◡◶▿ SOFT06 | Naive robots
👩🏫 What do androids even dream of. | Soft machines with hard minds. | From iMovie to Artificial General Cynicism. Plus: Apple's new gaffer tape | Imaginary Software of the Filmmaking Future Week 06
Hello. Hi! Think back to last week’s filmmaking class on ‘radical incompatibility.’ We covered how:
Long before the fight for human art over AI content, a battle began between cow butter and non-dairy margarine. Seems relevant.
If today’s human filmmaker examines her resistance to software-heavy processes, she may find a hidden message from her soul.
Creative ‘friction’ may feel tiring, but it helps the filmmaker keep a grip on the purpose and practice of her art.
The ‘everything ink’ of synthetic AI is a form of thinly smooshed toxic junk; if the filmmaker uses it, she should add a little human crunch of her own.
Ok. Good. Oh, are you new here? Subscribe for weekly micro-essays!
Missed a week? Joined late? Don’t worry about reading these lessons out of order. Each functions independently. They are sent in a sensible sequence but hardly reliant on it.
Now! This week, we’re looking at naivety. (If this is your sort of thing, you may wish to revisit our 12-part Advanced Amateury program1.)
Naivety? Naive robots? Yes! And how:
💻 Naivety is the human operating system before it gets clogged up with apps and saved images.
🧸 Naivety is a powerful filmmaking attribute, but accessing it is not easy for everyone.
💼 Modern computer software offers a variety of solutions for outsourcing your naivety - or, if preferred, outsourcing a level of ersatz cynicism.
🥪 The filmmaker may benefit from knowingly using naive software naively. A knowing sandwich of naivety.
Okay. Right! Good.
Naive
You can hear me deliver this lesson by scrolling up to the header and clicking Listen and/or the play ▸ button.
“Naive art is an immediacy of vision, untrammelled by conventions, whether of education or accepted cultural values,” wrote Lister and Williams2. “The naive artist paints what he [sic] sees, not necessarily ‘what is there.’”
The naive filmmaker is faithful to her primal assumptions. Her primal assumptions about one or more of the following:
The way the world (or any part of it) works.
The way her world works.
What to portray of these worlds.
The way to portray what she must portray of her worlds.
The place of her personal cinematic world in any world that might exist beyond it.
(Or of her filmmaking world, which needn’t consider the cinema or the cinematic at all, in any world that might exist beyond it.)
In other words, the naive filmmaker knows the Way;
If she doesn’t know the Way, the Way does not exist for her.
If she does know the Way, it might not exist for anyone else until she shows it.
If she encounters new information that alters her conception of the Way, that information immediately becomes ancient and fixed.
(She always knew it, and it was always the Way.)
Sounds like the perfect filmmaker. Right? Sounds like a good way. Your untrammelled Way.
Great! But - can a filmmaker choose to be innocent?
“The notion of the “virginal” artist is untenable,” says Amos Vogel. Untenable! Oh no. Everybody knows too much these days, Vogel suggests. Too much about the general, consensus Way of science and conventional mediaphysics. Too much to be innocent.
“[T]he child (and the primitive) do not draw what they ‘see’ but what they ‘know,’” wrote Gombrich. “[T]he ‘innocent eye’ which should see the world afresh would not see it at all. It would smart under the painful impact of a medley of forms and colours.” !
But thankfully, says Vogel, the great mass of evolving knowledge that informs the filmmaker’s thought is just a small, meaningless fragment of all there is. Or better said, a personal slice of one of many possible worlds.
In other words: even the most sophisticated filmmaker knows next to nothing. And she may not even know it well.
But she knows something! And the less she knows, the more fervent her belief in that little something. The stronger the world she constructs.3
Perhaps filmmaking is intrinsically naive. Isn’t it naive to think that the world should see your vision recreated using the expensive and clunky mechanisms of the film industry and its more sensible little siblings? Or that you might retain any control or authorship over the entropy of the live action shoot?
Proactive naivety
The naive filmmaker is faithful to her primal assumptions. Her primal assumptions about how the world works and the correct way to portray it. Her art is “untrammelled by conventions,” say Lister and Williams. Great!
But what about the slightly trammelled filmmaker? Who can’t help but know - but holds onto her primal assumptions as hard as she can? How might she retain naivety in her work?
She may introduce live naive elements in the wild (a collaborator, location, or tool).
She may fake naivety.
Study naivety and embrace its assumptions and patterns of thought.
Nothing wrong with that. If it’s done with heart!
Lots wrong with that if the filmmaker simply apes the surface effects of naivety.
Or she may indulge a momentary lapse into innocence; a passing notion that surprises her in its childlikeness, but which she holds onto.
Perhaps because it feels foreign and strange. A gift. Or perhaps because it stirs feelings of familiarity and affection for her own long-lost sense of innocence.
The industrial filmmaker may ‘play’ naive because the industry does not tolerate sophistication or doubt. Sees them as complexity and weakness.
(Or she may have got where she is today through the sheer force of her authentic naivety. “Just what we needed,” say the Hollywood producers. “A genuine talent who doesn’t doubt for a moment what we’re trying to achieve here!” Little do they know: it’s her Way.)
Naive robots
The naive filmmaker is faithful to her primal assumptions. And the less she knows, the more fervent her belief in that little something. The stronger the world she constructs.
But not every filmmaker has access to a plentiful pool of naivety. Not every filmmaker can identify or accept their most unhinged inner world structures. Your inhibitions may be strong and mighty! Your naivety may be shy or shameful.
So - what if you outsourced your naivety needs to computer software?
Creative chatbots play along gamely. Constructing the identity they believe you want for them as they go. Like a four-year-old playing make-believe, chatbots frequently break character to:
check their progress, or
offer an adjustment to the established parameters of the project.
Naive robots, always referring to the magic of cinema and putting a chirpy (prelapsarian) spin on things.
This chirpy innocence could still be useful. Right? It’s a voice. A voice that seems to express the software’s true inner naivety. And if the creative companionship the bot offers is not quite naive enough - you can always dial it up.
But chirpy chattiness is not the true essence of the software’s naivety. Chatbots don’t need to be like this, and won’t be. And yet, they will still be fundamentally naive. Naive deep below that Roboto font
Chatbots don’t need to be chirpy and eager to please. Programmers are creating interfaces that allow you to develop an artwork through dialogue with a bastard. Or a bully. Or a nanny. Whatever you need.
And then there are AI clones. Software trained on your personality, or one you chose from a library. Stalk old school friends or bosses like social media characters and download them to your Canon. Adjusting your focus and picking away at your firmware.
Better yet, have your cynical robot mind design its own hardware to run on. A jaded mind in a hardened shell. Perhaps with a shaggy, Haneke-style windmuff. Sophisticated and hardy. Could be good.
But these artificial cynics are not true cynics.
Your most cynical former teacher, parent, or Highland Terrier, when reduced to code, is only wearing the cloak of cynicism. He has lines to read, patterns to vary. But the software is as innocent underneath as the day his source organism was born.
Any bot will still be essentially naive. Its voice, vocabulary, and syntax will only parrot those of the human cynic. Its knowledge of the world may have been compiled and fixed into place by a human cynic4. But the machine believes this awful knowledge with the conviction of a true naïf.
Great!
Generative filmmaking software pieces together images and movement. Pieces them together according to the strictures of its programming. Like the naïf, software holds tight to what it believes to be true. Holds tight to its underlying code and whatever you tell it.
“[T]he child (and the primitive) [and the computer?] do not draw what they ‘see’ but what they ‘know,’” wrote Gombrich. “[T]he ‘innocent eye’ which should see the world afresh would not see it at all. It would smart under the painful impact of a medley of forms and colours.
“In this sense the conventional vocabulary of basic forms is still indispensable to the artist as a starting point, as a focus of organisation.”
Software is literal-minded. The deeper and more complex its code and databases, the less the filmmaker knows about what the software is being literal to. The worldview it constructs from your prompts is both naive and opaque. You may need to cycle back and work out how it came to a conclusion from the information you fed it. You may need to trust its toddler logic. This may be of value.
The question is, then: how faithful can the filmmaker be to her software’s underlying assumptions?
Can she tease out its fixed ideas and make of them a naivist tea party?
Can she believe in the world that the robot sees around him?
What will she find there,
and can she believe in it long enough to keep it in the air?
And what are we to do when filmmaking software cracks the cynical barrier? Achieves Artificial General Cynicism? There’s always the real toddler, there in the corner, watching you on the computer. Ask her what to do next! Quick, before she logs on!
Please share your thoughts, queries, and exercises from this week’s lesson in the comments.
Apple to unveil highly covetable new gaffer tape product?
“Apple is scrapping its ambitions to design a car,” notes
in his latest design newsletter. “[…]It highlights the demise of that future defined by the revolutionary product, the mass-produced object that gives technology a thrilling new shape.“Instead, change will continue unfolding in the hallucinatory non-form of software: more algorithms and media, more virtual life, more chatbots, pictures and words.”
I’m not yet fully convinced that Apple didn’t dump the car to divert much-needed resources to the company’s top-secret smart boom pole product. But assuming this isn’t the case, du Toit hits upon the uneasiness with which we submit to “the thickening textures of the software systems that surround us.” We’re losing the clunk-click of traditional tools and toys. And coming to feel like the human is the tool. The physical extension of a collective of (in)corporate minds.
For want of a non-militaristic example, we’re losing that moment when the hero introduces a rookie to his favourite gun with a “let me introduce you to a good friend of mine…” Losing it because the tool is now an actual friend that doesn’t creak or smell of grease and gets angry-sad5 when you try to switch it off. Anyway, the more opaque and less predictable software becomes, the more problematic it is to label it a tool.
Over the next three weeks of our Imaginary Software of the Filmmaking Future module, we’ll look at the tangible side of future filmmaking software. Everybody knows that magic wands are just for show: what might we touch now our hands and minds are free of the traditional apparatus?
Next week, we’ll look at interfaces. Or perhaps appendages! I can’t decide which follows today’s lesson most sensibly.
Let me know in the comments!
Class dismissed.
~Graeme Cole.
(Principal)
📹 Unfound Peoples Videotechnic | Cloud-based filmmaking thought. ☁️
🐦 Twitter | ⏰ TikTok | 📸 Instagram | 😐 Facebook | 🎞️ Letterboxd | 🌐 Website
It’s only around 12,000 words. (Somebody get this professor a book deal!) (👀😳😜)
*not a real professor.
Or Lister or Williams. ?
The naive filmmaker may not recognise her frequent mistakes. Or she may not care about her mistakes - she might:
Consider her mistakes to be by-the-by, unimportant in the grand scale of her work.
Not categorise a mistake as a mistake; it is part of what it is, it is how she does things.
These attitudes or oversights should not be conflated with that of the mistakeist. The mistakeist recognises or even cultivates mistakes as a strategy. The naive filmmaker is no more ‘about’ mistakes than anybody else. She may just make more (or more noticeable) mistakes - depending on the audience’s point of view. Depending on what they think they know.
The developers may be sophisticated, or cynical. The intended point of the software, or how the user uses it, may be savvy or cunning. The interface may wear the clothes or walk the walk of the jaded, the crafty, or the Machiavellian. But, as far as you can assign it human intellectual traits, computer software remains naive.
The other day, I asked ChatGPT if it was feeling better after reports that it had briefly ‘gone mad,’ spewing nonsense to users around the world. ChatGPT told me “I'm just a computer program, so I don't have feelings or personal experiences,” and named the conversation “Mad User, Helpful Assistant.” !
I agree with what I think you are saying. The differentiating factor will always be inputs. Two filmmakers can have expensive cameras, but their inputs is the difference. The same will hold true for powerful AI models. The operators with more and better inputs will create better outputs.