[ recap ]

Last week, we looked at George Spencer-Brown's Laws of Form and James Baldwin's The Creative Process. Both arrived at the same place from opposite directions: the creative act begins by drawing a distinction, marking one thing off from another.

[ thoughts ]

This month Meta decided what "original" means, and started enforcing it.

Facebook launched Creator Fast Track, paying creators up to $9,000 over three months to post. Instagram changed its algorithm so accounts that repost without "meaningful transformation" are no longer shown to anyone, except followers.

Both were announced as a gift to creators: the aggregators who made money reposting other people's work would finally lose their reach, and the people who actually made things would finally get the credit.

Aggregators stealing work was a real problem. Creators welcomed the change.

Look at what the tool became, though. To enforce "original," the platform had to define it, and the definition is whatever a machine can check:
Wholly created or materially altered, no watermarks, no speed changes. Commentary, parody, and voiceovers count.

Meanwhile the same Instagram added a button to its Edits app that generates video from a text prompt, and that synthetic clip counts as your original content.

The platform now penalises you for reposting a human's photograph untransformed, and rewards you for generating a video that no human touched.

So the machine decides what original means, judges whether you cleared the bar, and sells you the tool that manufactures the synthetic thing the rule was supposed to guard against. All at once.

In 1973, Ivan Illich published Tools for Conviviality. Illich was an Austrian philosopher and former priest who spent his life studying how institutions turn against the people they claim to serve.

His argument: every tool has a threshold. Below it, the tool serves the person using it, extends their capacity, stays under their command. Above it, the tool inverts. The person starts serving the tool, reshaping what they do to satisfy what it demands. Illich called a tool that stays below the line convivial. Past the line, the tool commands you.

I see the platform crossing Illich's threshold: someone at Meta decided that "original" should mean whatever a classifier can detect, because a classifier runs on a billion posts and a human editor cannot. That is the trade-off. Judgement does not scale, so judgement was replaced with a check a machine can perform.

As a creator, you now rearrange what you make to satisfy a definition written for a machine to read. Some creators are glad, because the aggregators who stole their work are finally losing reach. Some are upset, because the same system purges them by mistake and cannot see what they made. Both are right. The platform did not stop serving creators by accident or by malice. It stopped because serving them well would have required someone to keep looking, and looking does not scale.

[ creativity ]

So what is the thing the machine cannot define for you?

Richard Hamming spent a whole lecture trying to answer that.

Hamming was a mathematician at Bell Labs, part of the Manhattan Project, the man whose error-correcting codes make modems and CDs and satellite links possible.

In 1995, he taught a course at the Naval Postgraduate School called The Art of Doing Science and Engineering. Session twenty-five was on creativity. He opens by separating two words we treat as the same thing: original and creative.

Original just means not done before.

Hamming’s example:
Multiply two random ten-digit numbers and you have made something genuinely original, a product no human has ever computed. It is also worthless. Not done before, he says, is hardly enough to make anything important.

Originality, on its own, is nothing.

Creativity is different.

It is, he writes, "usefully putting together things that were not perceived to be related before," and the measure is the psychological distance between them.

The further apart two ideas sit in your mind, the wider the gap, and the more creative the act of joining them. That gap is the whole of the work. His own example: he switched the way he measured the distance between two strings of bits, and it turned out to connect to Shannon's information theory. The math was easy, but crossing the gap was not.

So how do you cross it?

Saturate yourself in the problem, think of nothing else for days. Get emotionally involved, because a detached mind only finds the conventional answer. Then drop it. The temporary abandonment is essential. It lets the subconscious find the approach the conscious mind kept missing. When you are stuck, he says, ask what a solution would even look like. And underneath all of it, the long habit. Turn every new piece of knowledge over as you learn it, look at it from many sides before you file it away, so that years later it is retrievable as an analogy.

He watched the statistician John Tukey do exactly this and realised Tukey simply had more hooks into more knowledge:
"We are, in a very real sense, the sum total of our habits."

Now hold a generative model against that definition. It has every field already loaded and connects them instantly, which means it crosses no distance at all. And the distance was the thing that made the act creative.
It cannot saturate and then drop a problem.
It has no emotional involvement to commit.
It cannot change its habits (it has none).

What it does effortlessly is produce the not-done-before at infinite scale, which Hamming already told us is the worthless part. The thing it cannot do is the thing he says creativity is.

[ making ]

In 1975, Brian Eno and painter Peter Schmidt published Oblique Strategies, a box of cards that came out of watching how they actually worked. Each card carries one short instruction. "Honour thy error as a hidden intention." "What would your closest friend do?" "Use an old idea."

The purpose, in Eno's words, is to derail normal thinking habits when they have stopped working. You draw one when you are stuck and apply it to the problem in front of you, however little sense it makes.

Twenty years later he kept a diary for the whole of 1995, published as A Year with Swollen Appendices.

Most of it is mundane on purpose. Lunches, arguments, the school run, studio sessions that went nowhere. But the year it documents is the year Eno was rewriting Oblique Strategies, so the diary entries are filled with new strategies caught in the moment they occurred to him. "Steal a solution." "What else is this like?" "When is it for? Who is it for?"

You watch the tool get made in real time.

Eno argues that pretending is the most important thing we do, because it is how we run thought experiments, how we find out what it would be like to be otherwise. Oblique Strategies are a way to pretend your way past the place you are stuck.

The deck does not solve your problem. It hands the problem back from an angle you would not have picked.

You still need to do the work, which is exactly what makes it convivial in Illich's sense. It extends your judgement instead of replacing it. It is the opposite of a tool that produces the finished thing so you never have to cross the distance yourself.

Take something you are genuinely stuck on and draw a card. Try to apply it, especially if it seems irrelevant, because the analogy that seems unrelated is often the one that gets you unstuck.

Did it make you throw something out, or add something, or look at the problem from the other end? Write that down. That note is worth more than the card.

[ sign off ]

Speaking of tools that came out of human decisions: I have been building Print Ephemera, an archive of pre-internet advertising on an infinite canvas. It opened with 120 Olivetti ads, filterable by decade and by designer.

Next week, the Whole Earth Catalog: Stewart Brand's 1968 version of the same idea, which someone has just made searchable in full.

Thanks for reading.

See you next Sunday, anonymous reader behind the screen.