Citat:
Modellerna är snabbt på väg framåt, det labben har är mycket mer än vi vanliga dödliga får prova.
https://thezvi.substack.com/p/ai-r-and-d-is-accelerating-towards
https://thezvi.substack.com/p/ai-r-and-d-is-accelerating-towards
Motargumentet, men även det är starkt om det stämmer.
Citat:
We’re entering the steep side of the METR task-horizon S-curve.
The published data through early 2026 still looks roughly exponential—p80 horizons (the length of skilled-human software/ML task an agent can finish with 80% success) have been doubling every few months. That’s amazing, but pure exponentials don’t last. Evaluation ceilings, scarce long-horizon demonstration data, and the messiness of real work all point toward a rollover.
I expect the measured curve to bend. My current best-guess logistic fit puts the inflection around now to mid-2027, then:
~65 hours by 2028
~550 hours by 2030
- approaching a practical plateau near 1,000–1,400 hours (~1.5–2 months of human time on tidy software tasks) by the mid-2030s
I’m open to being surprised. A strong self-improvement loop or new evaluation methods could keep the steep rise going longer and push the asymptote higher. Architectural or data bottlenecks could flatten it sooner. The current METR suite is already unreliable above ~16 h on the 50% horizon, so the next few years of measurements will be noisy either way.
The plot is the S-curve projection (green) against continued recent exponential (red). History says the green one is more likely. We’ll see.
Still this is a massive change for the benefit of humanity.
The published data through early 2026 still looks roughly exponential—p80 horizons (the length of skilled-human software/ML task an agent can finish with 80% success) have been doubling every few months. That’s amazing, but pure exponentials don’t last. Evaluation ceilings, scarce long-horizon demonstration data, and the messiness of real work all point toward a rollover.
I expect the measured curve to bend. My current best-guess logistic fit puts the inflection around now to mid-2027, then:
~65 hours by 2028
~550 hours by 2030
- approaching a practical plateau near 1,000–1,400 hours (~1.5–2 months of human time on tidy software tasks) by the mid-2030s
I’m open to being surprised. A strong self-improvement loop or new evaluation methods could keep the steep rise going longer and push the asymptote higher. Architectural or data bottlenecks could flatten it sooner. The current METR suite is already unreliable above ~16 h on the 50% horizon, so the next few years of measurements will be noisy either way.
The plot is the S-curve projection (green) against continued recent exponential (red). History says the green one is more likely. We’ll see.
Still this is a massive change for the benefit of humanity.