What does it mean that all schemas are wrong some are useful?

No schema perfectly represents reality but some are more useful than others for a given purpose.

Example: Newtonian mechanics is technically wrong — it ignores relativistic effects that compound at high velocities and in strong gravitational fields. But NASA still uses it as the primary framework for most orbital calculations. For the vast majority of spaceflight, Newtonian math is less than 0.0001% off from reality. It's wrong. It's also the most useful schema engineers have for getting things into orbit. Usefulness, not correctness, is the standard.

Try this: Pick one schema you use daily — a mental model, a planning framework, a personality type system, an architectural pattern. Write down three things it gets wrong or leaves out. Then write down three situations where it remains the most useful tool available despite those flaws. You now have a usefulness profile: a clear map of where this schema works and where you should reach for something else.