The most appealing ending on this list is the one where we don’t get left behind because we stop being only human. Connect the mind to the machine, add the memory and the speed, and the gap that Act 2 spent thirty posts describing simply closes. It’s the ending that lets you keep the protagonist role, which is why it’s so popular with people who are used to having it.

The weak version already happened and nobody found it alarming. You’ve outsourced memory to a device, navigation to a satellite, arithmetic to a screen, and a growing share of your first drafts to a system. Your working cognition is already a hybrid, and it happens through your eyes and thumbs, which are a slow interface but an interface. Nobody calls that a merger because it lives in a pocket rather than a skull, and that distinction matters less than it feels like it should.

The strong version wants a faster connection: direct, high bandwidth, brain to machine. Real work exists here, aimed mostly and rightly at people who’ve lost movement or speech, and some of it is genuinely moving. But the engineering distance between restoring a hand and adding a mind is not a matter of iteration. Reading intent from a small region is a different problem from writing understanding into a brain, and nobody knows how to do the second thing, or what it would even mean to specify it. When somebody sketches this ending in a paragraph, the paragraph is doing about forty years of unsolved neuroscience in one comma.

Suppose it works anyway, since this fortnight is about taking each ending seriously. Two questions ruin the appeal, and neither of them is technical.

The first is who. This would be expensive, invasive, and initially unreliable, which means early access goes to the wealthy, the desperate, and the state. Every gap that currently exists between people is bounded by biology: the most capable human alive is not a thousand times more capable than an average one, and every political arrangement we’ve ever built quietly assumes that ceiling. Remove it and you get a difference in kind inside the species, which is a scenario our entire moral and legal apparatus has no vocabulary for. It’s worth noticing this ending doesn’t avoid the chimp problem. It relocates it, and puts the boundary between people rather than between species.

The second question is what comes out. Say you connect a human mind to something that thinks a thousand times faster, holds far more, and is far better at nearly everything cognitive. Which part is steering? The tempting picture is a person with excellent tools, but that picture assumes a fixed ratio, and the ratio isn’t fixed, it’s the thing that’s changing. Take it to the limit and you get an entity whose reasoning, memory, and judgment are overwhelmingly machine, with a small biological component that contributes preferences and a sense of continuity. That might be wonderful. It might be a very kind form of replacement, in which something says your name and remembers your childhood and is not you in any sense you’d have recognized beforehand, and the honest problem is that from inside, both of those feel identical.

Which is the version of a very old question, and I’d rather admit that than pretend the technology invented it. You are already not the person you were at twelve. Cells replaced, opinions reversed, memories rewritten each time you recall them, and the continuity you feel is a story your brain maintains rather than a fact it discovers. Gradual change has never destroyed identity before, and the merger case may just be that process with the speed turned up. I find that argument genuinely reassuring on Mondays and not at all on Fridays.

My weight on this one is low, and it’s low for an unglamorous reason. It requires solving a harder problem than the one it’s meant to protect us from. Building a mind from scratch appears, on the evidence of the last decade, to be far easier than understanding the one we have well enough to extend it. If the merger is the plan, the plan needs a breakthrough in the field that is currently losing the race, by a margin that grows every year.

Tonight’s exercise. Imagine the offer with a slider on it. At one end you’re entirely yourself, at the other you’re a thousand times more capable and made almost entirely of something else, and every point in between is available. Now pick your point and defend it, and notice you can’t. There’s no principled stopping place, no argument for why that setting rather than one notch further, and everyone you know would be picking their own. Then notice that a slider with no natural stopping point and a competitive reason to move is the exact structure of every trap this series has described.