Idea · carried by 1 source
Inside the Exponential
The framing, attributed to Dario Amodei, that we have now entered the steep part of the exponential growth curve for AI capabilities. At this phase, every model improvement represents a massive jump — not incremental progress, but discontinuous leaps.
Where it comes from
Lenny describes Dario Amodei's framing: > “He kept talking about how we're going to do code... He talks a lot about this exponential that we're now on. We're like inside of the exponential now, which by definition means every improvement is a massive jump because we're like on that hockey stick part.” (~14:17) Dianne confirms and elaborates on what it's like operating inside this reality.
How it works
- Before the exponential: Models improve incrementally. Each release is a step forward but not revolutionary. Skeptics say “it's plateauing.”
- Inside the exponential: Each new model generation unlocks capabilities that were previously impossible. The pace is disorienting.
- By definition, each jump is massive: On the hockey-stick curve, the vertical distance between releases keeps growing. What felt like a huge leap last year looks small compared to this year's leap.
The most dangerous assumption right now is linearity. Planning as if model improvements will follow the same pace as last year means being blindsided by what's actually coming — because we're no longer on the flat part of the curve.