// Imagining the Traffic Jam
Technology has no inevitable trajectory.
Last of five, following on from An Ocean of Artificial Intelligence. What we didn’t imagine when we built the attention economy, what the same question means for the intention economy, and why the destructive path is always the easier one.
Also: a toddler who can infer what you meant to do, an installation at the Venice Biennale, and the moment we all lost, one more time.
The traffic jam
My career has always been about imagining the future and wrangling it into the present in accordance with my will.
There’s a line from Monika Bielskyte I keep coming back to. You can invent the car by imagining a world without horses. To be genuinely futuristic, you have to imagine the traffic jam.
The people who built the attention economy weren’t all cynics. They imagined connection. Access. Community at scale. They didn’t imagine what a billion of those systems would do to how people think and talk to each other. Or they chose not to.
The same question hangs over the intention economy. Both versions of it.
Holly Herndon and Mat Dryhurst spent this year making art about exactly this. Their installation at the Venice Biennale had one plaque. It read: “When machines are capable of building anything, the most difficult question is understanding what we want and why.”
Dryhurst calls what comes after the slop period an intention economy. Same phrase. Different decade.
Bielskyte makes another point that’s worth sitting with. Building something genuinely useful in healthcare, in education, in creative work requires accuracy, reliability and sustained effort. To spread disinformation or deepen social division, it only has to be cheap, fast and convincing enough. The destructive path is structurally easier. That’s why the traffic jam keeps happening, and why it’s most dangerous early on, before anyone fully understands what’s been built.
The builders of these systems have their own intent deficit. Growth targets. Engagement metrics. Conversion rates. Those are proxies for intent. The harder question is what kind of world we actually want to make. It rarely makes the slide deck. It’s not an OKR.
Some researchers are already building this differently. MIT’s CSAIL published work on an algorithm that infers what you meant to do, not just what you did. They modelled it on something a toddler can already do: watch someone struggling to open a cabinet full of books, figure out what they were trying to do, and open the door for them. The toddler doesn’t observe the action. They infer the goal.
The lead researcher put it plainly: “Ideally, the algorithms of the future will recognize our mistakes, bad habits, and irrationalities and help us avoid, rather than reinforce, them.”
The toddler doesn’t sell what it learns about you to the highest bidder in the next room. That’s the difference. Who the inference serves.
Technology has no inevitable trajectory. We can direct this toward strengthening human agency rather than exploiting human vulnerability. That’s an act of intent.
The intent has to live somewhere in the process. Make sure it does.
The attention span story is true. It’s just not the whole story.
We didn’t lose our attention.
We lost the moment where we decided what to pay attention to.
I'd love to tell you more.
// An Ocean of Artificial Intelligence
Optimise everything toward the same metrics and you get the same product.
Fourth of five, following on from Rocket Boots for Terrible Ideas. About what homogenisation does to trust, once every option looks and behaves the same, and why that makes leaving easy and staying meaningless.
Also: the same coffee shop in Berlin, Bucharest and Cape Town, enshittification, and why there’s no hook left for customer intent.