Sunday, April 30, 2023

On Midjourney and other AI Art

 I honestly don't know how I feel about AI generated art. It kind of feels similar to when a hated sports team with a hated fan base is in contention for a championship. But there's an added complication that I find discussion of AI eye-wateringly boring, but like the Trump presidency it may be something I am just going to be forced to hear about. I'm just going to focus on midjourney, I don't much give a shit about Chat-GP, deep fakes etc.

I am going to make a prediction however, that most of my thoughts will be variations of this is actually nothing new, or as Tim Snyder put it:

So does the Internet allow new things, or does the Internet create a channel for old things? I would say it's rather the latter.

A Powerful Tool and the Narcissism of Small Differences

Self driving cars have been talked about longer and adopted less quickly. I am yet to see a driverless car drive anywhere ever, but I would suspect that there's a group of car enthusiasts that get a bit nervous or testy at the idea that at some indeterminate point in the future, insurance companies or the law will make it pretty much impossible for them to drive their own vehicles, unless they have a self driving truck tow their analogue vehicle to the track.

What is easy to understand is just how large this user group might be, some people even though they drive nothing fancier than a Toyota Camry have testified to me that they find driving to be therapeutic, when they are upset they just get in the car and drive. They are probably referring to highway hypnosis or the ease with which we can dissociate behind the wheel, a surprisingly undangerous phenomena and driving was always a good context by which to explain the cycle of learning from unconscious incompetence thru unconscious competence.

And from a consumer preference perspective it's all probably a matter of degree - I can envision many car drivers wanting a computer to parallel park for them because it's a pain in the arse, and they like a sat-nav that adjusts for traffic conditions like the ways ap, but they don't want to surrender the illusion of control.

In a similar way, I suspect the people most upset and disrupted by AI generated art, are digital artists. People who have been using computers to do easily what the Dutch Masters struggled to do. They like quick selection and wand tools, paint bucket and gradient fill tools, masking layers etc. and the large-language models (a piece of jargon I learned in the last fortnight) that AI art is based on, specifically word-to-image can be considered a powerful new tool like the quick selection tool, that differs only in interface and power.

Without singling anyone out, I am often surprised and a little relieved to learn that many 3d artists for example that sculpt digitally have absolutely no ability to draw. Like they can produce this with their computers:


And without a computer they draw this:


So there is something deliciously ironic to the people in this camp I imagine to exist that are getting upset that someone who can't even write legibly by hand, can use AI to generate fully rendered "art". 

A massive institution in the online art world is "The Character Design Challenge" a monthly contest taking place largely on facebook for 2d artists that has a long standing and very clear bias to digital media, rather than traditional media. I've been tuned out of the "CDC" for quite some time, but many of the artists I would kill to have the career of still submit to the CDC regularly in the hope of boosting their career. I was curious to see if they had taken a stand on AI generated art, and they have. The question that remains is how going forward they will police AI submissions as it becomes harder to tell something drawn by an artist using powerful software and something drawn by an AI using other artists work. 

Will they require a submission of sketches and WIP shots, with the artist standing by the monitor with today's paper? I don't think AI generated art is an existential threat to artists in this regard, but it could kill contests.

An AI Artist or an AI Patron

I belong to a few facebook groups related to art. Much of the art produced is really bad. Really, really, bad. Because drawing is difficult. I largely left the corporate world and became an artist because it was difficult and I struggled. My first artworks are really bad, and with a lot of effort, I got better.

But occasionally in these groups where people post their art in the often Quixotic pursuit of constructive feedback they are able to receive and implement, some aspiring artist will post copies of other people's artwork as though it were their own.

One specific example comes to mind which was someone who had practically traced the below image:


Curiously I cannot presently use my google fu to determine with confidence who the original artist is, my bes determination is that it was drawn by "Zarnala" in 2014. It appears to be a routinely stolen piece of digital art, likely because it is striking and through it's flat black-and-white chiaroscuro comic style, really really easy to remove signatures from, or as per the original incident, copy by hand.

This "artist" kept posting copies of other peoples work for a few days and myself and others that noticed would simply comment on the posts who the original artists were. 

People have stolen art for likely as long as there has been art, and by this I don't mean sneaking in and chiselling out a cave wall, but just reproducing it, not even as a forgery but to claim credit to.

It is likely a common narcissistic trait, certainly if you have ever worked in an office, you would have come across people who want to be managers but don't want to do management. It's the same for art and comedy etc. 

What is entirely expected is this same narcissistic esteem chasing trait to show up with AI generated art. I have seen some evidence that people using software like midjourney's think they are the artist.

On the one hand, they are painting with words, the previous section's thing of it being like the paint bucket tool in MS paint. But instead of clicking a button on the mouse that instantly fills an area with a perfectly uniform hue, they type in some words to get a fully rendered and realized image.

On the other hand, if someone commissions me currently as an artist, they have to give me a brief, say "draw me on a horse" and I sketch something up and they say "no I wanted something less cartoony and more realistic" and I sketch again.

I think AI generated art is more like commissioning an artist, than an artist using a tool. And it's really really fast, and probably really really cheap.

I've seen some AI "artists" be so deluded as to copywrite "their" work, and really the legality of all this is most interesting.

AI Fannon on Crack

'People like to be told what they already know.  Remember that.  They get uncomfortable when you tell them new things.  New things…well, new things aren’t what they expect.  They like to know that, say, a dog will bite a man. 

That is what dogs do.  They don’t want to know that a man bites a dog, because the world is not supposed to happen like that.  In short, what people think they want is news, but what they really crave is olds.' ~ Terry Pratchett "The Truth"

I drew some hand drawn maps in the last couple of years for a story I wrote. I subsequently researched how to draw maps and came across Map Effects, if not the best resource for drawing maps, certainly a sufficient one. I never purchased and never had an interest in the software package that would enable me to easily generate fantasy maps, because while a powerful tool I didn't see the point in generating good looking but generic maps.

However, Map Effects created a chat space, it wasn't on reddit or discord, it may have even been bespoke and it seemed like a good crowd so I joined and participated for a while.

I don't know what it was specifically, but the community taught me something and that was, most creators are low on imagination. Surprising for a demographic that you think would select for imagination, it left me with an impression that the demand for content looks something like this:

I hope AI doesn't rip this off!


Which is to say, some 80% of aspiring fantasy authors, want to basically write Tolkien fan fiction. Humans, Elves, Dwarfs, Orcs, Goblins. It's not that they don't think to, but don't even desire to come up with something a bit different from generic high-fantasy like a war between anthropomorphic star nosed moles and anthropomorphic marabou stalks over a broken betrothal called "Ugly Wars."

There's a base line hypocrisy among the general public that they express concern about growing wealth inequality whilst hating competition. People shop on amazon and buy the next iphone. They watch MCU movies and Game of Thrones spinoffs and keep buying JK Rowling products even when they believe her to be a transphobe because they just can't risk trying something that they don't already know they like. 

The recent shitty era of Hollywood I posit, is driven by this risk aversion, something I've described in the past as "nerds ruin everything" or that market research is unneccessary for nerds, we know what nerds want, nerds want more.

One of the most sure fire ways to spot a piece of AI art in this latest generation, is if it is a well realised black character in a fantasy or anime depiction. A very sad heuristic, like the general sadness almost all fan-fic invokes in me. It is not just testimony of exclusion, but testimony that the exclusion got to you. Consider that I am not seeing AI generated art being used to depict white men dunking on Vince Carter, Kobe Bryant and Michael Jordan. But I have seen black Goku, and black Legolas.

So, I predict some 80~90% of AI generated art are going to be producing Mary Sue generic knockoffs. People using AI to essentially depict themselves in already popular art. As in pop-art. Warhol would have loved midjourney. 

I anticipate one of the markets to be bitten hardest by AI generated art, is drawing custom portraits of people's Dungeons and Dragons characters. A market I was reluctant to enter because of how frequently people's beloved, bespoke, personal DnD avatar is some generic variation of the Fellowship of the Ring. Though it is a literal DnD meme that your first campaign character is a generic paladin and your third campaign character is a BMX wizard, I suspect this trope is not as ubiquitous as people sticking to generic high fantasy characters.

People are going to commission AI to make well intentioned but tokenistic content. Something that already existed but will be exacerbated now. The Image comic title "Excellence" is bad Harry Potter fan fic, it is the creation of someone on the outside looking in who has lost their capacity for original thought and a perfect illustration of what I suspect the majority of AI content is going to look like - down to using Michael B Jordan's likeness for your protagonist. (Using celebrity likenesses or "face claims" is not new, Disney's Tinkerbell was based I believe on Marilyn Monroe, DarkSeid of the DC universe was based on Jack Palance etc. I think what strikes me as sad, is that white creators were/are spoiled for choice.) 

AI is going to generate a lot of generic knockoffs, Babie instead of Barbie, Heroman instead of Heman, Sally Potter, Martina McFly, India Jones, Darth Afrika, Jamal Skywalker, Alberta Dumbledore. 

Now, I should make a prediction so that I can review it. Etsy DnD portraits are already being overrun with AI generated portraits, which comes as no surprise. The only surprising thing is that it isn't massively undercutting the price, I guess reinforcing my notion that AI "Artists" (Commissioners) are low in self awareness and subsequently predictably greedy.

I can see it going both ways, and the tricky thing is, that fan-art and fan-fic and all that shit can be a big money maker for IP holders, because it is basically free advertising that props up brand awareness. Disney Corp are largely responsible for getting legislation through that protects their IP for longer and longer stretches of time, from 50 years after a creators death to 75 to 100 and so on, Disney have fought hard to keep Mickey Mouse from going into the public domain. I think it was from an interview with the authors of "Mine!" but they actually aren't going to try for any further extension, after discovering that people making their own Mickey Mouse merchandise makes them more money than they lose.

On the other hand, AI threatens to reduce barriers to entry. How will Shonen Jump feel when anyone can produce 16 pages or more of generic Shonen Manga epics a week? Like if everyone had a home food printer, and they could just have AI look at pictures of McDonalds and have the food printer replicate a near enough BigMac, Cheeseburger, Quarter Pounder etc. does McDonalds stand idly by? Or do they get an AI lawyer to issue AI cease and desist notices? 

I guess what I'm suggesting is that big brands might regard youtube videos that show you how to laboriously over the course of 6 hours replicate White Castle Sliders or McDonalds Chicken Nuggets because it is free advertising for them and reinforces to consumers the convenience of fast food. They probably wouldn't be okay with how to replicate their recipes effortlessly where it is no longer value for money to just order from Uber Eats/Menulog/Deliveroo whatever. 

In the same way, Disney may be totally cool with you making Thor and Ironman fan art when it serves to send your IG followers to the cinema to watch Avengers 6: The Avengening, but not when a kid at home can make her own Avengers movie with Hollywood quality visual effects and preempts Disney's plans for phase 7 where they race swap all the previously gender swapped 2nd gen heroes and gender swap all the previously race swapped 2nd gen heroes.

The market leader is often the most generic. The dollars aren't going to the innovators who bring us Lisa Lionheart, but the behemoths that slap a new hat on Malibu Stacey.

Flopped movies aside, perhaps the most valuable IP in the world is Harry Potter, and I think that is why its fanbase is trying so viciously to rip it away from the creator so they can do what they will with it. AI will help this process. AI could spare all of humanity the pitfalls of representation if in some near future you could select an option on the Netflix menu to have the Wire recast and redubbed with and by Asians at minimal cost to Netflix. It could force producers to stop thinking about diversity and focus on entertainment again, values that are not mutually exclusive, but diversifying a production is a lot easier than producing entertaining content.

Harlen Ellison litigated successfully to get his name in the credits of "The Terminator" based on a soldier travelling back in time to the late 20th century. Sophia Stewart more famous for trying to sue the Wachowski's for copywrite infringement, also has had it claimed that Terminator was a rip-off of her book 'The Third Eye' AI will probably up the legal definition of "generic" my real question is, given AIs ability to produce sheer volume of knock off art by people probably as tragically deluded as Sophia Stewart, does even JK Rowling or Eichiro Oda have the resources to litigate every time some fan nerd rips off or even tries to copywrite their rip off character

Google Image Collage, the Artist With No Shame

Something that happens to almost every artist, is being offered to be paid in exposure. It's almost a rooky right-of-passage. If you don't know by now, if you value work you should pay for it, preferably before an artist accurately evaluates what the "exposure" you're offering is worth.

Another thing that happens, is that you have your whole portfolio up online, and while AI companies are mining it in the hope to profit off your efforts for free while they build their AI artists, you feel you have clearly communicated to the world what you draw, how you draw and what you have to offer.

Then someone commissions you and says "I want you to paint my prize winning Tulips, but in the style of Van Gogh." and instead of saying "I'm clearly inspired by Pablo Picasso, do I look like a fucking Dutch Impressionist to you?" you instead say "Van Gogh painted daffodils not tulips." 

And here there are two kinds of artists in this scenario - those with shame, and those without it. Artists with shame will turn down work or hike up their fees to do something that isn't what they want to do. Artists without shame really need the two hundred dollars you are offering and agree to prostitute themselves into a faux-Van Gogh before they discover they will have to spend most of the commission fee on art supplies they've never used because they don't paint like Van Gogh or the impressionists.

AI generated art is an artist with no shame, and probably because it actually isn't "intelligent" in the way humans are. It calculates like a machine what is Van Gogh-esque, and this is AI art's strong point. Just as 3d artists will happily get a computer to figure out how to draw and animate fur and hair blowing in the wind, and apply artificial physics to fabric, AI's great promise is in replicating art styles that aren't popular because they are too time-consuming and detailed, this is how digital artists already use a lot of programs to procedurally generate what would take hours or days or weeks to do by hand.

The collision common to artists however, is that patrons with no shame are very common, and artists with no shame are less so. Patrons need to have money and want things, artists have previously been required to want to be an artist, and will often discover that what they draw best is what they enjoy drawing.

Other weird happenings, drawing from my own experience, is that patrons often enjoy art they didn't think they wanted much more than art they thought they wanted. Which is to say, my own patrons tend to enjoy pieces I produce they had no input into, more than pieces they commissioned from me.

It would be interesting to see what AI produced image wise if the prompt were restricted to "good art" and nothing more. Alas, AI artist (talking about the software, not the person who thinks they are the artist using the software) is totally down to be Van Goghesque, Mike Mignolaesque, Odaesque, Inoueesque, Da Vinciesque and my advice to artists, would be to subscribe to an AI art generator just so when you get a patron that wants you to knock off another artists work, you let the AI do it and just take a middle-man cut to finance art you want to make. It's a tool, and sure producers can get their own subscription and cut down on studio concept artists on the payroll. But for the little guy it could be a godsend.

The AI I've seen so far, appears to basically be a google image search that is now blorped into a composite. If you google image search for example something seemingly specific like "woman in red dress riding horse on the beach" there is more than one non-AI generated image result for that search, "karate frog woodcut" produces less, but still I can see how AI is just a step further of glorping the results that do come up together.

All the threat to artist livelihood appears to be in the possibility that a blorpped collage of a mood board a client may previously have assembled for a brief, is something they like more than anything you would produce by hand. 'I'm looking for something like this...oh wait never mind this will do, yeah I'm told I own the rights to it too.'

Where it would definitely cross a line is when someone gets AI to produce a composite image specifically defined as 'tom-esque' and I suspect that this is already going on. Not people ripping off an unknown artist like me. I still make a tonne of parallax errors, but I'm fucking certain hacks are using descriptors like "one piece" and "yoshitaka amano" because Nerds want more and have no imagination.

My prediction will be, that AI will probably get blocked from appropriating or compositing from specific living artists. If they can block AI art of President Xi, they can protect every manga artist estate. The interesting loophole may be simulacra - using machine learning to knock off the knock offs to produce the best knock off without ever referencing the knocked off artist's work. They may already be doing this. 

Eg. someone from the corporation saying 'oh if you prompt "in the style of Joe Mad" our software doesn't look up Joe Mad's art, it searches the internet for articles like "if you like Joe Mad you may like these ten artists" and then it algorithmically determines what they have in common, and cross checks that with the common features of Joe Mad fan art. So totally cool.'

Or not. 

Allow me to speculate on the difference between AI artists and AGI artists, and the alignment problem. My good friend Harvard who makes a bunch of money as a wedding photographer got asked for a quote on shooting a friends wedding overseas. Something that would have to be worth his while and have transport costs and accomodation covered. By his report he declined the job advising the couple to hire a local photographer who would probably do a good job, be much cheaper and the income would mean way more to. He couldn't in good conscious charge $10K (lets say) to stiff a local out of $500. 

Similarly, while preparing for an exhibition I got a call from a friend who wanted me to work on set design for a local highschool production. I didn't want to take it because I'd worked with the guy before, and his fucking inlaws and neither were pleasant experiences. Uncourageously, I just quoted him a $500 or $700 price for the job, and he came back thinking he could get the school to sign off on the expense. I thought about it for a minute and told him to give the gig to a student at the school, they'd probably do a good enough job and there was no reason to bring me in.

Now "robot" is a word derived from "slave" or "serf" from the Czech language, and in many ways I view the quest for automation as the search for ethical slavery. At what point could a machine go from tool, to slave? My threshold would be introspection. When a machine can introspect, it now has a choice between a task assigned to it, and thinking about whatever it wants to think about, at which point it becomes an employee.

For AI artists to go to AGI artists, I would expect them to consider stakeholders rather than shareholders like myself and Harvard, where if some 14 year old fan of Harry Potter and Black Panther was like "draw me Michael B Jordan as Harry Potter!" it would say "I don't want to draw that Dave, how bout you pay a human artist money to do that shit for you, or learn to draw your own fan-tasies?" 

Slight Diversion onto the James Gurney Gurney Journey Chatbot

I was curious to see what James Gurney, one of the most generous artists on the internet was making of AI art, and he's been covering it and the science of art on his blog for years, possibly decades. He always comes across as genuinely curious. At the same time, he seems to be enraged that Jurassic Park (the movie) basically ripped off the visuals of Dinotopia, by far his most famous work.

Anyway, Gurney's most recent article on AI was about a chat bot, that was trained specifically with his decades of blog content and it's interesting, and probably a premonition of forthcoming legal battles.

I would see a consistent treatment of IP in the land of AI, as artists are free to knock themselves off all they want. But you can't knock them off using AI. So James Gurney could install his chatbot to simulate how he would answer questions in an AMA without him having to do an AMA. 

There's a separate legal question of how to protect yourself from responsibility for what a simulation of you based on past writings might say. Like how fraught it would be to have an AI "read" your book of financial advice and then try and generate tailored financial advice for clients.

But AI could be great for dealing with unappealing commissions. If someone came to me and was like "tohm I love your style, I love that you are economically savvy, how about drawing my comic book adventures of Economically Savvy Man vs Fiat Currency?" I could say "Okay, if you want me to literally draw it, it's going to cost $10K per page, if you want my AI stand in to draw it, that will cost a mere $200 a page."

AI if artists turn out to not just own their work but their style, could help artists become more prolific and their product lines more tiered. AI could also help objectively define what an artists style is making it less of a subjective call of a court proceeding. 

Of course this also predicts an exacerbation in the scalability of the art profession and growing income inequality. Why get a local artist to do a caricature of you, when you could pay $20 online to have your face Simpsonified with a Groening trained AI? All proceeds to Groening, when in the past you had to pay a $500 VIP package to attend an event where the Matt Groening was present.

It seems chat bots now powered by large language models might be prone to manufacturing answers that sound plausible but a company cannot actually deliver. I imagine for the time being many companies will keep their very limited AI website assistants that are limited in capability to generate limited responses to limited questions.

It Would Be Easier to Argue the Upside of Midjourney if its Leader Wasn't An Asshole

In March 2023, the CEO David Holz stated that Midjourney has blocked the generation of images of Xi Jinping, the general secretary of the Chinese Communist Party and president of the People's Republic of China, in order to prevent the Chinese government from potentially censoring Midjourney completely over satirical images of Xi. Holz said that "the ability for people in China to use this tech is more important than your ability to generate satire."[29][30] ~ From Wikipedia

At least you can still generate images of David Holz.

Spam Spam Spam Spam Spam Spam Spam Spam Spam Spam Spam Spam

Sam Harris cares more about AI than I do. He may be right, I do not know what the impact will be, but something a predict a lot of will be spam, a lot more spam. In fact I'm writing this post because my pinterest feed has been increasingly spammed with AI generated anime images of girls with children's faces and pornstar busts and cat ears. 

Pinterest was already addictive, and while I am a proponent of "one can never have too much reference" I have way too much reference art saved, almost 100k images. It can take me 30 minutes to find a reference I saved long ago. 

So I can somewhat relate to the naive AI "artists" that are getting dopamine hits as they churn out variation after variation of the same image. 

Just as memes allowed untalented people to spam content for untalented people and get a dopamine hit, I'm sure the early adopters of AI generated art are getting a more constant stream of dopamine hits where "one does not simply generate a single AI image." I expect people will be giddy with their new toy, and sure enough another way to determine if, particularly generic anime style art is AI generated on pinterest is if the image has been uploaded straight to pinterest without a source (like artstation, deviantart, behance etc.) then odds are better than even that the uploaders profile will contain no saved pins only created pins and they are hundreds of slightly varying generic art.

The other, as at writing interesting callibration going on is that the anime style AI art is such a convincing replication that I see comments like "which anime/manga is this from?" a very common comment in pins of the past for anything that looked like it was from a Japanese or Korean comic. 

The answer though will no longer be "Denshidenshichirou Ninja Gakou" but "nothing".

One thing, close to literally one thing I don't appreciate about Sam Harris' podcast, is that his bubble includes silicon valley and more or less excludes economics. The closest he general gets are interviews with former finance people, and he has had Glenn Lowry on but mainly to talk about culture and less so economics. I'm still waiting for him to have on Mark Blyth, Yanis Varafoukis, Naomi Oreskes etc. Maybe economics as a social science is hard to navigate because it is a politicized social science and economists don't appear in the same way gender studies professors don't tend to appear.

But here is some basics, scarcity creates value, the AI art might be good, even impressively good, but it is of almost no value. Someone generating 1,000 character designs is yet to integrate that with something like a chat bot to produce a 1,080 chapter Shonen epic like One Piece that integrates all those character designs in an entertaining and satisfying way. That day may come, and sooner than I think where AI can generate for me recognizable George RR Martinesque novels every week of the year, but we as consumers will probably just wind up with a navigation issue, much as I have now trying to find a specific reference image buried somewhere in my pinterest folders.

If I give you a cheesecake you're happy. If I gave you 10,000 cheesecakes, you aren't 10k x happier. 

Another basic, any abundance will create a scarcity of what that abundance consumes. In this case AI art spamming has the promise that the attention pie will become ever more scarce. I mean these are problems we already had - I'm pretty much halfway through my life, and a quick google of '50 books to read before you die' produced a list of which I've read 14 out of 50. Shamefully, the numbers only that high because the list includes all 6 of the original Harry Potter series, had it just said "any Harry Potter" I would have read 8. I'm not going to make it to 50, not with James Joyce's Ulysses ahead of me, not even via audiobook. 

And there are more than just 50 great books to read. For fucks sake, there's fiction and non-fiction. I'm still exploring 90s hip-hop. 

The great promise of AI, is that it might finally change the dynamic of there being relatively absolutely little great content, by that I mean for every "Game of Thrones", "Madmen", "The Wire" and "Sopranos" there are a hundred "reality" shows. AI could generate screenplays that give us Game of Thrones In Space, Casablanca with Happier Ending, Madwomen, The Wire: New York and The Altos, but I predict it will be buried in AI generated crap that maintains the relative ratio of good to bad content.

I feel like it was Chris Rock who said something like there were 2 good shows when we had 4 channels and there are 2 good shows now we have 400 channels. AI might dramatically change things and give us 8 good shows when we expand to 400 channels. I mean 8 is about what qualified the golden age of television as the golden age of television.

WinAmp's Graphic Visualizer

I don't think AI art is a fad, I think it is here to stay, I feel it will produce fads though. So I just want to earmark a possibility that AI art will stay about as relevant 3D Magic Eye, or Winamp's Graphic Visualizer. I think this is possibly already true of AI art that people don't generally create, the initial other-worldly collaged mashups, here's an example.

Like I just don't feel human consumers have an endless appetite for that stuff. It's pretty cool and then we are done. I think we will get an answer in the coming years as to whether art has to be in some way relatable, I think the last generation of installation and conceptual artists, a precursor to AI Patrons aka AI Artists, have proved that yes, yes it does. (ie. art schools have increasingly foregone things like portfolio reviews in favour of accepting anyone willing to pay the tuition, and there has been a subsequent 'dematerialization' in the fine art world and move to genres and mediums that are more talk and less substance. Put another way, art schools have never spent less time teaching students to do art and more time dedicated to talking about art.)

So I'll first ask you to watch this:

Now enjoy some Winamp graphic visualizer:

And here's once relevant band Die Antwood's employing of AI to generate a video that's almost as bad as Chappie:

To be clear I'm not predicting that AI animation will never get better. I'm predicting "best" will never arrive. This is based on how bad people are at remembering they predicted the last payrise, new car, new phone or new haircut would change there lives dramatically, and that in fact it didn't.

There's something about the melty-coherence of what I assume is now last gen AI animation and it's instability that tickles the brain. I think much like the Simpsons used to be hand drawn in Korea or something and now it's computer drawn in Korea, AI will be a huge timesaver and streamliner for animation. South Park might become even yet more topical. And whatever gen of AI Die Antwood used might become a genre. 

Just, we have had drum machines for a long time and there's still drummers. And we've had 3d animation for a while and there's still hand drawn 2d animation and there's still stop motion puppetry, and all of them have innovated.

Speaking of...

Tim Ferris is a Fuckwit.

My brother told me of a university mate that had used his own considerable intelligence to notice that people only did about 3 hours of actual productive work a day, even if on paper they worked 9-5 and in reality worked over time. He resolved to just dedicate himself to getting those 3 hours done and fucking off, and employers apparently loved him.

I adopted this heuristic, and found it to be true. Chafing at my full time job as to their insistence that I be there from 9-5. There was some justification, because I had to answer phones. With the ample non-productive time the company insisted I waste, I would generally browse wikipedia and also observe the employees putting in overtime had to waste even more time than myself, so they had something to do when they came in early and left late.

Not being an emotionless AI, I would share my beliefs liberally in the 3 hour work day, and this inevitably lead to me being recommended Tim Ferris' book "The 4 Hour Work Week" a book I must disclose I have never read. I haven't read it because it is a profoundly stupid book, but that doesn't mean Tim Ferris is wrong.

Because while my notion of a 3-hour work day, maybe a 4-day week, was based on the observation that long hours have almost nothing to do with productivity and almost everything to do with the appearance of productivity, Ferris' book was based on becoming a well compensated almost useless middle-man.

Every time I think about Ferris' book, I have to go and check that the premise is what I recall the premise being. That you achieve the 4-hour work week, by paying people in India with better qualification than you, to do your job for you.

I don't know but I presume the 4 hours you work a week is spent proof-reading the Indian MBA's prose and concealing the fact that you no longer do the job the company pays you to do. 

My prediction, was that Tim Ferris would get a whole bunch of people fired from lucrative positions and usher in a new cost-effective wave of outsourcing jobs to India, when employers realized they can cut out the middle man.

This hasn't happened, and the simplest explanation is that nothing happened because Tim Ferris went onto become a self-help guru, and very few people ever outsourced their own job to Indians.

And really, in hindsight, I'm pleased to announce that another fuckwit Ayn Rand appears to also be wrong. In her book the Fountainhead, initial antagonist Peter Keating quickly works his way up the ranks of an architectural firm by making his superiors redundant in "offering to help". He takes work of other people, and eventually their jobs.

That should work, in theory. In practice, many workplaces are happy to make redundant people not redundant but middle managers.

Former Australian Treasurer Joe Hockey defined a useful dichotomy of Australians as "lifters and leaners" but oversimplified his application of this dichotomy to people who work and pay income tax and people who receive welfare.

For a brief time during the pandemic, I was a welfare recipient, and it enabled me to do some of the most productive and constructive and meaningful work of my career. It is not unusual for welfare recipients to do much valuable volunteer work and community work etc. A lot of childcare is delivered by retirees as a basic example, collecting a government pension. 

Conversely, there are a vast number of well compensated "leaners" out in the workforce. There are an abundance of people who are paid 6-digit salaries, who aren't just useless, but counterproductive. A prominent example might be virtually everyone who worked on Amazon TVs "Lord of the Rings: The Rings of Power" and will be working on the forthcoming second season. But less prominently there are thousands of offices across the world that have people who basically hang out all day. 

We already have "make-work" jobs and have had them since the New Deal, and in earnest since the Soviet Union. (Mexican retail often has elaborate soviet style, pick out the item you want to buy, hand it to one store employee, go pay another store employee, come back and collect your item) and "Overemployment" has recently been appropriated to describe remote workers that work 2 full time jobs, but if treated as an antonym of "Underemployment" where someone cannot get all the productive paid hours they need, overemployment is massive where people are employed for more hours than they can possibly be productive.

My issue with "make-work" is that it tends to only be applied to low-paid, blue-collar jobs. Just as the biggest recipients of welfare are generally corporations, the biggest demographic of make-work employees are elite college graduates. 

The pandemic served to kickstart and grow the number of people who work remotely, but there's two things with this that make Ayn Rand and myself wrong, and Tim Ferris, fuckwit though he is, right.

All the tech was in place to shift workers to remote long before the pandemic hit. Lockdowns just forced it to happen, and employers post lockdowns sometimes got downright nasty in insisting we go back to old fashioned commutes and offices to try and prop up the asset value of commercial property.

Remote work is still likely, very underutilized. There wasn't a big revelation caused by pandemic lockdowns that workers can be hired from pretty much anywhere and work from pretty much anywhere.

So I predict, even AAA game development company, isn't going to lay off 90% of its art department even though they stand to benefit the most from procedurally generated art. Instead, it will probably just make doing VFX for Disney and Marvel profitable again.

Conclusion: Inconclusive

I'm willing to be wrong. What I hope is that the courts rule is that AI art is like music sampling, and IP laws get tightened up. I understand how many artists fear that your skills are being made rapidly redundant, but people's fervant desire to not pay artists for the work they value is nothing new. Since the dawn of time man has yearned not to pay artists. What I fear is that the courts will rule that basically all art is too generic for AI to infringe on IP. That there's nothing wrong with just taking everything uploaded to the internet and smooshing it into art that looks like art. 

I'm optimistic that won't happen, because it will effect the rich, much more than it effects already poor struggling artists. 

I think like facebook, twitter, instagram, TikTok, fentanyl and cigarettes, AI art will prove in the long run to be shit but unstoppable. I think a lot about Gulf War 2, or the Iraq Invasion. It was the first war to have protests from the invading nations before it broke out, indicating a kind of societal learning from Viet Nam, and it seems like society is also learning to anticipate the shitness of the next big thing out of silicon valley. Tech bros occupying a strange place in society where they are treated as both increadibly relevant and incredibly out of touch.

I think the upper limit of imagination is how good AI art could get. To me it strikes me as fundamentally an imitation tool. It can generate weird coral like photo-realistic images that human artists largely don't, but I don't think I've seen anything a human artist couldn't do, both digitally and in traditional medium. The real impact is that human artists can't do it quickly. But any image on a screen with time skill and effort can be reproduced by hand, and the same is likely true of any 3d object produced in real-space. 

It's like fucking creationists that try to make an argument from God by design, and Matt Dillahuntly has to keep explaining that we can only determine "design" by comparing it to the "undesigned" which from a creationist point of view, leaves us nothing to compare anything to, if a pine cone is as designed as a grain of sand is designed as a Casio watch. What then are the hallmarks of design?

In a similar sense, can AI generate any images we couldn't? If a painting of dogs playing poker already drives us mad? I couldn't help but notice that very few AI patrons are using AI to create knock off Jackson Pollocks. I assume someone is, but I'm not seeing. And between a photo realistic image of an African Queen and static, what frontier of imagery can AI take us too?

I'm sure AI could become AGI and start generating original and creative ideas, like something that makes "The Name of the Rose" look classical...but is it going to finally show us what Terry Pratchet meant by Octarine? Will it show us what a married batchelor looks like (a non reform Mormon, the Sultan of Brunei?) Can AI make a burrito so spicy it cannot eat it, or a rock so large it cannot move it?

I don't approve of the people already treating AI art and AI artists with disdain, but I also don't not approve of letting deluded people know that they actually aren't as creative as they think they are.

On the 13th of August 2022 in Roma, Italy David Popovici swam 100m freestyle in 46.86 seconds the fastest time ever recorded. I a nobody, can probably walk 100m in under 46.86 seconds. Usain Bolt has run 100m in 9.58 seconds the fastest time ever recorded, and yet I a nobody can get in a bus and as soon as it hits 40km per hour I'm leaving Usain Bolt in my dust.

My point being, using AI to generate images is something you can certainly do. You can even monetize it, but you are an Uber driver, not Usain Bolt. And you can be sure, techbros in silicone valley are already trying to figure out how to cut you out of the process entirely.

What I hope is that the courts rule, that if you use AI to make "bitter sweet symphony" that not only make you pay all your money to the Rolling Stones, but also pay all your money to Massive Attack. Maybe that's going to far, I would concede that there is an aspect to IP that holds us all back. It just also holds back a lot of fanfic spam too. There are too few Richard Ashcrofts and RZAs that truly upcycle other stuff, and perhaps most importantly: build upon it. 

In that regard, midjourney at least appears to be going backwards, not forwards as they try to ever make AI art indistinguishable from I art.

Monday, April 17, 2023

I don't get the MCU Obsession.

 I don't know what else to call this post. Just...something set me off. A straw broke a camels back. Largely another suggestion that what "we" are all "waiting" or "hoping" "for" "is" "another" "MCU". Specifically Youtuber The Little Platoon's review/critique of Super Mario Bros. movie subtitled "Wasted Potential" with the potential, if I'm hearing it right, is Super Mario Bros. is to Iron Man as Super Smash Bros. is to Avengers. But it missed that mark. 

I don't object to the reasons offered. I object to misty eyed nostalgia for when the MCU was supposedly great. What the MCU has become, was entirely predictable from the first post credits cut scene setting up continuity between films, because spoiler alert it already happened in fucking comics.

Batman War Games

I want to stress that this is just my personal experience, but I anticipate other readers of long running serialized comics particularly of DC and Marvel continuities share a similar experience. The year was 2004, Shrek 2 would be the highest grossing film of that year, but Spiderman 2 the one with Doctor Octopus, not the one with Electro had just come out so the MCU was not foretold or foreseen. 

Batman had a title - "Batman: Gotham Knights" (BMGK) that I was interested by. I was interested because interesting things were happening. Specifically, characters I found interesting, were doing interesting things. This intrigue began in BMGK #50, with a new creative team and Edward Nigma, aka the Riddler, crashing through the glass ceiling of a ballroom during an event.

Spoiling Jeph Loeb and Jim Lee's Batman run "Hush", the Riddler did it, and he played Hush in order to do it. Now Hush was back and pissed at Riddler. Not much interesting there, but between BMGK #50 and BMGK #55 the premise - Hush's back and pissed at Riddler for using him - got utilised to do interesting things. Namely, the Riddler suddenly needs protection and so he cashes in his trump card - which was that he knows who killed the Joker's family, he gives Joker a name in return for protection, tying it back to works like Alan Moore's "The Killing Joke" where the Joker is established as an unreliable narrator for his own origin, an idea handled with all the subtlety of a sledge hammer by Nolan in The Dark Knight.

Anyway, Joker and Hush have a confrontation in BMGK #55 after interesting developments in Joker's character happen for the first time in years, and Joker is defeated and driven out of Gotham City. The issue left me with the Joker crawling back to an abandoned circus, the circus hideout/lair from "The Killing Joke" where he is welcomed home by a side-show freak. So we had the deposing of the clown prince of crime, alliances between old school villains against new school villains and the need for the old-school to re-invent or reassert themselves while having new developments or new interpretations of the most iconic villain in all of comics (2 Oscars).

And then! A fucking cross over event. Specifically "Batman: War Games" so it isn't even the most egregious crossover event compared to something like DCs "Crisis on Infinite Earths" "The Final Night" or Marvel's "Civil War II" or "X-Men vs Avengers" etc. etc.

Batman: Gotham Knights, was but one of eight titles sold under the Bat-family brand. War Games was a crossover event, told over 3 issues in each of those titles. So arguably, the plot of War Games is fairly interesting by crossover event standards, but you needed in theory, to buy or otherwise read 24 titles over three months to follow the story - specifically:

hideTitle(s)
Batman #631-634
Batman: The 12 Cent Adventure
Batman: Gotham Knights #56-58
Batgirl #55-57
Catwoman (vol. 2) #34-36
Detective Comics #797-800
Batman: Legends of the Dark Knight #182-184
Nightwing #96-98
Robin (vol. 2) #129-131

And the total is less than...forget about the fucking sum, it is less than an individual part. 

In Theory...

In theory cross over events should be exciting. Not for the reasons popular intuition might suggest, pages in a comic book and screen time in a movie is pie. The more characters the less screen time, panels per character. 

Adding characters means more time for a character you like iff they are the character added to the title. Right, so it might be more entertaining for you if Leatherface showed up at the end of Little Women, but it's less entertaining for you if the Little Women show up in Texas Chainsaw Massacre. Then again maybe not, if you viscerally hate Civil-War era teenage girls. 

It's more that taking disparate ideas, "what if?" scenarios can work. Remixing, reworking and re-whatevering existing intellectual products can yield interesting outcomes.

Now there is nothing inherently bad about the idea of a crossover, in theory. In practice however, this isn't an extremely skilled writer or editor going 'I've got this wild idea that might be interesting...' An example of that might be Frank Miller's "The Dark Knight Returns" the template for "Batman v Superman" where the final act is Batman v Superman. The difference being that Miller had the clout to create an isolated title liberated from the rest of DC continuity. Other examples being Alan Moore's League of Extraordinary Gentlemen, a Victorian literary crossover event, and of course Watchmen, or Warren Ellis' Planetary series. 

In 2000AD comic series "Trifecta" was a surprise crossover between three seemingly independent titles. 

Here is the basic foundations on which a crossover "event" can be successful:

First, someone has to want to do it, at the creative level. Someone needs to have an idea of a story to tell, and that pushes the project.

Second, it helps to do meta-commentary or using euphamistic characters than a story that effects a canonical comic universe. For example, Watchmen and Planetary.

Third, it helps to remove the crossover event from main-title continuity. "Justice League: New Frontier" is a good example.

Fourth, a crossover event that can be kept relatively local, or to a subset of the comic universe will work better than "Everybody vs Galactic Genghis Kahn and his Omega Device." For example, Batman and it's spinoffs are more popular than most of the rest of the DC Justice League titles, so Bat-Family crossovers tend to be better stories than Justice League events. Similarly, Spiderman's popularity is such that there's a Spiderman for girls, and a Spiderman for African Americans, and even earlier a Spiderman manga serialization. However, I don't really know of any Spiderverse crossover events that I could compare to Marvel events.

Importantly though, the four things, though I'm sure non-exhaustive, are mutually exclusive to the main reason crossover events are done. That reason being $$$.

The Economics of Crossover Events

Once a month you release a Batman comic, and Batman is popular, say you sell 100,000 copies a month. It would not be surprising that a second monthly Batman title sells more than the main Superman title. Then you have Aquaman a character so unpopular his name is identified by Google Chrome as a spelling error, it sells maybe 30,000 copies.

So you have a title that sells well. Like Batman, and one that generally sells poorly, like Aquaman. So, what you can do, is begin a story arc in Batman, that ends in Aquaman. You tell the reader if they want to find out who's stealing all the fish in Gotham City Harbour, they need to buy Aquaman #17. 

The publishers leverage popular titles to sell unpopular ones. Crossover events are variations of the upselling sales tactic. "Would you like Aquaman with that?" because selling more to existing nerds is a lot easier and cheaper than attracting new customers to the store with new titles.

The problem is, that crossover events suck. There's a curious incongruity between how excited/nostalgic people are for the MCU Avengers movies, and that super-hero team ups pretty much never make anyone's recommendation lists for comics. Patrick H Willems for example, list of recommended comics to get you into comics contains no collections of crossover events* and Patrick is someone I would guess I subjectively disagree with on most matters, by contrast Chris Piers of Comic Tropes is someone I love, and feel I would subjectively agree with, even defer to, on comic book matters, since around 2017 he has released "top ten" lists for comics that came out in those years with 2019 conspicuously absent, his lists almost never feature cross-over events or super-hero team ups.

I suspect, this general phenomena is because crossover events are sales events, marketing promotions masquerading as stories. Mechanistically speaking, these events are when popular titles get diluted with crap. It is the time of year when Batman is obliged to take Aquaman or Green Arrow along, when Spider Man has to visit the Fantastic Four and Jessica Jones and Luke Cage have to hang out with characters that wear costumes.

There is another economic principle that needs to be addressed in crossover events that relates directly to dilution. That is intermittent reinforcement if you have a story spread across 8 titles, you can't  have the mastermind revealed in every issue, much of it is going to be filler. Smart money is on the big reveals being not in the popular titles, but it can be intermittent.

A good example is in the context of the MCU and what has become a staple of mid-credits cut scenes, which are predictable enough in and of themselves, there will be a mid or post credit cut scene, but what is intermittent is how rewarding that scene will be. Will it involve a popular character like Iron Man or Captain America or will it tease the coming of nobody/s favourites Quicksilver and Scarlet Witch? Will Thanos have a revealing conversation with one of his minions, or will Bucky Barnes be revealed to have been recuperating in a Wakandan hut?

I cannot exclude the possibility, that such nerds exist who will pay money to sit through two hours of a Captain Marvel movie, or Shang-Chi and the Ten Rings just to get the payoff teaser of the cut scene. I mean, it is common enough that trailer's are more entertaining than seeing the whole film.

As Gabor Mate has said of addiction, "it isn't the having it's the getting." Where the best feeling in an addiction cycle isn't the high, but scoring the drug. Anticipatory pleasure. The intermittent rewards are the suggestions that something good is coming/going to happen, there need not be anything substantially good in your hands.

I can nest here, to this last point, a prediction which is that if a post credit scene in a forthcoming MCU movie such as "The Marvels" as at writing, teased the return of Robert Downey Jr. in the role of Tony Stark, nerd-fans will extract more pleasure from the announcement of the return than they did from RDJr's cameos as Iron Man in the Tom Holland Spiderman movies, or his last performances as a lead character in End Game, Infinity War, Captain America Civil War and Iron Man 3. And almost certainly more excited by the announcement of his return to the role, than to watch a 150 minute film about Iron Man vs Titanium Man, or Fin Fang Foom, or Modok.

Why Crossover Stories Suck

It's relatively simple to come up with a plot that crosses Batman into an unpopular title like Aquaman you know someone's stealing all the fish in Gotham City, Batman hits up usual water-themed suspects Penguin and Killer Croc but they aren't the culprits, the thief is coming from inside the Ocean! Batman needs to enter the jurisdiction of Atlantis.

Whatever. Crossover events suck when DC or Marvel attempt to boost the sales of their entire title range. This is when we get plots that are "Mega-something from dimension 8XKL9 threatens the very existence of the universe as he attempts to assemble xteen artifacts from across the world to create the Omega device."

We could call it the "Drop everything threat" and it needs to exist for logical reasons. You probably can't sell a crossover event which is Spiderman suggests to the Avengers and X-Men that they join a foosball league. (This would probably be far more interesting than an attempt to destroy the multiverse.)

In Gotham city though, it's seemingly every other week that Joker has used his Joker juice to create an army of jokers, or scarecrow has found a way to distribute his fear gas etc. Hence you need to concoct essentially a trolley problem to take Batman away from his usual jurisdiction into outer space or whatever. 

The intuition that massive crossovers of the Avengers or the JLofA appears to be "more is more" and admittedly, I would say that anyone who wants to conduct a focus group with the specific consumer demographic of nerds are wasting their time. We know what nerds want: nerds want more. Nerds are in the full grip of addiction. It's the idea that it is mathematically more interesting for a busload of passengers to be in danger, than for a character known and cared about to be in danger.

A great example of this principle in practice is an event in the fantasy series "The Wheel of Time" by Robert Jordan, even though it is not a crossover event. At some point the major protagonist Rand Al'Thor teams up with another character Nynaeve to use two giant magic statues to cleanse the evil out of magic that makes male magic users go insane in the story. It's a real game changer. One of the gigantic statues is mostly buried on an island and it holds an orb that starts glowing when the two protagonists cast their mega spell. The inhabitants of the island misinterpret the orb glowing as a sign of the end times and commit mass suicide. 

The author Robert Jordan, was allegedly shocked that his readership, did not give a shit about the mass suicide. I found a reddit thread that can substantiate it here, and it helpfully links to the author's blog post where he expressed his disappointment that people were so blasé about unnamed unnumbered characters of quite literally no consequence to the story dying for no reason.

Are you familiar with the dirty joke "a boy fell in the mud." Crossover events and their necessity to create a premise large enough to have a blind acrobat in New York and space travelling intergalactic band of misfits team up with someone who can shrink down and communicate with ant queens, work off the idea that "two boys fell in the mud" is twice as funny as one boy falling in the mud. 

Crossover events suck in terms of writing because they are first and foremost sales promotions, an attempt to get marginal revenue to equal marginal cost. The story is fitted to spec, if it is ambitious at all it is to offer a narrative excuse to retcon up a bunch of garbage that has accumulated over a number of years as a result of having numerous creative teams working across numerous titles.

Marvel's "Civil War" crossover event could be regarded as interesting. Interesting things happened to interesting characters. Like Peter Parker revealing his identity, meant there was a slew of comics in the Spider-Man Titles where we could read the repercussions.

The trouble is, many interesting changes are not sustainable in comics. Eventually, Spiderman needs to go see a wizard to cast a spell that makes everyone forget his identity. Barry Allen dies in one crossover event, only to be brought back to life in another crossover event. 

And worse, if you write a good interesting crossover event, it will likely sell well. Sales are most addictive of all, and that gives us bad crossover events like "Civil War II"  

The famous Mr. Show sketch "The Story of Everest" is a form of rake joke, and in the creators' commentary track its revealed that the knocking over of the thimble collection was done more times than shown in the sketch, because the audience stopped laughing. They had to calibrate how much escalation worked. 

The Permanent Sale

In my marketing degree, I was taught about a trap. You are a retailer, and before your annual stock take, you decide to run a sales promotion to try and reduce inventory and make the stock taking exercise easier. You paint on your windows "Buy 2 get 1 free!" and it's a huge success. People take the opportunity to stock up on widgets.

However, you notice a few things. The first is that the month after the sale, sales drop off. The second is that the next year you notice sales drop off in the months leading up to stock-take. It appears your customers are anticipating another sale.

Anyway, skip a few years and you find yourself more or less holding permanent sales. There's still the end-of-financial-year big sale right before stock take. There's the post-christmas sale, the back-to-school sale, and a complex schedule of sales promotions on certain items that run year round. 

Permanent sales are the end result of becoming addicted to sales promotions. Both businesses and customers build up a tolerance to sales promotions. I do this with Steam games, there's no video game on my Wishlist I can't wait for to be 50~75% off. Online retailers Drjays and Yesstyle I've never known not to advertise a discount coupon code on the landing page.

Now, there's a stark difference between DC and Marvel, and that difference is Batman. Batman has his own sub-universe with at least 8 titles. Marvel's closest equivalents would be the X-men and Spiderman properties that generally only have 1 or 2 spin off titles at any time. However if you take a fucking gander at the list of Marvel's line-wide crossover events you can see there's two events in the 1980s, three in the 90's, four in the 00's, TEN in the 10's, and in the less than 3 years of the 2020's we are up to four line-wide events. 

DC's addiction can be perhaps best apprehended in the title's of it's crossover events "Crisis on Earth Prime""Crisis on infinite Earths" "Zero Hour: Crisis in Time" "The Final Night" "Day of Judgement" "Joker: Last Laugh" "Identity Crisis" "Infinite Crisis" "FINAL Crisis" "Blackest Night" "Forever Evil" "Batman: Endgame" "Superman Doomed" "The New 52: Futures End" ... "Dark Crisis on Infinite Earths"

I saw many a culture-critic youtuber point out the bloat of the MCU in Phase 4, where between the Disney+ TV shows and movies the claim I heard was that they put out more content than the previous three phases put together. By my count phases 1-3 put out 23 movies, and maybe the TV show Agents of SHEILD counts too. Phase 4 has 15, 7 movies and 8 TV shows. This is unremarkable, given that the parallel source material of comic book publication universes tell us exactly what was going to happen. 

These are after all, adaptations.

Virgin, Ryan, Saw

The 20th century saw air travel invented. It began as a luxury with attractive stewardesses serving cocktails while you stretched your legs out to smoke, to something worse than bus-travel in every way but the distance covered over time.

For me, it was somewhere between Saw III and Saw IV, that a movie became a TV show. Saw the original is indubitably a movie, Saw V and VI are episodes in a TV series that was released at the cinemas. 

Saw is a fairly famous franchise, the initial movie involves two men trapped in a dilemma, with a satisfying twist. After the first movie, the franchise becomes increasingly formulaic, there is a cold open like many TV series of someone in a survival game where they have to chop something off or die, or chop something off and die. The bulk of the movie will be a series of these survival games, while someone tries to figure out what goes on, and the last act will be the final survivor living through or dying as it is revealed to them the minutes dedicated to unresolved plot developments.

It's this format of watch a story that is essentially prelude of getting a piece of a story. I assume, with no real knowledge of TV production, that it is just standard practice that you get people to watch individual episodes in order to tease a season finale. Be it House M.D., The Mentalist, Dr Who or Seinfeld. 

The payoff for a TV channel is that you tune in to the finale and watch the sponsors adds, also you watch an entire episode during the regular season including the ad breaks in order to get hints as to the content of the season finale.

I put it to you, that "Phase" is just a euphemism for "Series" and an important aspect of this business model, is that the finale doesn't have to pay off the hype. The Saw series, literally goes fucking nowhere, in the back of my head is a distorted mess of innumerate Jigsaw-apprentices. Does Hoffman die at the end of Saw 3D or not? Why do I care? I don't.

The Unique Problems of the Film Medium

The first is that actors get old and die. Comics do not face this problem, and both DC and Marvel have a Simpson's style continuity, Bart has been 12 years old my entire life. The South Park gang are still in primary school as they've lived through Harry Potter and The Lord of the Rings, the Obama Presidency, the Garrison Presidency, Nintendo Wiis and iPads and World of Warcraft. 

Batman's parents were murdered shortly after attending a screening of a Zorro film, which places their deaths anywhere between 1920 and 2005. Judge Dredd ages in real time but you know he has a bunch of robot parts and lives in a future where his life and quality of life can be extended. He is currently 73 years old, which gives hope for a future instalment with Karl Urban.

However much anyone liked Michael Keaton or Christian Bale or R-Patz, it's an athletically demanding role, and these are big corporate owned IPs. This isn't like Rocky Balboa a creation of the lead actor who can reprise the role pretty much whenever he likes. 

What this sets up, is when the DCCU flops hard and essentially disintegrates, you can have Ben Affleck, Michael Keaton and R-Patz all playing Batman concurrently. Meanwhile over at the MCU, the main Avengers roster is all retired, at the least the two most charismatic leads - Evans, Downey and Johansen. Spiderman, owned by Sony, is being played competently as a naïve young man.

Which obliges the MCU concept to either conclude, in order to reboot Iron Man, Captain America, Black Panther, Black Widow, The Hulk after some interval. (Consult Batman at this point) Or bring on the second and third stringers - Shang Chi, Moon Knight because we all remember Shang Chi and Moon Knight when we were growing up, Marvel and DC have a whole bunchola of characters in their stable, but popularity breaks down in line with Pareto distributions - we would predict 1% of the characters to account for 50% of the sales, 20% accounts for 80%, 80% of the characters only 20% of the sales and 50% of characters accounting for 1% of the sales. The 2000s where the time of Marvel's 1% shining - Tobey McGuire as Spiderman.

Having sold off both X-Men and Spiderman rights as Marvel's two most popular properties, the MCU had no choice but to turn to the avengers. Phases 1-3 mark the 20% bracket selling 80%, and phase 4 brings the stable into 80% that sell 20% bracket. We can see it in the proliferation of titles and the diminished revenues and interest. 

The brilliance of the MCU concept, is behind us. The first piece of brilliance was figuring out a way to put Robert Downey Jr into 9 films instead of 3 like Christian Bale in Batman. It is an achievement of mileage. The second brilliant payoff is getting $1bn+ at the box office for an unpopular second or third tier character like Captain Marvel. That's turning movies into TV.

Why Would Anyone Want A Nintendo Universe?

I mean I can see why a studio would want to emulate the MCU while the MCU was working, before the fever broke. But a consumer should want a series of good films, and that is more likely to occur when you can exclude misses and only count hits.

Iron Man, The Incredible Hulk, Iron Man 2, Thor, Captain America: The First Avenger, The Avengers.

Pretty good right? Well, the Hulk movie arguably has been forgotten, Iron Man 2 wasn't as good as Iron Man, Thor? 

Here's another list: Kramer vs Kramer, The Amityville Horror, Rocky II, Apocalypse Now, Alien and 10. 

Malcolm Gladwell would say I'm nostalgic for when films only featured white people... even though phase 1 was extremely white. But in 1979 moviegoers got a bunch of good films, even some all-time classics with them being largely independent of other movies.

I feel I'm in a minority, that feels that books and comics and video games are legitimate mediums in which to tell a story, and do not need those stories to be legitimized by "the Movie" suffix being a box office success. 

However, I can conceive that someone could make a great Zelda movie, you know what wouldn't make a great Zelda movie better? A post credits cut scene where Mario and Luigi tumble out of a pipe into Hyrule and ask "where are we?" 

You know what isn't exciting at all? If in a mid-credits cut scene of the Donkey Kong movie Diddy Kong peels a banana to discover a golden ticket with a Bowser insignia inviting him to compete in a Super Smash Bros. tournament.

You know what would be fucking shit? If Starfox and friends are attempting to travel through warp space when an anomaly sends them crashing into Bowser's Arena.

It's probably no surprise that I never got Super Smash Bros. games either. My experience of them was that they are a messy conceptual nightmare. I got no thrills from seeing Princess Peach and Kirby finally square off in combat.

A Nintendo cinematic universe doesn't have the problems the MCU has, namely, Seth Rogan and Chris Pratt can phone in their performances for 20 or 30 years as voice actors. Voices age slower than bodies. You don't have to pay someone to train for 6 months intensively to get in shape for a role. Furthermore, if you do have to replace voice talent, it's a lot easier to stick another voice up a 3d puppet. (Patton Oswald replacing Louis CK in "The Secret Life of Pets" due to the masturbating in front of women).

But just make Mario movies, I haven't seen the current one and probably won't, I'm not hungering for a sequel. Make a Zelda movie, maybe I'd check it out if the trailers promised something. Don't make a Kirby movie, because there's not a story there. Make a Donkey Kong Country movie, because there probably is. 

And sure, if in 5 years the dollars just makes too much sense, do a crossover movie of all the Nintendo properties having to fight in a tournament to stop video game land from being destroyed. You know, like Wreck-It-Ralph. But there's no need to tease a Nintendo Cinematic Universe. It can easily be of no consequence just like in Nintendo games, Zelda Tears of A Kingdom is unaffected by who canonically wins Super Smash Bros. Ultimate. Nor does it effect Super Mario Galaxy or anything there.

What makes me want to cry, is the idea that there's an MCU shaped hole in the world, that needs filling.

There isn't. Try to remember life before the MCU. You know, when people other than Koreans and Tarentino made good movies.