Stream

The Tao of Programming predicted the age of AI coding — and it did it by accident

Geoffrey James wrote a satirical manual about programmers and their managers decades ago. Dry, deadpan, fortune-cookie sentences.

Read now through the lens of agentic coding, and it takes on a different meaning.

The Tao warned us: complex systems resist rigid control; programs succeed by accident; the people issuing instructions rarely understand the work they manage.

Sound familiar? Today, the developer is often the manager, and the AI agent is the coder — the roles have quietly rotated.

What the Tao couldn’t foresee is that the entity writing the code now has no skin in the game: no intent, no pride, no 2 a.m. indignation about an edge case.

That absence isn’t a bug in the old philosophy; it’s a new reality we need to design for — one that assumes the coder has no ownership.

Re-read the Tao. It’s a warning and a guidebook, if we learn to read it for the century of agents.

Understanding often feels real right up until the moment you try to explain it.

If you can’t explain something in your own words, you probably don’t understand it yet.

You might recognize the terms. You might nod along. But that’s not understanding that’s familiarity pretending to be knowledge.

One of the fastest ways to puncture that illusion is to think out loud. Write the idea as if you’re explaining it to a curious beginner. No jargon. No shortcuts. Just plain language. The moment you try this, the fog shows up exactly where your understanding is weak.

You can go even further by teaching the idea—to a friend, a colleague, or an invisible audience on the internet. As Feynman believed, teaching isn’t a byproduct of understanding; it’s the stress test.

When you get stuck mid-explanation, that’s not failure. That’s the location of learning.