.@danshipper's "Automation Paradox" perfectly explains the feeling we are all having - while AI is automating our work, we are still somehow busier than ever.

Let's look at the two frontier ways of working with AI right now:

Delegation — you invoke an agent to accomplish a task end-to-end. Orchestration — you work alongside AI in something like Claude Code or Codex, steering as you go.

Both look automated. But both are quietly dependent on humans.

Delegation only works when someone owns the agent and keeps tuning it. Orchestration shines only when a human stays in the loop. Remove the human and you get volume, not value.

Here's the paradox: AI makes yesterday's competence cheap. So cheap competence floods the market. Anyone can produce a "decent start," so everyone does — and you get a glut of output that all looks the same.

That's what slop really is. Not the em-dashes or the purple gradients. It's the sameness you feel after enough AI output.

And when output goes up and sameness goes up, the demand for different skyrockets.

Who produces different? Experts. They either build systems that reliably pull high-quality work from AI, or use AI for the draft and layer their expertise on top.

So more automation doesn't mean less human work. It means more. Every agent needs a human to be good — and experts are the only thing standing between your work and the slop pile.