The Operating System
22 Frontier Play

Project 20 Years Out, Then Work Backwards

Generate direction by imagining the world two decades ahead and reasoning back to what has to be built now, surrounded by people who take wild ideas seriously.

You want to be able to project yourself 20 years into the future, and then think backwards from there.

The method behind Altman’s wave-picking, from Idea Generation, is long-horizon backcasting: “You want to be able to project yourself 20 years into the future, and then think backwards from there.” Start from the world you believe is coming, then reason back to what must be built today to get there, a way of generating direction that ignores near-term noise.

Environment is an input, not a backdrop. Altman is explicit that who you think alongside shapes the quality of your ideas: “Stay away from people who are world-weary and belittle your ambitions.” Ambition is partly a function of the room you’re in.

Doesn’t transfer if you can’t survive the near term. Backcasting from 2045 only pays if you’re still solvent in 2026. It’s a luxury of the well-capitalized; a startup that’s default-dead in six months can’t optimize for a two-decade horizon, and telling it to is advice from someone who no longer has that constraint.