supposed to somehow remain completely unphased by this state of affairs.

While researching this subject, I found an AI of Taylor Swift in what I guess you'd call her AI digital birthday suit. Not bad. I can't post it here but if you Google for "Taylor Swift AI nude" and turn off the safe search filters, you will find a couple. More than a couple. Excuse me while I reinsert by eyeballs inside their sockets.

I thought that article by William Dembski was very good. It's nice to read a piece that uses careful and descriptive language to describe something that is somewhat complex, at least philosophically. I liked what he terms "prissiness":



Having bias is unavoidable. Pretending it isn't is dishonest. As long as these LLM's are wielded by tech giants, they will be no more ideologically neutral or honest than the MSM. He holds out hope that the LLM that will be deployed by Musk will be better:



I think his central point is solid. And that is that "...LLMs can have no such experience [of the real world]. They consist of a neural network that assigns weights to relations among words and sentences."

That is, much like a postmodern or "intellectual," they parse language, and can be very good at it. With both AI and an "intellectual," that language stays unconnected from the ground of reality. It's exists referencing only itself. But at the same time, the more astute among us can often spot that it is all gibberish meant to bamboozle, not enlighten. But most do not, will not, or cannot.



My brief experience with ChatGPT has left me not desiring to use it at all. I mean, it has its uses, so I frequently read. But it just seems inherently fake, phony, and untrustworthy. Worse than that, it has an air of flighty unreality to it.

As Dembsky says, for finding the capital of North Dakota, it can work. But for larger subjects, it feels no more useful than sitting down and talking to Rachel Maddow or John Stewart. It's just another purveyor of left wing baloney. What is called a "hallucination" in AI we call a "delusion" in an MSNBC talking head. In fact, the concordance between Biden and the occasional "hallucination" from an LLM is apt.

The upside of the LLMs is that (as Dembsky noted) their bias is potentially correctable….either by how you carefully cajole and use them, or if some non-left-wing company deploys an LLM. That's at least something.

One of the difficult aspects of LLM "bias" will take Dembsky-like powers of articulation to explain. And I don't have those but I think I can summarize (what I've never seen summarized anywhere):

I've read several short articles on the web about the problem of LLM "bias" and how to minimize it.

I put "bias" in quotes because most of these people mean "bias" to mean "going against my world view or expectations of the way things should be."

One example used was if an LLM assumed that one is a man if told that someone was a doctor. Or vice versa, if the LLM was told that one was a nurse and assumed that meant female. The only correct answer according to nearly all articles I've read on this subject is that a lack of "bias" means not assuming that a doctor is a man or that a nurse is a woman.

2021 data says that 88.3 percent of nurses are women and about 63.7 percent of doctors are men. There are plenty of exceptions. But the unbiased "rule" (according to how the word, "bias," is being used) is that you can't notice what's going on in the real world.

Saying that it's more likely that a doctor is a man and that a woman is a nurse does not break the laws of probability. What it breaks with is Cultural Marxist demands.

One wonders that if the typical LLM assumed that an NBA player was black or that a golfer on the LPGA tour was a woman if that would be "bias"? Technically, yes. A normal person would call it "bias" if the assumption was complete color-blindness and, at least in regards to the NBA, assumed most NBA players were Asian. The "bias" we are talking about is really the commandment to ignore reality.



A good question is whether we can expect more out of the AI large language models than we do people. As it is now, at least on most issues outside of black-and-white "What is the capital of Nebraska?" questions, AI is as unreliable regarding the truth as nearly any reporter in the mainstream media.

Regardless one's opinion on AI, one thing is certain. A huge amount of money is going to be scammed and blown on it.

Some of this is understandable, especially early in the article.

www.nature.com/articles/d41586-024-01314-y