Corn-Poop Phenomenon

Source: A primitive for enabling environments: early work on machine-generated prompts

After the GPT moment at the end of 2022, I lost track of why I'd spent years studying engineering and marketing. With a bit of despair, a question kept running through my head: does it still make sense to learn on my own? If AI is better at it than I am? As the startup grew, I had to manage people who weren't Japanese, so I put time into English and got somewhat competent — but I wasn't sure (at the level of a whole life) whether spending money and time on that was worth it.

Even after burning out at the startup and starting otiose corp., I felt that self-directed learning and education might be "the biggest challenge of the AI era." Through the bulletin, by connecting with people in different countries, I thought I might grasp something — and what I ran into, right then, was Andy Matuschak's writing.

"The point here isn't really to memorize. It's to be changed—to metabolize an experience so that I feel or act differently in the future."

I keep listening to YouTube videos on AI, engineering, and product development at 1.5× speed. I paste whole videos into NotebookLM and have them summarized so I can ingest them fast. My NotebookLM account has already hit its limit. And yet — I swear — none of this consumed content has changed how I feel or how I act.

In Andy's terms, I've only memorized. I haven't metabolized. Crude analogy: it's like eating a ton of corn one day and finding most of it still intact in the stool the next — undigested. Maybe this isn't just me. Maybe it's a common condition for people in the AI era, hooked on dopamine and hype. I'll call it the corn-poop phenomenon.

Andy's article has ideas for getting out of the corn-poop phenomenon. The language gets a bit hard in places, but I'd like you to read it. On the other hand, to make learning something that actually changes how I act, I have a hunch the hint is to deliberately slow down. Like this — writing with a pen.