The Flat Curve Society

Tech IndustrySoftware EngineeringCareer & LifeRants & Opinions

Yegge argues that AI capability has hit an inflection point where the most powerful models are becoming dangerous enough to be locked down like nuclear weapons, with access restricted to a tiny few via government and supply-chain control. For everyone else, the intelligence curve will *appear* to flatten—not because progress stops, but because frontier models become inaccessible and because most users hit a 'discernment horizon' where they can't tell models apart or verify superhuman output. He calls this the 'Flat Curve Society': the exponential continues invisibly while we operate as if it's plateaued. He contends this stable plateau is actually good—it lets SaaS rebound, makes software a durable craft again, and keeps engineers relevant. The bulk of the post pivots to AI literacy, citing a Netflix training study showing employees can jump from non-users to single- or multi-agent users in ~5 hours, measured by daily token spend. His prescription: train your workforce up the literacy curve, then teach them token efficiency, since mastery flips from spending tokens to conserving them.

AI's exponential progress will continue but feel like a plateau because the dangerous frontier gets locked away and most users can't even verify superhuman output—so the real competitive skill becomes AI literacy and token-efficient craft, not access to ever-smarter models.
  • 7

    The intelligence curve is as real as the Earth is round, but just as flat from where you stand. Welcome to the Flat Curve Society.

  • 6

    Superhuman means unverifiable.

  • 6

    You can't hand out an intelligence engine that nobody can supervise. It's pointless to own because you won't know if it's helping you or walking you off a cliff.

  • 8

    The demand horizon is where you can't tell two models apart because you don't have a hard enough problem.

  • 6

    That router is the discernment horizon encoded as infrastructure.

  • 7

    It is a craft that only gets harder, and more valuable, the weaker your models are.

  • 8

    Well then it's not that I'm not using AI, I'm just extremely token-efficient.

  • 7

    AI Literacy does not come for free. The only thing you get for free is AI Anxiety.

opinionated, speculative, conversational