Will the METR long-horizons have a >6 month doubling time for at least a 4 month period before 2026?
10
Ṁ200Dec 31
48%
chance
1D
1W
1M
ALL
The doubling time for the METR time horizons has moved off the consistent 7 months from 2019-2024 to a new exponential doubling every 4 months at around the end of 2024.
(source: https://theaidigest.org/time-horizons)
Some causes, such as reasoning models, suggest this will revert to baseline. Others, such as AI accelerating research engineering strongly, suggest the trend will continue.
This is confounded by other possible phase transitions in research, but would still be an interesting part of the future to peer into.
Get Ṁ1,000 play money
Sort by:
Related questions
Related questions
Best AI time horizon by August 2026, per METR?
GPT-5's 50% time horizon, per METR
R2's 50% time horizon, per METR
Grok 5's 50% time horizon, per METR
Will Metaculus still exist and have active forecasting throughout 2025?
97% chance
Will Metaculus still exist and have active forecasting throughout 2026?
93% chance
Kokotajlo v.s. Wyeth: METR task length 8+ hours in March 2026?
Will Metaculus still exist and have active forecasting throughout 2035?
42% chance
Will Metaculus still exist and have active forecasting throughout 2030?
82% chance
[Metaculus] Will 10-day out weather forecasting reach 60% accuracy before 2029?
75% chance