There is a concept in theater. Suspension of disbelief. Its required because we all know the actors are playing roles and nothing we are watching is real.

I am not sure that many people aren't now operating in that mode most of the time because it is expected of them so much. Some of the popular implementations of AI are just another example of that inability to distinguish between the artificial and the real.

That said, there is another version of AI. Which is a very sophisticated computer program that can process huge amounts of information and apply it to solving problems.

LLM's are mostly the former and the US consumer companies exploring them reflect that. China, reportedly, is focused on the latter. Creating AI for specific industries and tasks. There is certainly no clear line between the two. But the latter is measured by its productive output. The former is characterized by its marketing hype.

I hadn't thought of it in those simple terms. The boss's kid isn't on the list of people who need to sign off to go forward. AI will move forward with approval of those required. The kid goes off and the project is sidetracked. Carl knows the kid and consults with them anyway. Project moves forward seamlessly. Carl just changed the metaphor.

But the larger issue is that some people believe there is a special group of people with special knowledge who are the only ones really make any difference. AI fits perfectly into their world as a super version of one of them.

Well that was my poor choice of language here is the original article:

How is that not being dissatisfied with the existing metaphor? His point is that AI is incapable of being dissatisfied with Newtonian mathematics and creating an alternative way of looking at the problem that matches a new set of observations.. He may be wrong, but I think that is his point.

Einstein is an extreme example. Carl may be able to innovate as well. I think the larger danger would be that AI will end the development of new knowledge. One of the article's point is that AI is never going to be dissatisfied with what exists. It can only use the knowledge and metaphors it is trained on.

Maybe we should just say "AI thinks like a calculator".

Well yes there is. AI is a computer programmed to give the most believable answer tailored to its apparent audience. Perhaps we need to inform people that the answer they get from AI may be different than the ones it gives other people who ask the same question.

Do they? Its not clear that there value is critical thinking, which is why they are offered as examples of highly trained jobs that can be replaced. Doctors have a lot of medical knowledge that AI can easily also have and share. Therapists are essentially human chat-bots. They are trained to respond to prompts from their patients. AI can mimic that.

They would be right. I think that was the point of the article wasn't it? AI simply rearranges human thought. It doesn't generate any new thought.

What does over-confidence in LLM's have to do with anything? Plenty of people have over-confidence in all sorts of sources of information. If you want them to lose confidence you simply have to point out when it is inaccurate, incomplete, wrong and misleading. Just like you do with any source of information. And with about as much likelihood of success.

The problem you have with AI is that there are some very powerful forces at work to convince people its answers are always reliable and complete..Even its failures get portrayed as unusual aberrations to be disregarded.

Your brain hasn't fully developed until you are dead. Everyone accepts that young people can run faster, jump higher and lift more. I am not sure why they assume they can't think better. What is obvious is that they have less experience to use when thinking and likely less polished skills in describing their thoughts.

It wasn't "the hippies", it was coined by Jack Weinberg, apolitical activist in the early 60's before "hippies" even existed as a thing. It was popularized by the Yippies who weren't really "hippies" either.

This study certainly could support the proposition that after 30 people's minds change. And there have been plenty of observations that people's genius tends to peak before age 30, right along with their physical development. I am not sure that necessarily means they are "adolescent". There seems to be an assumption in that language that the changes observed are improvements rather than signs of deterioration.

Its interesting that the criticisms of this don't appear to be different outcomes. Is there any real difference between the content and presentation of a lecture by AI and one created by a human? It seems to me that is the only important question for the student.

And the question of "why do I need a college to provide me that?" is a legitimate question. If you break a college professors job down into component parts its not clear that most of the actual tasks can't theoretically be done by AI. It can create and organize a syllabus. It can write the text. It can compose the lectures. It can even interact with students and answer their questions. It may be able to do all those things better than the average professor. And do it in a way that is individualized for each student.

What it can't do is recreate the interactions with other students. It may be that the assumption of professors that their knowledge is at the center of students learning is going to take a hit here. If they have a role, its going to be their human skills at engaging students, not their technical knowledge. Which is likely true even now.

It may be that students will end up having far more control over their education with AI guiding them in setting and achieving their personal goals. The need for a college to organize their education may just disappear.

No, what is happening with coal plants is pretty simple and obvious to anyone who has ever built a fire. It takes a long time to build a fire hot enough to get the boiler hot enough to produce electricity. If you let it die out it will take a long time to get it going again and get the boiler hot. Some natural gas plants that use steam turbines have similar issues.

Which is why natural gas is used for peaker plants. But there are also base load natural gas plants that can't be.

Was the idea there was an underground sea of water every a reasonable conclusion from the evidence available? Or just the most interesting possibility. My guess is that the scientists always understood it was just an interesting possibility. The version that reached the rest of us through the media made it into something more. It certainly peaked interest in sending a mission to Mars though didn't it?