Another demonstration of chatGPT 4.0 capabilities

It all leaves me wondering if we have moved so very far from Lorem Ipsum or Waffle generators.

Lorem Ipsum Generator

Waffle Generator (reload page to regenerate)

Prompted by your comment here, a few days ago I looked online for some AI-generated fantasy people. It was disturbing to see how many were somehow deformed or freakish, people with three arms appeared several times, hands with about fifteen fingers in a tangled knot and even one where the arm ended at the elbow. The associated forearm and hand were floating mysteriously in mid-air several inches away.

That is so common.  The AI has real problems with hands and fingers especially.  Lazy eyes are very common too.

This is my Deep Dream gallery.

https://deepdreamgenerator.com/u/silentstar/account




.

Here’s What Happens When Your Lawyer Uses ChatGPT

( https://www.nytimes.com/2023/05/27/nyreg...e4358ee2df , paywall )

I really like the following disastrous episode of a supposedly smart lawyer using ChatGPT to compose an important brief for his client. Needless to say, it didn't go well. It looks like the old legal saw that "an attorney who represents himself and uses himself for his lawyer has a fool for a client" now has to be extended to include "or who uses ChatGPT or similar AIs to do his research"

"The lawsuit began like so many others: A man named Roberto Mata sued the airline Avianca, saying he was injured when a metal serving cart struck his knee during a flight to Kennedy International Airport in New York.

When Avianca asked a Manhattan federal judge to toss out the case, Mr. Mata’s lawyers vehemently objected, submitting a 10-page brief that cited more than half a dozen relevant court decisions. There was Martinez v. Delta Air Lines, Zicherman v. Korean Air Lines and, of course, Varghese v. China Southern Airlines, with its learned discussion of federal law and “the tolling effect of the automatic stay on a statute of limitations.”

There was just one hitch: No one — not the airline’s lawyers, not even the judge himself — could find the decisions or the quotations cited and summarized in the brief.

That was because ChatGPT had invented everything.

The lawyer who created the brief, Steven A. Schwartz of the firm Levidow, Levidow & Oberman, threw himself on the mercy of the court on Thursday.

Mr. Schwartz, who has practiced law in New York for three decades, told Judge P. Kevin Castel that he had no intent to deceive the court or the airline. Mr. Schwartz said that he had never used ChatGPT, and “therefore was unaware of the possibility that its content could be false.”

He had, he told Judge Castel, even asked the program to verify that the cases were real.

It had said yes.

I'll bet he greatly regrets making that decision, and not realizing that despite really knowing nothing about ChatGPT he was risking that it could be a pathological liar willing to do anything for a good-reading composition. 

He'll probably be lucky not to be disbarred. 

So it seems it's everybody, not just naive young students, who are foolishly using ChatGPT (and similar AI systems) to compose important research and other papers.

[Laird]

What could the advantages possibly be if the information, the most important bit, is unreliable?

If you reread the quote that I provided from the article, you might see for yourself!

Laird posted a link a while back, that I think is extremely valuable. I'd recommend that everyone should listen:

https://podcasts.apple.com/us/podcast/a-...0592835492

As my own analogy, it is as though someone had a child who was diagnosed with a very rare disease that nobody knew much about. The parent might hit the internet in a big way, and try to search either for a doctor that could help or some form of treatment that was believed to help.

After doing that, his brain might be filled with a shallow mass of knowledge. He might also be able to quote whole paragraphs of medical information from the internet. Nevertheless, he probably wouldn't try to treat the child himself - unless perhaps the treatment was something very safe - he would want to find a medical professional who could devise a treatment schedule based on that knowledge.

The parent knows what he doesn't know, and can act accordingly. The AI system doesn't know what it doesn't know, nor does it care (about anything), and you use it at your peril.

Some of the discussion centred around marrying symbolic AI with the ANN approach. I have a hunch that might work for a bit, but then it would hit another, more subtle, brick wall.

Needless to say, I think the only real way to build an AI, would be to find a way to attract a spirit inside it - much as I think we inhabit our bodies for a while and then move on.

David

If you reread the quote that I provided from the article, you might see for yourself!

I find it difficult to believe that using AI for information then fact checking the information is more efficient than direct research if the AI is producing the amount of junk the writer of that quote is implying.

Needless to say, I think the only real way to build an AI, would be to find a way to attract a spirit inside it - much as I think we inhabit our bodies for a while and then move on.

I'd say it is necessary to separate the idea of AI from that of consciousness. In my understanding of what you suggest here, you are advising on how to construct something which has consciousness but not necessarily intelligence.

One of the interesting things I keep noticing as I'm exploring Midjourney, is how interesting it's fashion design is. I've been exploring using aggregated sliced images to seed the AI prompts. Here is an example image...




On the far right of this image, you can see a partial slice of the photograph of a male model in a Zip-up outdoor jacket, and more often than not, the inclusion of this photograph forces Midjourney to produce a jacket in it's output images, and the design of that that jacket often attempts to incorporate some of the other odd image slices within the aggregated image. Some of the jacket designs have been spectacular, I've often not saved them, as one gets image overload on Midjourney, and they've not been the primary purpose of what I'm playing around with.

Here are just a few images I've cropped showing jackets Midjourney has created... Some are on the extreme, but others seem relatively easy to produce... There is a very rich historical database of fashion/clothing photography available, and this seems to allow Midjourney to excel in  coming up with some amazing colours, fabrics, textures, accents, designs... where do I think this might lead... away from mass produced clothing designs... and towards more bespoke designs... more variety... an increase in smaller manufacturers... taking advantage of new technology... and/or larger manufacturers offering more customization... I find it very exciting...

One of the interesting things I keep noticing as I'm exploring Midjourney, is how interesting it's fashion design is. I've been exploring using aggregated sliced images to seed the AI prompts. Here is an example image...




On the far right of this image, you can see a partial slice of the photograph of a male model in a Zip-up outdoor jacket, and more often than not, the inclusion of this photograph forces Midjourney to produce a jacket in it's output images, and the design of that that jacket often attempts to incorporate some of the other odd image slices within the aggregated image. Some of the jacket designs have been spectacular, I've often not saved them, as one gets image overload on Midjourney, and they've not been the primary purpose of what I'm playing around with.

Here are just a few images I've cropped showing jackets Midjourney has created... Some are on the extreme, but others seem relatively easy to produce... There is a very rich historical database of fashion/clothing photography available, and this seems to allow Midjourney to excel in  coming up with some amazing colours, fabrics, textures, accents, designs... where do I think this might lead... away from mass produced clothing designs... and towards more bespoke designs... more variety... an increase in smaller manufacturers... taking advantage of new technology... and/or larger manufacturers offering more customization... I find it very exciting...

I might give that a go on Deep Dream.  It looks like a positive use of the AI and human designers might still be needed to polish the designs.  I think AI is really good for inspiration.