Assembly Required.

AI may matter less than how we decide what we believe about it.

The AI debate has split into two tribes. One treats any use of it as a moral failure. The other treats every use as inevitable progress. Both feels like they're driven more by vibes than understanding.

I refuse both. I only trust what I have tested with my own hands on work that actually mattered, such as debugging and writing code, copywriting, design, and the like. The results have left me more skeptical than hopeful.

AI, I have found, rarely delivers the 10x leverage its promoters promise. On creative work, it produces the average of what already exists. It's impressive only because humans already did the foundational thinking. The quiet danger is subtle, yet insidious: it's easy to stop noticing when the tool is executing your judgment and when it has started making the judgment for you.

Where this technology does help is narrow: untangling broken code, parsing opaque documentation, drafting clearer docs. While this is useful leverage of this technology, it isn't a magical push-button solution and it still requires human tinkering. I am a programmer and a writer. I can see the parts of my work it can (and should) absorb. I can also see where it fails completely and would be a catastrophic substitute for human insight.

After genuine experimentation, I am not worried about my livelihood, and neither should you be worried about yours. Even in coding the human must stay in the loop, and much more than the hype admits.

I sometimes wonder if the revolutionaries calling it civilization-scale and the doomers calling it worthless share the same flaw: confidence without the scars of real use. I keep thinking of the Luddites. They did not fear machines in the abstract. They were skilled practitioners who understood exactly what the looms would do to their craft and their families. Their resistance carried weight because it was earned from the inside. I wanted the same standard.

After my own experimentation, I land mostly against AI. You might be on the opposite side; if you've gotten there honestly, then I can respect and appreciate that.

I also arrived at another conclusion: the process of arriving at my position, through direct experiment rather than tribal loyalty, perhaps matters more than the conclusion itself.