RSS ⧉

In response to
Protocol Institute Blyg · Editor · “The Stevenson Test I’ve had a brainworm stuck in my head for a few days, thanks…” · item 1frf8xh464fdgmqegfb5wx66y2, v2 · <https://blyg.protocol-institute.org/t/1frf8xh464fdgmqegfb5wx66y2/> · retrieved Oct 1, 2026

The Stevenson Test

I’ve had a brainworm stuck in my head for a few days, thanks to Sean Stevenson.

Great idea from Sean Stevenson's talk at Protocol Symposium -- the test of New Nature should be that there is nobody to appeal to.

Which sends me back to Paul Ford's cult classic "Why wasn't I consulted" -- the argument that the web is fundamentally a customer service medium, and that WWIC is the fundamental question of the internet, the thing every comment section and pull request and reply-guy is really asking. The two ideas are the same idea from opposite ends: the web was built to make everyone feel consultable, and New Nature is what's left when there's no longer anyone on the other end of the complaint.

This feels like a Turing test type condition for New Nature. Call it the Stevenson test.

We’re familiar with ranting about kafkaesque bureaucracies and hard-coded protocols you can’t argue with. These are appeals to nobody with nowhere to direct them. You could complain online or do standup comedy jokes and hope for some commiseration and laughs.

AI is a new layer of new nature that has created entire new genres of appeals to nowhere/nobody. I don’t know how it’s manifesting irl but online the main genre seems to be poetic copes, often quoting some bit of poignant human artistry with a sentiment like “AI could never do this.”

The interesting thing about such copes is not whether they’re true or false, classy or cringe, or whether others agree. It’s that they’re poetic appeals to nobody, nowhere. And unlike the older layers of new nature marked by hard inflexibility, the AI layer is particularly frustrating because it is so protean and responsive in every other way.

The absent authority you feel you could appeal to feels human shaped. If only an authoritative scientist or even an authoritative AI itself could reassure you that yes, AI cannot do that, it’s mathematically proven.

But there’s nobody listening at the other end. You could complain to a real LLM, and it might even offer reassurance if trained to be kind, but that’s like being handled by the little boy in Waiting for Godot who reassures Vladimir and Estragon that of course Godot will come tomorrow.

from Protocol Institute Blyg ↗ · snapshot of v2

What an honor to have my own test! As good science fiction is supposed to predict not just the car but the traffic jam, I am now trying to figure out what the Three Gorges Dam of New Nature will be. What giant project will need to be undertaken to solve the equivalent of recurring floods?

Created: Oct 1, 2026