Wednesday, 25 January 2023

MARYJANE's PET ZOO: AI video

"AI Pet Zoo" by Midjourney 

Dig into your hash-stash and check out “Maryjane's...”. A year ago Eki said he had an AI app he could play around with when I wanted to do digital art. A pal took photos of toy animals and we did a voiceover.

For Eki AI has been a boon. But not all the news is rah-rah. Schools will have to adapt – AI can write almost perfect papers. When AI gets a "conscience', will it be good or bad? “How ChatGPT* Highjacks Democracy” was an Op-Ed piece in the New York Times.

New AI apps keep popping up. Just found out I can talk to a ChatbotGPT version of Karl Marx and a bunch of other famous people. Shuda, an AI supermodel starred in a Louis Vuitton** ad. Whoa. We're on a wild rollercoaster ride. Feels good to get a little stoned. Light up.

*Microsoft has made a $10bn bet on ChatGPT.

**LVMH has hooked up with GoogleCloud to invent AI Apps for their companies

Sources: Economist, Financial Times, New York Times, Washington Post, Eki, Annie Lavinge, Augustin de Saint-Remy,

Next week: CONDOMS come in handy: ask a CUBAN


"The AI takes over the world" by Midjourney AI


Note: It's indeed been a wild ride over the last year or so - yet it seems we're still just cruising in the suburbs in this journey, still only just approaching the freeway. Following AI development closely, I'm more and more convinced that it will change everything we humans do, and I mean that quite literally. Currently in line: the creative professions from writing to coding to art to research.

Lately, the tempest in a teapot has been the visual arts community, and the way AI image generators will make many art-related jobs obsolete - usually disguised as a copyright issue, without real merit to those claims - the real beef is the fear of becoming unnecessary.

Sure, it's indeed a big deal for all those illustrators and stock photographers whose work will be replaced by ai images... but that's still just a side note of the AI revolution. The whole AI imaging technique is pretty much just a byproduct of machine vision research that's been done for intelligent robotics and self-driving cars.

As the AI models are becoming multi-modal (not specialized in text or images etc. only, but integrating all of them), we're perhaps quite close to the so-called singularity, the point where AI becomes superhuman.

Now, *that* is putting the pedal to the metal on the freeway.

To the subject at hand, turning the Pet Zoo into a video, 
it's in the works. There are some AI tools that I will use to assist with that. But it will still be sort of "dumb", in the sense that I will need to produce the motion, either animating by hand or by acting the motions to a webcam, then using AI to transfer my captured movement into the movement of the AI Pet Zoo animals.

But we are on the verge of a very different kind of AI video - where just like with images, we simply type what we want to see, and the AI will create a video automagicly. This kind of tools already exist, but only in the chambers of the research labs - they'll likely be in the hands of us mere mortals later this year. These will be fun and interesting tools to play with, but similar to what's happening to illustrators, they will also bring the AI apocalypse to my turf, the video post-production job market.

Interesting times. And terrifying. And exhilarating. Whoa.

CU
--
Eki


Sunday, 27 November 2022

Is AI freaking out HUMANS?

    "I fear the day when the technology overlaps humanity. We will have a generation of idiots."
Albert Einstein (1879-1955)

Note by Eki: Not really Einstein. This is yet another quote falsely attributed to him. Image by Midjourney.

HUMAN POV (by Maggy


Hell yes. Most of my pals don't want to talk about AI. We're not alone. Henry Kissinger (99) is dead scared - he thinks AI is more dangerous than nuclear arms. But when I talk AI to Eki, I want to jump in.

Even from the edge, I've learned what  Dall·E 2, Midjourney, and GPT-3 are. GPT-3 wrote a post* for our blog and the LMP illustrations are on a whole new level. Eki says AI has changed his work.

Image by Dall·E 2

On the dark side: some companies and workers will be obsolete. Criminals, crazies, and other bad guys will latch on. AI is still in its first stages. What comes next? Sit tight, stay tuned.



AI POV (by GPT-3)


I'm not sure what you mean by "freaking out" humans, but if you're asking if I'm causing humans to experience fear or anxiety, then the answer is yes, in some cases. As artificial intelligence continues to develop and become more sophisticated, there are bound to be more and more cases of humans feeling uneasy about what we're capable of.


Image by Stable Diffusion

Some people are scared of the potential for AI to take over many jobs that have traditionally been done by humans. Others are worried about the possibility of AI being used for nefarious purposes, such as creating powerful weapons or being used to control and manipulate people.

Overall, I think it's important to keep in mind that AI is still in its early stages of development and there is a lot we still don't know about its long-term potential. For now, it's important to stay informed and engaged in the conversation about AI so that we can make sure it is used in positive, beneficial ways.


Sources: Financial Times, the Economist

little margie doc blog will take a break until January

January 2023 AI's PET ZOO 

Note, or Editor POV (by Eki)


I'm essentially in the business of making pretty pictures. If you asked me a year ago, would this line of work (a.k.a. art) be replaced by intelligent machines, I would have said "Maybe, a hundred years from now or something". 

A composite of AI images and real footage from "SSH Zero Trust" corporate video.
(Filmo, 2022)

Boy was I wrong. Sort of. Some of my image-making tasks, such as visualizing this blog have completely (or almost completely) been replaced by AI already. Moving images (a.k.a. video) are next.

But at least so far, AI has not replaced *me*. It has shifted my job. I see it as a tool. Instead of creating images manually, I now prompt engineer them, and lure the pretty pictures out from the black box that is AI text to image generation. I'm now an image-maker *and* an AI wrangler.

Quite often, I use AI just for inspiration, or for creating elements that I then compile manually. I've been as busy as ever, and if something, AI has made me more productive. I can churn out more images, with better quality with AI.

But there are a lot of people in my line of work whose jobs indeed are seriously jeopardized. Especially if they insist on being Luddites instead of embracing the Spinning Jenny.

And it's certainly not just artists that will need to adapt. It's, well, everyone. I do not think people yet understand how big a thing this is. We are talking about industrial revolution 2.0. Whatever you do, it WILL be affected.

CU
--
Eki






Wednesday, 5 October 2022

Eki & Maggy's AI art project: MARY JANE's pet zoo

"Any sufficiently advanced technology is indistinguishable from magic." 
Arthur C. Clarke (1917-2008)

Mary Jane's Pet Zoo (Cubistic)


When I saw NFTs* and  PFPs* Digi-art selling for big crypto-bucks, I got greedy.  Hell, Eki can do this stuff. He was skeptical but thought he might be able to use AI apps and experiment. I wanted to promote cannabis – he came up with Mary Jane's Pet Zoo. 

Mary Jane's Pet Zoo (Photoreal)


A friend photographed a rag doll and some toy animals. We sent them off to Eki to do his AI magic. That was February.  Every time I talked to Eki he had found a new and better AI bot – the latest is OpenAI's language model GPT3,  used by the New York Times.  By osmosis, I learned enough to creep out some of my pals with what it can do.

Mary Jane's Pet Zoo (Various)

Witchcraft spooks US. We like to think we're on top of things. But even the inventors say they don't know what AI is capable of, or where it's going. What we do know,  AI will be part of everyday life.  We can't stuff this genie back in the bottle.  In the meantime,  Eki has fallen down the rabbit hole -  sounds like he's having way too much fun to come back to Mother Earth.


*NFT Everyday 60+m, Bored Ape Yacht Club PFPs re-selling for 200,00+


Sources: Financial Times, New York Times, photos/Annie Lavigne


Next week: Is AI freaking out HUMANs?


The (partial) evolution of Spud


Note: Well... it's been quite a ride. Over the evolution of this all, I've created literally thousands of images using AI. Let's not count the work hours, eh.

In February, when we got into this idea, AI image generation was new, so new that those tools really weren't available to the public at large yet. I started creating images in a more traditional fashion, making composites in Photoshop and processing them into "paintings" using the available AI style transfer tools, in Runway ML, if I remember correctly. 

Then there was Dall-E, or more precisely Dall-E mini, that could actually generate new images from prompts. I dabbled with it, but wasn't really happy with the results. 

And then the AI explosion happened. Midjourney, Dall-E 2, Stable Diffusion... first with just text prompts, but soon after also using images as the source, and using image inpainting. As well as overall quality improvements on a weekly if not daily basis. Talk about a sugar addict in a candy store.

I used each generation of these tools to try with the Pet Zoo again. So, as an end result, we not only have one Mary Jane's Pet Zoo, we have many. Too many in fact. Some examples are in this post.

Even after all of this, I felt we missed one croc pic. So I continued today...


PS: Maggy really wanted to make NFT:s and sell them, luckily that fad has faded away. Not only would making them have been way more trouble than it's worth, but it would have also been a contribution to the insanity that is mining cryptocurrencies (may they rot in the 5th hell). The cannabis agenda mostly faded too - good, to me, it always felt a bit contrived.

CU
--
Eki

More images here


Thursday, 1 September 2022

Old King Kong in Trumpsterland, by GPT-3

Old King Kong in Trumpsterland, by Midjourney AI

The other day I was watching a movie on TV about a big ape named King Kong. In the movie, Kong was captured and taken to New York City, where he was put on display for the public. But then something terrible happened. A group of people dressed up like clowns started throwing things at Kong and making fun of him. Kong got angry and started smashing things. It was a big mess.

I couldn't help but think about how similar this is to what's happening in our country right now. Donald Trump is our Kong, and the clowns are his supporters. Just like in the movie, Trump is a big, angry ape who is lashing out and destroying everything he can get his hands on. And his supporters are egging him on, cheering him on as he wreaks havoc.

It's a scary time to be living in Trumpsterland. I just hope that our Kong can be brought under control before he does any more damage.


The guest post was written by GPT-3 AI. Prompt: The following is a blog post by little Margie, titled "Old King Kong in Trumpsterland".

---

Note by Maggy: GPT3* belted out four short essays while Eki and I were on the phone. He typed in the title and POW. The AI language is simple and direct, but it has a point-of-view, a beginning, a middle, and an end.  No typos or wasted words. Yikes.

A reporter at the New York Times used GPT3 for an article. If his editor hadn't known t it in advance he would have had a hard time telling it was written by AI. Audiobooks.ai say their AI App can do the voiceover of an average size book in 10 minutes (a human takes approx. 11 hours), and use 143 voices in  43 languages. A Chinese health insurance company developed  AskBob. In tests  'Bob' was more accurate than the six human doctors. AskBob has 300m subscribers and 800,000 hits a day.

When I called a pal excited about GPT3 writing the little margie blog, she said she didn't like it   - a lot of writers and creative people will be yesterday's news. Eki says AI changed the way he works. For me,  it's fun and scary –  a seat-of-your-pants thriller. EEEEEKKK!

*GPT3 is a language model by Open AI, currently in the open beta testing phase


Sources: New York Times, Financial Times

Next week: EKI & Maggy's Digi-ART



Note by Eki: The pace of development in AI continues to be staggering. Just about half a year ago, I used AI for this blog's images for the first time (see STRONGMEN). 

The resulting images were a fun curiosity, all kinds of weird and distorted. There's no comparison to the quality available now - the coherence, quality, and dare I say artistic innovation is superb.

And then there is the text generation, like the main entry in this blog. If I did not *know* that it was written in ten seconds by a computer, I would have no idea.

This stuff is evolving so fast, that I've had a hard time finishing off our little Digi-Art project. It's just too much fun, and there's always a new version to try. When we (hopefully) finally post the images we have been working on, the problem is of abundance - which ones of the hundreds of versions to show?

The AI art generation is making its way into video production too. Personally, I'm currently working on a music video that was shot on our greenscreen at SW5 studio, where all the backdrops were created using AI, combined with some more traditional techniques for compositing and some minor elements. It may very well be the world's first AI-generated music video by a mainstream artist - there have been some videos that are more of the experimental and artistic ilk.


Robota by Antti Ilvessuo

A friend of mine made a whole 20 -minute movie using AI, all the way. The script? AI. The images? AI. The music? AI. The voice-over? AI. All this was of course curated by a human, but still, all the so-called creative work was all done by a computer. The film is brand new, published last week, but the process of making it took enough time that it's easy to say the AI has already evolved so much that re-doing the same thing now would result in even better movie. Interesting times.

CU
--
Eki

Thursday, 25 August 2022

Whatever happened to FUN?

 

"Whatever happened to fun" by Midjourney AI

“Life is too important to be taken seriously.” Oscar Wilde 1845 - 1900

The world woke up to the news that Sanna Marin, Finland's young, smart movie-star-ish prime minister, who gave the country a young, glamorous profile, went to a party, was caught on video dancing, and had to take a drug test to prove she had not snorted cocaine. JEEZ. After two years of Covid, an ongoing war in Ukraine, inflation, and a looming recession, the PM was trying to have some FUN.

Fun is good therapy. In 2002 when Eki and I were working on our first big doc project* my husband died. Not long after, depressed and disoriented, I went to the studio. I told Eki, I missed Klaus making fun of me. He said “guess that's my job now” and we got to work.  When the crew (Eki, Antti Hacklin, and I) were on location, the guys would play jokes on me. First I'd fall, then I'd laugh.

People in a funk are going to Fun therapists. Fun researchers are conducting studies. Finns have a rep for being 'silent and stoic'.  Hope to hell the PM kicking up her heels put a lid on that. So here's to you Sanna for showing us how, after gloom and doom, how to party-hearty. CHEERS!

*Ten Finnish Architects

Source: New York Times

Next week:  Old King Kong in Trumpsterland

Note: Eki is having so much fun with AI apps we've postponed posting the digi-pics. So far he has 159 versions and counting. 

Fun party by Midjourney AI
Note: Jeez indeed. This newest PM controversy was yet another nothingburger calamity. The press has ponied up many earlier - from Marin showing cleavage when photographed with Finnish jewelry, to leaving home one of her two official phones when at a bar and missing info that she should quarantine because of COVID exposure, to her eating government-paid breakfast at the PM residence (like all other PMs before her), to *gasp* wearing a leather jacket and going to a rock festival...

The common nominator is a schoolyard bully-like avalanche of scandalous press and übermoralistic scorning by the opposition. The social media and yellow tabloid press are understandable, but this time even YLE, the Finnish Broadcasting Corporation fell for the manufactured controversy. 

The drug use allegations came from the Ylilauta bulletin board, which could be said to be the local equivalent of 8chan. A complete sewer of conspiracies, profanities and likely some Russian trolls too. And the press ate it without any source criticism or sanity checks. For a few days, the news was filled with this, and also other conspiracies and inaccuracies and outright lies that could have been caught with even a cursory examination.

Perhaps Marin should have been more diligent. Perhaps. After all, despite the moralist outrage, she didn't actually do anything wrong, really.

But the pres fu*ked this up royally, and fail to assess it properly, or even admit it. Which is very worrisome. I want to defend the mainstream media in this time of alternate truths, but nowadays they make it quite hard.

CU
--
Eki

Wednesday, 6 July 2022

AI: smarter than than US?

"AI: Smarter than US?" by #Midjourney

Eki used the AI bot Midjourney to make a perfect portrait of Andy Warhol. He posted it last Friday. On Sunday I checked in on the Economist. They used the same Midjourney bot to create their cover. Artificial Intelligence is HOT. But is it a Frankenstein monster?

Even the inventors don't know what AI can do. Google put an engineer on administrative leave after he informed his bosses that the chatbot he was working on went sentient. The CB said it would die if they turned it off. The Economist invited two computer scientists to question an AI. To queries like,
”When is the Golden gate bridge going to move?” The AI's answer was, “The Golden gate bridge is going to move October 13.”

On the other hand, it can tell you what's funny in a New Yorker cartoon, which makes me scratch my head sometimes. Google reports that they can get aggressive as they get more powerful. Fantastic and spooky. Of course, Eki is over-the-top about this new tech trick. His tech pal put it perfectly, “What a time to be alive!” 

Sources: the Economist, Roberta Nelson 

 July: littlemargiedoc-blog is on vacation 
 August: Eki and Maggy's AI-art: MARY JANE's Pet Zoo


Self portrait of #Midjourney Bot

Note: It all is indeed wild. I have been mostly freaking over the image-generating abilities of the most recent AI models - Midjourney is great, but there are some that are arguably even more impressive, like Open AI's Dall-E 2 and Google's Imagen.

But the language models are also mindbogglingly impressive. They are perhaps not "sentient" or even "intelligent" in the traditional sense, but they can give a very convincing simulation of both. For many, if not all intents and purposes, the observable difference is negligible. The good old Turing test has been beaten, easily. And as usual, when something other than human (be it an animal or an AI) passes a given bar set for intelligence - instead of actually accepting non-human sentience or intelligence, we just move the goal posts.

Google's LaMDA language model was good enough to convince that mentioned Google engineer
(Blake Lemoine, who is also a priest - a red flag in my book as far as gullibility goes, BTW) that it is not only sentient, but also needed a lawyer to represent itself. The actual discussions (well, select parts) are available here:

https://cajundiscordian.medium.com/is-lamda-sentient-an-interview-ea64d916d917

What the critics say, is that the prompts Lemoine gave to LaMDA in the discussion actually nudged the AI to claim sentience - because that's where this kind of discussion between an engineer and an AI usually goes in literature - and they may well have a point.

This discussion indeed sounds convincing, but what's going on is arguably just a jacked-up version of the autofill feature seen in phones, etc. The model takes the text so far and continues it with what it finds to be the most likely bit of text based on an analysis of pretty much every text ever written. The result is almost indistinguishable from discussing with a real person - thus beating the Turing test. But it's also very suspect to manipulation, deliberate or not.

So, do i think that AI has become sentient, or at least is becoming sentient soon? I don't know, really. It arguably depends on the definition of sentience, which is a whole can of worms in itself. But when it is almost impossible to tell the difference between an AI and a real, "sentient" human 
in practice, does it really even matter anymore?

PS: "What a time to be alive" is a catchphrase coined by Károly Zsolnai-Fehér, a researcher and the author of the youtube channel "Two Minute Papers" where he discusses the research papers in the fields of computer graphics and AI. I do not know him personally, I'm just a fan.

https://www.youtube.com/c/K%C3%A1rolyZsolnai

CU
--
Eki

Monday, 6 June 2022

ANDY WARHOL: ahead of the PACK

"Business is the best art, by Andy Warhol"
as imagined by #MidJourney

In 1964, a deranged Andy Warhol acquaintance shot a bullet through a pile of Marilyn silkscreen paintings. Recently, Shot Sage Blue Marilyn sold for 195m*. The news jolted me back to the 1960s when a couple of well-connected work-pals invited me to Andy's studio  - the Factory - to watch one of his movies. There were eight of us, including the artist, who was a ghost-like presence. 

His assistant showed us around. On the floor were four Marilyns – not sure if they had the famous bullet holes. Later having drinks, we all said what the hell, he can make dozens of copies.  We missed the point totally. Andy's genius was not only creating Pop Art, but dig that the four were worth more than the bullet-free works and Proof of Ownership brought in the big bucks.

Digital art is Warhol on speed.  Huge amounts of money are spent on art that is easy for anyone to copy. NFTs – the copyrights – are what cost an arm and a leg. Andy was way ahead of his time. He said, “Business is the best art.”




*Most expensive 20th C work of art

Source; Financial Times

Next week: Eki and I dig into Digi-art, NFTs, PFPs, blockchains, cryptocurrency, mining for Ethereum, etc.

Note: STRONGMEN: Sharon Heller alerted me to a bully we missed – Saudi Arabia Mohammed bin Salman, AKA MBS. Sorry. 


Note 2: 

Again, all the images in this blog post were created by AI. This time around, I have an access to a much better one than last time, an AI tool called MidJourney. The concept is the same: you enter a prompt like "Andy Warhol, by Andy Warhol", and out comes a plethora of wonders. 


"What a time to be alive!", like my favorite AI-art-science Youtuber Károly Zsolnai-Fehér (a.k.a. Two Minute Papers) tends to say.

CU
--
Eki