Wands From the Sky

The Wands

Wands are falling out of the sky. Most people haven't noticed. The ones who have are busy arguing about what kind of wand it is, or whether wands are safe, or whether wand-users will replace non-wand-users in the job market.

Almost nobody is asking the interesting question: where do you point it?

The cost of intelligence has collapsed. Not completely — not yet — but enough that a single person with a laptop can now do research, analysis, and synthesis that would have required a team and a budget eighteen months ago. The bottleneck is no longer capability. It's direction.

And direction, it turns out, is revealing. People say they'd never behave like the debauched billionaire, but most of them have simply never had the opportunity. Power doesn't corrupt so much as it exposes what was always there. AI isn't giving everyone wealth, but it is giving everyone access to a kind of leverage that used to require teams, budgets, and credentials. For the first time, a very large number of ordinary people have something that functions like power — and what they do with it is starting to diverge.

The Sorting (Continued)

I wrote previously about picking up the stone — the moment when a new capability appears and people sort themselves by response. Rejection, absorption, despondency, adaptation. That sorting is accelerating.

But within the "adaptation" category, a second sorting is happening. Among the people who've picked up the wand, the question is no longer whether to use it. It's what for.

Someone I know — a relative — picked up the wand and immediately thought: I can build a SaaS in thirty minutes. He's been chasing entrepreneurial success his whole life, following a father who chased it too. I don't knock him for it — I think it's great that he's trying. But his excitement is about the capability itself, not about where to point it. He's dazzled by the wand. The question of direction hasn't arrived yet.

I think it will, for him and for a lot of people. Once the novelty of building things fast wears off — once everyone can build things fast — the question that remains is: what was worth building?

The Privilege of Direction

Here's something I need to say plainly: I am writing this from a position of extraordinary privilege.

I have a job. I have a roof. My daughter is healthy. I'm not fleeing anything. The fact that I get to sit here and think about purpose and direction and what to do with AI is itself a luxury. Maslow would put me near the top of the pyramid — self-actualisation territory, where the questions are about meaning rather than survival.

I'm not going to pretend otherwise, because pretending otherwise would undermine the argument I'm about to make.

That career is under real threat — I'm a software developer, and I haven't written code in months because AI writes it better and faster. But even that anxiety is a privileged anxiety. I'm worried about career obsolescence, not about clean water.

Aligned With What?

There's a conversation happening in AI research about alignment — making sure AI systems do what humans want. The question is treated as technical: how do you specify objectives, how do you prevent reward hacking, how do you maintain human oversight.

But there's a prior question that rarely gets asked: aligned with what? What is the thing we want AI aligned to? What do humans actually want?

I don't have a universal answer. I doubt one exists. But I have a heuristic — a reasonable one to begin with:

Help people climb the hierarchy.

Two billion people lack reliable access to clean water. Four hundred million have rare diseases that take nearly five years to diagnose despite existing diagnostic tools. Four hundred and thirty million need hearing rehabilitation that already exists and is already affordable.

These aren't technology problems. The technology works. They're deployment problems, institutional problems, recognition problems — the overhang this site is built to map.

If collapsed intelligence costs let me investigate those problems more deeply, map them more clearly, and make them more visible — that seems like a defensible use of the wand. Not the only defensible use. But a defensible one.

The heuristic: use the new capability to help other people reach the point where they get to worry about purpose and meaning, instead of survival. Push the floor up. Make the privilege of direction less exclusive.

Where This Site Fits

This site is my attempt at an answer. Map the places where technology works but deployment doesn't. Investigate why. Make the stuck points visible so that someone — maybe me, maybe someone else — can act on them.

I don't know how effective this will be. I'm one person with a laptop and some subscriptions, not an institution. But when I'm old enough to look back at the moment the wands fell, I'd like to have pointed mine somewhere worth pointing.