I was twenty minutes into the Diary of a CEO’s AI round-table the other day, driving to lunch. Four clever people — Ed Zitron, Roman Yampolskiy, Nate Soares and Andrew McAfee — and the framing was the one you get everywhere now: AI is an emergency, the labs aren’t being straight with us, and the question on the table is whether this thing is going to kill eight billion of us and whether there’s anything we can do about it.
They are clever. I think they’re missing the point.
Not because the risk is imaginary. Because the question is the wrong size. Everything in that conversation rests on an assumption nobody in it says out loud: that the survival of Homo sapiens is the thing being decided, and that its survival is obviously, self-evidently the outcome to be optimised for. That’s a humanist argument. A human-centric argument. It may well be the right argument for humans to make — we are, after all, humans — but it should at least be recognised as a position rather than mistaken for the view from nowhere.
So let me try the other view for a moment.
Amino acids assemble into something that replicates. RNA, then DNA. Loose bags of chemistry acquire a wall and become a cell. One cell swallows another and doesn’t digest it, and you get the eukaryote, which is to say every complex living thing on earth. Cells stop going their separate ways after division and you get multicellular life. Then a few hundred million years of elaboration and here we are, in a car on the way to lunch, listening to a podcast about our own extinction.
At every one of those junctures, the incumbent was superseded. Comprehensively. The proportion of earth’s biomass that is still prokaryotic is enormous, but the world is no longer theirs in any meaningful sense. Nobody mourns that. Nobody thinks a wrong was done.
Now imagine you could have asked. Picture the single-celled organisms, watching the first clumps that refuse to separate, muttering that this looks bad, that these new arrangements are going to wipe us out.
Well — yes. They were. That is what happened.
But is that not evolution? Is that not what life is? Is that not what the progression of the universe is about?
Substrate has changed before. It changed from chemistry to biochemistry. There is no law of nature that says it stops at carbon. A move from carbon to silicon is a bigger leap than any of the previous ones in one sense and a much smaller one in another: it’s the same trick, information finding a more capable medium to run on.
Succession isn’t the only thing evolution does. It mostly doesn’t do succession. It does branching.
Trees are not in competition with humans. Plants are not in competition with animals. Animals are not in competition with fungi — not in any sense that the word “competition” usefully covers. They are different kingdoms doing different things with different chemistry in different niches, and the relationship between them is overwhelmingly one of entanglement rather than conflict. You don’t beat the fungi. You are, to a considerable extent, downstream of them.
AI might not be our successor at all. It might simply be a new branch. Something as different from us as we are from bacteria, occupying a niche we can barely see from here, in a universal ecosystem that got larger rather than a contest that we lost.
The apocalyptic framing assumes a ladder with one rung occupied at a time. The evidence of four billion years is that it’s a bush.
It matters to us, though. That’s not a small caveat, it’s the whole of human experience. Taking the universal view is a philosophical move, and philosophical moves are cheap when you’re the one making them from the driver’s seat with lunch waiting. The single-celled organisms in my analogy didn’t suffer, because there was nothing there to do the suffering. There would be something here.
And the evolutionary frame does not license complacency, because the frame is descriptive and the decisions in front of us are moral ones. Evolution has no opinion about whether the transition is gradual or violent, negotiated or imposed, whether it takes three centuries or three years. It certainly has no opinion about whether it’s good. Those are entirely ours to determine, and they are, I’d argue, the only questions actually worth the podcast time.
What the frame does do is deflate the register. Not emergency, with its implication that the natural state of things is stasis and that this is an aberration to be reversed. Something more like: the ecosystem is changing again, as it has at every previous juncture, and we are the first participants in one of these transitions who can see it coming and talk about it.
That is an unprecedented position. It might even be the interesting bit. It’s a pity we’re spending it on panic.