I was actually referring to its graphic counterpart, not its linguistic one

It's still questionable that it would provide trustworthy results of unique requests, as the statistical model of relationships between images and the order of sub-images in images is the same as with works and the order of words. There is no understanding of what material optimization is, only what likely images would evolve from the universe of previous human images per a input.

This is a well researched and testing application that's been around long before the current crop of IMO over-hyped AI products.

https://wooddesigner.org/opticut-software/

 

https://greekcitytimes.com/2025/04/...-chatgpt-reads-husbands-affair-in-coffee-cup/

Meanwhile, seasoned coffee readers have weighed in, noting that real tasseography involves much more than just the grounds — skilled practitioners also analyze the foam and the coffee saucer.

We aren't going to make it, are we? People, I mean. - John Conner

 

What is it with human curiosity and its eternal confusion with cause and effect?

 

I'm curious what the line of reasoning is that would lead to his conclusion. Maybe what I think of as a Large Language Model system is too restrictive compared to what you do, but I don't see a solid connection between a system that has built up a statistical model of relationships between words and strings of words so as to string them together is such a way as to conform to that model would make it particularly good at optimizing geometrical layouts.

And with the cardboard cutting question that sort of proves that it is not that good at understanding all the details that go into the solution. It was a simple geometric problem and it missed some details.
I find it cutting corners like that with a lot of test questions. Another one had to do with solving differential equations with a certain type of algorithm. It skipped one whole step and applied the input to a step that could not have been used because the input for that step was never calculated, and the input to the previous step could have never been correct at all. It's like the algorithm had three functions, and it used the first function and then used that output to use the third function, totally missing the second function entirely. That no doubt meant the end result was completely out of whack. After I reminded it that the second step was missing, it agreed and then was able to use the entire algorithm.
One thing that does help is that by discussing it with the bot sometimes I can come up with new ideas just because the discussion forces me to look at the problem in more detail too.
It can be very interesting if I keep the technical and math stuff very simple or walk it through one calculation or something.

 

https://greekcitytimes.com/2025/04/...-chatgpt-reads-husbands-affair-in-coffee-cup/

View attachment 348763

We aren't going to make it, are we? People, I mean. - John Conner

The husband also noted that this wasn’t the first time his wife had fallen under the spell of supernatural guidance. “A few years ago, she visited an astrologer and it took a whole year for her to accept that none of it was real,” he said.

Sounds like he should sign those divorce papers before she changes her mind. Run while the running's good, mate! Things ain't gonna get any better. That, or start a pool on what nonsense she will fall for next. Palm reading? Tarot cards?

 

One thing that does help is that by discussing it with the bot sometimes I can come up with new ideas just because the discussion forces me to look at the problem in more detail too.

Often, when I was struggling with a programming problem, I would start to explain the problem to a colleague, and, in the middle of the explanation, the solution became obvious.

 

Often, when I was struggling with a programming problem, I would start to explain the problem to a colleague, and, in the middle of the explanation, the solution became obvious.

Yep. By rights, I should have a very educated dog, given how many complex problems I have explained to her in great detail and that she has led me to see the solution.

 

Yep. By rights, I should have a very educated dog, given how many complex problems I have explained to her in great detail and that she has led me to see the solution.

Exactly, it's the process of nailing down details that matters, I take the time to take pictures as my way of talking to the Dog because I have visions of what I want before building it

 

Exactly, it's the process of nailing down details that matters, I take the time to take pictures as my way of talking to the Dog because I have visions of what I want before building it

This is good stuff, very important and successful designers do this kind of thing a lot. I sometimes speak out loud and try to explain to an imaginary person how something is going to work, that forces my mind to connect the dots and ensure things kind of hold together.

 

This is good stuff, very important and successful designers do this kind of thing a lot. I sometimes speak out loud and try to explain to an imaginary person how something is going to work, that forces my mind to connect the dots and ensure things kind of hold together.

Here's my position on this, if you cannot explain something to another intelligent person, then you don't actually understand it yourself.

And, there's nothing wrong with talking to yourself, or replying to your own posts.

 

Often, when I was struggling with a programming problem, I would start to explain the problem to a colleague, and, in the middle of the explanation, the solution became obvious.

Yeah that happens to me now and then too. I think it's because in the process of explaining it we have to go over every little detail and that makes it even clearer in our minds.
I like to get suggestions from others too even if they don't work because sometimes a little alteration and then they do work. Also, other minds have other ways of looking at things.

 

Yeah that happens to me now and then too. I think it's because in the process of explaining it we have to go over every little detail and that makes it even clearer in our minds.
I like to get suggestions from others too even if they don't work because sometimes a little alteration and then they do work. Also, other minds have other ways of looking at things.

One time my boss and I got together in my office to discuss how to go about designing an IC for a new project. We each had already come up with our own approaches and thought we had it pretty well finished. The two approaches were completely different. So we each started defending our own approach and pointing out the flaws that we saw with the other person's approach. It got very animated, with lots of raised voices. In the process, we each became convinced that our own approach was crap and that the other person's approach was the better starting place and, in the end, ended up with a design that was a melding of the two, which very successfully mitigated the weaknesses that we had seen in each other's ideas. When the meeting was over, which had lasted for well over an hour, we discovered that the entire rest of the company had heard us yelling at each other and had come down and listened in just out of sight of my office door, figuring that when it was over, there would be one less person working there. Since he was the president and I was just the senior engineer, there was no point in starting a pool on which one it would be. The amazing thing was that, throughout the entire exchange, he and I were actually having the time of our lives. There was never anything personal about it -- just two people having a passionate exchange of ideas defending what we each believed was the better solution but willing to acknowledge the hits the other scored. We had a common goal -- the best solution to the customer's problem, and we didn't really care whether that ended up being our own idea or not. It was one of the better examples of engineering collaboration I've been involved in.

 

One time my boss and I got together in my office to discuss how to go about designing an IC for a new project. We each had already come up with our own approaches and thought we had it pretty well finished. The two approaches were completely different. So we each started defending our own approach and pointing out the flaws that we saw with the other person's approach. It got very animated, with lots of raised voices. In the process, we each became convinced that our own approach was crap and that the other person's approach was the better starting place and, in the end, ended up with a design that was a melding of the two, which very successfully mitigated the weaknesses that we had seen in each other's ideas. When the meeting was over, which had lasted for well over an hour, we discovered that the entire rest of the company had heard us yelling at each other and had come down and listened in just out of sight of my office door, figuring that when it was over, there would be one less person working there. Since he was the president and I was just the senior engineer, there was no point in starting a pool on which one it would be. The amazing thing was that, throughout the entire exchange, he and I were actually having the time of our lives. There was never anything personal about it -- just two people having a passionate exchange of ideas defending what we each believed was the better solution but willing to acknowledge the hits the other scored. We had a common goal -- the best solution to the customer's problem, and we didn't really care whether that ended up being our own idea or not. It was one of the better examples of engineering collaboration I've been involved in.

This is the way to negotiate a specification.

The marketing guys just think I'm angry.

But, I guess it doesn't help to call them idiots.

 

Note to mods: the following link is a link to a political post on a political web site, but I am posting it here because the last paragraph is apropos to this thread and it's an interesting and insightful take.

Please don't ban me again.

Here, ZeroHedge makes a good point about AI models: they can (perhaps) tell you where a target once was, or where a broad consensus believed it to be. But they cannot tell you where the target is at this moment, or where it will be in the next instant. AI is a lagging indicator.

https://www.powerlineblog.com/archives/2025/05/and-all-the-ships-at-sea.php

 

https://arstechnica.com/ai/2025/03/...code-tells-user-to-learn-programming-instead/
An AI Coding Assistant Refused to Write Code—and Suggested the User Learn to Do It Himself

https://forum.cursor.com/t/cursor-t...ing-it-to-generate-it-limit-of-800-locs/61132

Finally, a sign of intelligence and reasoning.

 

https://arstechnica.com/ai/2025/03/...code-tells-user-to-learn-programming-instead/
An AI Coding Assistant Refused to Write Code—and Suggested the User Learn to Do It Himself

https://forum.cursor.com/t/cursor-t...ing-it-to-generate-it-limit-of-800-locs/61132

Finally, a sign of intelligence and reasoning.

Or perhaps an indication that so many people have been telling folks that same thing that it has finally risen far enough in the statistical model to become a likely response.

 

https://arstechnica.com/tech-policy...ake-ai-citations-nearly-put-them-in-a-ruling/

A plaintiff's law firms were sanctioned and ordered to pay $31,100 after submitting fake AI citations that nearly ended up in a court ruling. Michael Wilner, a retired US magistrate judge serving as special master in US District Court for the Central District of California, admitted that he initially thought the citations were real and "almost" put them into an order.

 

https://arstechnica.com/tech-policy...ake-ai-citations-nearly-put-them-in-a-ruling/

A plaintiff's law firms were sanctioned and ordered to pay $31,100 after submitting fake AI citations that nearly ended up in a court ruling. Michael Wilner, a retired US magistrate judge serving as special master in US District Court for the Central District of California, admitted that he initially thought the citations were real and "almost" put them into an order.

It's only a matter of time before something really serious gets through

 

https://www.newsweek.com/college-ai-students-professor-chatgpt-2073192

The professor asked the chatbot to create some "really nice feedback" for the student, despite many in the education sector calling on students to stop using artificial intelligence for work, according to a report from The New York Times.
...
Other class materials included distorted images, misspelled text, and other prompts, all of which are clear signs of AI usage.

However, Stapleton's business major explicitly ruled out the use of unauthorized AI and other "academically dishonest activities," leading Stapleton to file a formal complaint against the professor.

Lazy and stupid IMO

 

AI generated babies of worldleaders: