The question everyone's skipping past
When I was studying philosophy of mind, the Turing test was an argument. Something you debated in a seminar room, batted around as a thought experiment about whether a machine could ever be mistaken for a mind. Today about a billion people run that experiment every single day, and the machine passes. Not as a hypothetical. As a Tuesday.
I don't say that to marvel at the technology, though it is marvelous. I say it because it's the moment I started asking a different question than the one everyone around me was asking. Everyone wants to know if AI is aligned with our values. Fair question. But underneath it is a bigger one I think we're skipping past: has any technology, ever, actually made us better at the thing that produces progress in the first place? And if this is the first one in centuries with a real shot at it, what would it even mean to build for that on purpose instead of stumbling into it?
Technology is the gas pedal, not the wheel
Here's the claim I want to put on the table, and I know it's not the popular one: human progress was never about technology. It was about understanding, morality, organization, education — and technology has always been downstream of that, not the source of it.
Nuclear energy gives you bombs or power plants. Gunpowder gives you fireworks or war. Dynamite digs tunnels or levels cities. The technology doesn't pick. We pick. Technology is the gas pedal — it makes whatever direction we're already pointed in go faster. It's not the wheel, and it's certainly not the car. The danger of forgetting this is that we hand more power to more people whose misalignment with each other doesn't shrink just because their capability grew. It usually gets worse.
The technologies that actually bent the curve
Every once in a while, though, a technology comes along that doesn't just amplify what we could already do — it changes the terms of how we do it together. Phonetic writing democratized who could read and argue and record. The mechanical clock gave whole societies a shared unit of coordination they'd never had. The printing press didn't just make books cheaper, it made a public argument possible at a scale that hadn't existed before.
Compare that to what we did with social media, which promised something similar — a tool for a more connected, better-organized society — and instead became a very efficient machine for extracting attention. Nobody serious defends the premise anymore that it actually helped us deliberate. I don't want to build the AI version of that. I want us to build the AI version of the printing press.
Deliberation is the actual mechanism of progress
Here's the reframe I keep coming back to. If technology isn't what produces progress, what does? I'd argue it's deliberation — our collective, unglamorous process of fumbling through the dark together and figuring out what's false, what's wrong, and where to go next. Science works because it built a rigorous method around exactly this. Democracies function better when they build real organization around it too. Socrates spent his whole life doing one thing to the smartest people in Athens: making them say out loud what they didn't actually know. That pestering is the cornerstone of Greek philosophy, and arguably the cornerstone of science itself.
A smart person with bad motives has always been dangerous. That problem predates AI by all of human history. What's new is the scale — and what I think is genuinely new is the possibility that a technology could, for the first time on purpose, help us get better at the fumbling instead of just going faster in whatever direction we were already fumbling.
Can machines give us more fingers?
I think about deliberation as feeling your way through a dark room. All you have is what's at your fingertips — where the wall is, what doesn't move, what gives way. That fumbling doesn't go away. No amount of AI hands us divine revelation. What I'm asking is narrower and, I think, more answerable: can machines give us more fingers?
Not machines that think for us. Machines that serve the thinking — quickly, accurately, with real feedback — the way a good friend does when they know exactly when to push back, when to redirect, when to help what you're trying to say land with the right nuance at the right moment. Most of us have that friend, if we're lucky, for a few hours a week. What happens to a society's deliberation — its policy debates, its geopolitical strategy, its arguments from the municipal council up — if that friend is available all day, to everyone, at once?
That's not a rhetorical flourish. It's the actual discovery journey I'm on with this lab: can we generate large-scale, regulated, rule-based debate that accelerates the deliberations a society actually needs to have. I believe the answer is yes. I don't yet know what the mechanism looks like at scale, or how you keep it honest once a billion people are doing it every day instead of a few hundred people in a seminar room. That's the part I'm building toward next.