Plans in Decades, Execution in Weeks

One of Altman’s cleanest operating maxims, from his 2024 essay: “Plans should be measured in decades, execution should be measured in weeks.” Decade-scale ambition keeps you from optimizing for noise; week-scale execution keeps the vision from becoming an excuse for slowness. Most people collapse both horizons into the same medium one, and lose both benefits at once.

It’s how he runs OpenAI. AGI is the decades-long goal; a relentless weekly ship cadence is the tempo. The two horizons run in parallel, not in tension.

Doesn’t transfer without stability. Holding a genuine decade plan requires institutional footing (funding and autonomy) that most companies lack. For a startup living raise-to-raise, “decades” is aspirational cover; only the near term is actually real, and pretending otherwise can be a way to avoid the uncomfortable short-term math.