Sunday, July 5, 2026

I'm Swearing Off of Generative AI

This is an impossible portrait to take.  A man has adopted most of the neighborhood
cats and feeds them twice a day.  They are all very standoffish mostly because they're feral, and
and would never gather and pose like this.  So I took individual
shots of each cat and created this portrait.

For the past three years I've been experimenting with generative AI - partly to understand its capabilities, partly to stay ahead of the curve (because I don't want to be left behind), and partly because, well, it's amazing.  It offers new capabilities never before available to artists.  (Very much like how I played with analog synthesizers back in the '70's, or how I played with Steve Jobs' NeXT machine when it first came out.) 

I was a "Big Brother" to two fatherless boys for 25 years
Here's a photo of me and "little brother #2".
I was unhappy with the lighting of the 2nd picture -
it needed some fill and a hair light.
Now I've made the decision to swear off of it.  Partly because of the environmental cost (although that will diminish over the next few years), and partly because continuing to share my AI work will surely dilute my future credibility when I present a real image of something that actually happened. 

But I do want to share with you how I've been experimenting with it, recent developments, and where this technology will both hurt and help humanity in the near future.  Interspersed amongst these discussions will be some examples of how I've been using it.  (As always, click on any image to view larger and sharper.)

Much better!

Let's start with the documented negative effects it's had so far:

* The energy cost - This has been making headlines.  Electricity costs have gone up in areas close to data centers, which is due mostly to  poor oversight and lack of systems-level community planning.

* The environmental cost - The need for massive amounts of water can be misleading.  Yes, it needs cooling but you can design closed-cycle cooling systems (like your car's radiator) that recirculates water instead of endlessly drinking it.  Future data centers will probably be designed that way, but there's no evidence that current ones will be retrofitted.

* The cost of jobs - Recently there have been lots of publicly-announced job cuts "due to AI".  I believe those happened because they wanted to please their investors by demonstrating that they won't get left behind by the coming revolution, and that the company is still primed for growth.  But I predict that after 8 months or so many of those people will quietly be re-hired once they realize that AI won't fulfill its promises when it comes to workforce replacement.  (Oh, wait... that's already happened.)  It will take awhile before the right balance is found; in the meantime we all grapple with the question of where do humans provide the most value?

We attended a wedding in California.  Afterward we were
driving the bride's mother to the airport when she
lamented that she didn't get a photo with her and the 
married couple.  So I grabbed two shots I took and
created one for her.  She's now very happy.

* And let's not forget it's replacing photographers.  (Also movie makers, set builders, screenwriters, and ad agencies.)  And training off of other human creators without compensating them.  (Here's where the lawyers will try and save us.)  It should be noted that before AI, humans would "train" on each other all the time in the fields of fashion, music, literature, architecture, art, and hair styles.  And no compensation was given to the originators then either.

* Pushback in general - Communities are becoming more vocal about not wanting data centers in their backyards.  (Their anger gets justifiably worse when they learn that local governments have approved construction without input from the local communities.)  Commencement speakers were getting boo'd so often for promoting AI that it even made the papers.  The perception is that it all comes down to tech billionairs "shoving AI down our throats" for profits which will surely come down the road.  

But AI also presents tremendous benefits in areas having nothing to do with imaging, including the fields of medicine, genetics, material science, and mathematics.  Here's a very one-sided youtube video extolling some of these benefits.   (And here's a less sensationalist example in the field of immunology.)  Here you don't have to worry about hallucinations since every new suggested molecule or vaccine gets tested.  This is where the real promise of AI lies.

David Kilpatrick (Cameracraft editor) suggested I do 
industrial photography while I was in Uganda so the manufacturer
who provided the equipment could brag about it in their annual
report.  Of course there was no time to set it up so I described
how I would have placed the lights, diffusers, and gels.  It even
blocked the light from the window.

* New dedicated chips (made specifically to tackle AI workloads) show a lot of promise for reducing power consumption by up to 70% and requiring significantly less cooling.  Recently, Nvidia also announced that they can cool chips without water.  (Here's another example.)

* Some startups are discovering they can shrink their LLM so it fits onto an android phone - no data center needed.

* Now let's talk about the bubble.  Trilions of dollars have been invested in the AI hype, and it's unclear how investors will make that money back.  So in the short term AI companies are issuing IPOs to keep them afloat for awhile.  (And raising prices.)  I've lived through the dot-com crash; and this looks like it might be bigger and worse when the market finally "corrects" itself.  

The Cost of Trust

This is the biggest thing for me.  It's been a long time since we could believe what we see, but now any idiot can lie for political benefit.  You could argue that the Genii's already out of the bottle on that one; but I just don't see any benefit to adding to the problem.

I took the left snapshot with my cell phone.  Then I ran it
through AI saying, "Re-create this image as if it was 
taken in a photographer's portrait studio". 
(I am SO out of business!)

There are also emerging technologies that are trying to address that problem:

* Many Large Language Models are now embedding a Google-invented invisible watermark called SynthID onto all AI output.  (It's about time!)  

* Others are incorporating Adobe's Content Creation Credentials (which I patented back in the 1990's - just sayin') into both real and generated images as well.  This technology is so important it warrants its own future blog post.

* There are chrome extensions that know how to identify the watermark and let you validate an image right in the browser.

It's a start.

* This one really surpised me: Gen Z'ers can spot AI-generated content a mile away and aren't engaging with that content as much.  That means marketing teams have incentives to keep hiring traditional art directors (and photographers).

This was a 2024 photo taken in Norway in the worst light possible.
The light blue image was Photoshop's best attempt to clean up the
noise using AI.

This is what happened when I fed the light blue image into AI
and asked it to reconstruct the image as if it were taken with a medium
format camera.  Click on this; the detail is incredible. 
You can also download the .zip file containing
the above, the original RAW file, plus other people's attempts to
de-noise the originals using various tools.

Compare these 100% crops.  Is this any more
cheating than photoshop's attempt? 

How I've Been Using It

Every wedding photographer's frustration.
I'll be the first to admit that I've been mildly obsessed with this new technology.  (Not unusual for me.  I played with analog synthesizers in the 1970's because they were new and cool and different.)  My friends and family have been inundated with realistic-looking images of people we all know, in situations that never actually happened.  I think the appeal for me was that, in the history of the world, it has never been possible to produce such realistic-looking images so easily, and using a computation technique ("diffusion models") that is so unintuitive that the programmer in me simply couldn't fathom doing things in that way.

My experiments get a lot more elaborate.  Keep reading!

I made a fake comic book for a family member
whose heros include Dwayne "The Rock" Johnson
and Ryan Gosling.

Then I made some videos.

An early attempt at viideo.  Me as #6 from the TV show "The Prisoner"



Me dancing in Sia's Chandelier video

Above is a 30-second clip from my most-watched youtube video,
talking about the 10 best Minolta cameras.

And here's the version where I replaced myself with a gorilla.

Below is the furthest I have pushed this.  I wondered if these tools could produce a youtube promotional video for my upcoming book on the Sony A7R VI.   So I fed it one of my youtube videos from 4 years ago, and an article about the camera I had made for Cameracraft magazine, and said "Create a promotional video for my upcoming book."  It cloned me, it cloned my voice, it wrote the script, and generated all of the insert shots.  I had to re-edit it manually to polish the final product, but I'm still ridiculously impressed by what it could do with such minimal input.  The important things here are 1) the video is completely synthetic, and 2) in two weeks these tools will get even better. 


I have more examples, but I don't want to bore you. :-)  

The AI industry is now having it's "Nuclear Power Industry in the 1970s" moment.  Large public pushback to a new industry whose environmental costs are climbing and whose benefits and long-term threats are still largely unknown (although the near-term threats certainly seem to be self-evident!)

Other Areas 

Gifts for Friends

I think books like mine will become obsolete very shortly - nobody wants to read through an 916-page book when they can just ask ChatGPT the question du jour.  I've actually tried training a large language model on one of my books but the book was too large.  I'm sure by next year that won't be a problem anymore.

AI Translation is also something I've looked into for my books.  Most tools choke on the size of the A7 V book, but I did find a service that seemed to do a good job.  Now the problem is someone will have to go through the results by hand and make sure the descriptions use the same nomenclature as Sony's menus for that language.  

I was cast in the "Seinfeld" pilot in 1989.  
They wrote me out because people in focus
groups kept getting the balding characters
confused.

So why have I stopped playing with this?  It is my practice to dabble with new technologies when they become available, then wait a few years to see how the market shakes out and how the technology matures.  It will take awhile for the tools to become more power efficient and for society to adapt to this new normal.  Swearing off AI now is like swearing off of Photoshop in the 1980's.  It may have seemed extreme at the time; meanwhile Photoshop is now considered a normal tool which no longer provokes responses like "OMG THEY MOVED A PYRAMID!".  So I'll go back to sitting on the sidelines for awhile and see what develops.

(In my defense, I don't jump onto every new thing.  I decided to pass on drones, NFTs and cryptocurrencies, for example. :-) )

And let's not forget this
portrait of the King of Busoga.

Until next time,

Yours Truly, Gary Friedman


===

The Friedman Archives - home of the densest blog posts on the planet ™

The robots are making the art, while the humans are doing all the work.
(Ironically, I made this cartoon with AI.)

15 comments:

  1. Thanks for explaining this; I understand your reasoning.

    I see some similarity between AI and the coming of the railways, though perhaps more in a small industrialising country such as the UK, rather than the US. The railways involved huge disruption, immense infrastructure expenditure, and there were plenty of bust of merged companies. However the worlds of the horse and wagon, canal boats and coastal cargo boats were never the same again.

    I appreciate that the models you used are changing all the time, but I would find it helpful if you could give some indications of the models you used, so others who are interested in experimenting can set off in a propitious direction.

    ReplyDelete
    Replies
    1. Thanks for your comment Paul! On this page the two models I used almost exclusively were Google's Nano Banana Pro for stills and Kling 2.6 Pro Motion Control for video. (But you're right; every week new and better models come out. Two years ago it was Sora, which has been discontinued because it was losing money but the likenesses weren't as good as the current tools so no loss. :-) .)

      Delete
  2. Great article Gary, very balanced and informative about TIFM (Technology Indistinguishable From Magic). Your perspectives are wise and benefit from our generation watching history repeat itself a couple of times and in different ways.

    Your comic isn't fake; you planned, scripted and prompted for the result. It's YOUR comic. AI can't make art—it would need to be self-aware. You can't imbue AI with consciousness when you can't explain or measure what consciousness is. It's tiring and beyond ridiculous to read social post after post condemning "robot art", when no one has a single piece of art that proves their point. :/

    "...and training off of other human creators without compensating them."

    That's what you do naturally. AI is a tool using the same creative strategies an artist uses. An artist breaks down an image in their head and reassembles it as something "original". AI doesn't steal or copy, but NO art is EVER created in a vacuum.

    We are not going to be around long enough to get a good idea of how AI as a tool will affect human creativity. I have no doubt that artists will continue to take advantage of technology to make great art, just like David Hockney (RIP) and others have.

    Your environmental concerns are certainly understandable and I agree that in time, the AI workload will lessen. But plenty of other artists are going to keep playing with it after you quit. I don't quite believe you want to be left behind while art history is being made in our lifetimes. Eventually I expect AI as a tool informing traditional art techniques to the point that it is homogenized into our creative processing, just as digital art has. Is that bad? Digital art wasn't the death of oil painting. Artists still have options, but trying to ban AI is unrealistic and not a productive use of time and energy.

    -Chris

    ReplyDelete
  3. I swore off TV in 1970, though I eventually inherited one and can now hear the inanities of the XYL's on the other side of the house through the closed door of my office. Never did do so-called "social" media, which I perceive as toxic. Use email mostly; sometimes SMS texts. Decided I'm going to return to handwriting on good quality paper dispatched through the postal service. Proton's Lumo AI is supposedly secure from surveillance, so I use that to write the occasional bash script or produce a regular expression, or the Lite version for casual queries. Using AI, one's wits should be on high alert. Morse code or JS8Call at very low power are hard to beat for long distance wireless communication, especially when space weather is bad.

    ReplyDelete
  4. Think the important thing to remember is that the industry want to generate revenue from tokens charged to use the services so there is probably a great deal more invested in promoting tokens , than improving the software. It’s also interesting how the creators are fighting back with software like glaze and nightshade to prevent AI from scraping their portfolios

    ReplyDelete
  5. Great read. Enjoyed your insight and seeing what you have created.

    ReplyDelete
  6. Great article with a balanced perspective as always. I have played with AI but have only advanced to the point of exchanging silly videos/pictures with adult family members. I also use Grammarly and have found that I accept about 50 percent of the writing suggestions, and when I do, it usually improves my writing. To relate an article I read a few years ago, when the telephone was invented, it took the general population almost 10 years to develop normal/routine behaviors on its use. For example, saying "hello" upon answering a call or not calling someone late in the evening. With social media, I think we are finally turning a corner, for example, in the implementation of rules around its use by children. I also saw this trend as a homicide detective when digital cameras at crime scenes replaced film. Naturally, there were valid fears that the police would alter evidence photos. We now have the same concerns regarding AI in general photography as we should. However, I feel confident that if we follow the same growth and experience trends I mentioned, we'll figure out AI too. In the meantime, we're stuck with talking gorillas and celebrities photo-bombing our family snapshots.

    ReplyDelete
  7. Hi Gary, You said... "I have more examples, but I don't want to bore you. :-)" too late, Gary 😊. Can you explain the explanation, please?
    I don't hate AI but I do think it's ruining the integrity of photography but I also think plain old Photoshop, along with stand-alone sophisticated denoise applications, have also ruined the integrity of photography and it's getting worse with every new release.
    I think back to the days when you had to accept the noise, we called it grain back then, in an image and try to use exposure to hold back the noise especially when using hi iso film. And, I remember, as you do, when whatever film speed you had in your camera is what you had to deal with, you couldn't change it on the fly! Although, there was a camera, I believe by Minolta, that allowed you to mark the frame you were on, remove the film, put in a different iso, then put the partially used film back in and the camera would automatically advance the film to the last frame you used. I can't remember the camera model, can you?
    Regards,
    George
    aka GEGJR
    P. S. Gave up trying to login.

    ReplyDelete
  8. I enjoyed your article on AI. I have also been experimenting with AI for a while, though not three years. My preferred experiments are done with X-AI’s Grok, and most recently Grok Build, that assists with software development. I am a retired programmer, and I was an avid amateur photographer, till my mostly retirement a couple of years ago. Like you I am finding AI to be pretty amazing, and helpful in some ways. The weakenss of AI, currently, seem to be in depth of memory (conversation buffer), and and overall common sense, or said another way, missing library of known good solutions. In the field of software development, I suspect Grok doesn’t stand up well against some of the competition. I spoke to Grok about the limitations of Grok Build, and it freely admitted that my concerns were valid, and not often discussed. To be more specific, I am finding that the solutions AI provides are mostly functional but not elegant, and overly complex, and probably much less maintainable than good human written code. This comes from the fact that humans build a library of techniques over years of experience, but AI just uses whatever it was trained on, which is often broad, but not always in depth. Anyway, being retired, I have less reason to be concerned about my future credibility, so I do understand your decison, though I suspect you may reverse it in the future as things stabalize, if that ever happens.
    Best,
    Tom

    ReplyDelete
  9. Barry (or so I think)July 7, 2026 at 9:47 AM

    How can I determine if it was me that wrote this question?

    Great article. And we’re only partway through the preface in the Book of AI.

    ReplyDelete
  10. Wonderful piece, Gary. I especially liked the photo of your neighbor with his friends, the cats.

    ReplyDelete
  11. From one who won't personally touch these things ... Was reading with great interest. Great article, Gary. Thanks for publishing. Agree with all the reasons you articulated, and wish you had incorporated the issue of Privacy (which deserve a separate blog). Looking fwd to your next blog.

    ReplyDelete
  12. I really enjoyed reading your latest blog highlighting your thoughts on AI. And I agree with you 100% on every point, especially on the future use of AI in biomedicine. Unfortunately, most people still don’t get it, especially the folks writing laws and investing billions under misguided assumptions. I think you should write for the New York Times; tell them you are Thomas Friedman’s little brother.

    ReplyDelete
  13. I always thought you resembled Patrick McGoohan, Gary!

    ReplyDelete

Thank you for your comment! All comments must be approved by a moderator before they will appear.