It's Sad How Much You Want To Automate

I’d been sending my wife screenshots all week.

Mostly of things that were broken. A wall of red. A thing that was meant to run overnight and just didn’t, for reasons nobody will ever know. Here, look at this. She has no context for any of it. I sent them anyway, in the way you talk at someone while you’re thinking.

At dinner she said it’s sad how much I want to automate. And that it could work.

I didn’t defend it. I’m not really trying to convince anyone of this, including her, including me. But the second half is the interesting one. She wasn’t saying it won’t work. She was saying the sad part is that it might.

I don’t feel the sadness. That’s the honest report.

I didn’t ask what she meant, so I’m not going to put words in her mouth. What I can talk about is the belief sitting underneath the whole subject, which is that the doing is the point. That there’s something real in the work itself and taking it away removes something.

Loads of people hold this and it makes sense that they do, because when you’re an employee the hours genuinely are the product. You get paid for them. Worth attaches. It would be strange if it didn’t.

It just never installed in me, and not because I saw through anything. I simply never wanted to take orders.

Early on I’d have told you it was about money, and it was, for a while, in the way it is when you’re still counting. That’s mostly gone now and it went quieter than I expected. What’s left underneath is smaller and less impressive. I don’t want to be told what to do. The work was never the objection. The work just tends to arrive attached to instructions, and that’s the bit I’ve been dodging my whole life.

Which isn’t a principle. It’s the position of a fifteen year old. I’ve just had longer to arrange things around it.

Now the part I’d rather skip.

What’s left for me sounds impressive if I say it in a certain voice. Direction. Taste. The calls that need a human. There’s a version of that sentence where I’m wearing a gilet.

The truer version is that what’s left is the bit I haven’t worked out how to hand over yet. Some of it is real. A lot of it is me squinting at things to check they’re fine, because there’s currently no other way to know.

That’s not philosophy. That’s a missing feature.

If the thing could grade its own work against something, actual tests, checks, whatever they end up being, I wouldn’t need to look. I’d trust it further because there’d be a reason to. Right now my trust extends about as far as my own eyeballs, which is roughly one screen wide. Everyone building this stuff talks about evals. I have built precisely none.

So I’ve made a factory, skipped quality control, and have been calling the gap judgment.

One more thing worth flagging, mostly to myself.

I’ve decided that the doing-is-the-point belief is something people got taught, and that mine is strategy. Very convenient. It’s also exactly what I’d say if it were the other way round, if I were the attached one, dressing it up in the language of leverage. From in here the two look identical. Anyone who tells you they can spot the difference in themselves is doing a bit.

Anyway.

None of this got written down. It came out at six in the morning, into my phone, while I was having a shit. Ten minutes, and I say the word freeing about nine times.

That’s where the thinking happens now. Not on a walk. Not at the desk I built the whole system on. A man on the toilet, talking to himself about org charts, technically working.

And the last thing in it, after all the lanes and the plans, was this.

The thing works and it’s grey. Status lines, progress bars, text. Correct, and with nothing in it.

What I actually want is little cartoon characters running around doing the jobs. Minions, but clever ones. A tiny design department that shows up for design work and argues about something. Small silly workers, visibly busy. I want to watch them.

I sat with that for a second, because it is obviously ridiculous, and then thought no, that’s the most honest sentence in the recording.

I’m not building a company. I’m building the thing that builds the company. Maybe a few, if I enjoy it. Which sounds unhinged out loud and is nonetheless what’s happening. It runs on nobody.

And the first thing I want to do is put some people in it, even if they’re drawn.