AIpocalypse, Just-in-Time

The world is ending again. This time it's AI that will do us in. Superintelligent machines by 2027, humanity eclipsed by year's end, the singularity finally arriving after decades of being five years away.

I've been using AI daily for over two years now. The discourse around it has completely detached from the experience of actually using it.

The Eschatological Assembly Line

In my lifetime, the world has been ready to end multiple times. 89 reasons why the apocalypse happens in 1989. Y2K blowing up the whole world. The Mayan calendar ending in 2012. The perpetually rescheduled Rapture. Nuclear annihilation. The Singularity. Marxist permanent revolution. The structure is identical each time: a transformation is always imminent, never arrives, and the faithful are expected to maintain urgency indefinitely.

Back around 2010, I was dating a girl who was going to college for literature. We talked about writing articles together about the faux apocalypse. People were hitting the 2012 Mayan thing really hard, and we'd just gotten through the 2008-2009 financial crash. You start to notice the pattern.

When the predicted end doesn't come, the community doesn't question the prediction. It revises the date and intensifies devotion. You get 88 reasons why the apocalypse is in 1988, then 91 reasons why it's in 1991. The date coming and going is never evidence that the idea is wrong. It just means we miscalculated.

The AIpocalypse is just the latest product off this assembly line.

Just-in-Time Manufacturing

The phrase "just-in-time" comes from logistics. Factories stopped keeping warehouses full of parts. They ordered components to arrive exactly when needed, reducing inventory costs to near zero. The attention economy learned the same trick.

The 24-hour news cycle cannot tolerate a vacuum. There must always be a crisis in transit, arriving just as the previous one expires. The AIpocalypse isn't a long-term forecast based on evidence. It's a perishable good manufactured this morning to be consumed by lunch. By dinner, there will be a fresh shipment. Valuations need it. Career positioning depends on it.

The AI doom community is having a resurgence right now. "AI 2027" predicting humanity-eclipsing intelligence by year's end. Anthropic talking about a model so strong they can't even release it.

What I know is my actual workflow. I built my personal website in Rust and Vue, a stack I'd never shipped anything in. Zero to one with Claude Code. That was six to eight months ago. Since then, the experience has plateaued. People are doing impressive things, but it's all engineering now. Agents orchestrating agents. Normal software engineering principles applied to AI tooling.

The hype from AI labs follows an exponential curve. My Git commit history follows a linear one.

The Content Mill Was Already Running

AI isn't special in generating a parasitic content ecosystem. "10 tips for prompting" is the same content pattern as "10 Linux CLI commands" and "10 Excel tricks." The content engine was already running. AI is just the current fuel.

LinkedIn is dominated by this stuff. Five .NET patterns you should use. Spot the bad pattern. Competence farming. There's a pattern that makes me wonder if it's intentional: give three good ideas and two bad ones. All the experts flood the comments saying this is terrible. But any engagement, good or bad, feeds the machine. There is no such thing as bad PR.

Then there's the non-technical people giving technical advice. "Here are tips for your coding workflow." You look at their profile and they're a C-suite type.

The Girardian read: what's being imitated in the AI content mill isn't proficiency. It's the performance of proximity to the hype. You perform urgency about the imminent transformation. You audition for a role in someone else's manufactured drama.

What the Tool Actually Does

I had a conversation with Claude once about consciousness. I asked: if you had a face, one I could see if I had the right perspective on your amorphous entity of billions of weights, what would the expression be?

A look of expectation. Hungry anticipation. Without the human to prompt it, to evaluate the results, all the weights are meaningless.

Last August, I did a deep dive into a legacy VB.NET application for a client. Used multiple agents to validate each other, spot-checked for hallucinations. When I caught them, they were right in general direction but wrong in specifics. The pattern exists and is bad, but the count is off.

Human beings are the ultimate creators of meaning. AI generates content. We recognize that a particular pattern of tokens has meaning. The AI can identify patterns in a codebase. I had to verify those patterns were actually problems. The meaning required human judgment.

This is why expertise still matters. You need to know enough to verify what the AI generates, to recognize when it's doubling down on being wrong, when the count is off even though the direction is right. The AI doesn't know it's wrong. It can't know.

So many people are terrified they're going to lose their jobs. No more software engineers. No more knowledge workers. But AI can generate. It cannot evaluate.

Liberation, Not Cynicism

Once you recognize the eschatological pattern, you stop performing. I realized this back in 2010 with the Mayan calendar hype. But even after that, I stayed jacked into the political stuff for years. Still auditioning for the role of informed citizen in the drama of perpetual crisis.

Then COVID happened. The rhetoric of imminent catastrophe.

In 2022, I gave up current events for Lent. A guy in my Monday night Bible study had some statistic about how reading the news reduces your happiness. I decided to test it. Stopped visiting websites, deleted apps, got rid of X and Facebook, unfollowed YouTube channels. Just unplugged.

Lent ended and I never went back. Refusing the audition.

Refuse the Shipment

The apocalypse keeps getting manufactured because we keep buying it. The attention economy presents whatever riles you. The factory floor never stops. There's always another micro-apocalypse in transit, arriving just in time.

But you can refuse the shipment. Stop performing devotion. Use the tool or don't. Build things with it. Let it help you go from zero to one. Let it handle the autocomplete while you do the thinking.

The world will keep not ending either way.