Today is October 16, 2026. Tags: [VERIFIED]…

For the two most recently completed doublings of frontier training-compute efficiency before 2031-01-01, what will be the ratio of the time taken by the later doubling to the time taken by the earlier one?✓#1

ratio0.96dimensionless ratio or multiplier (1.0 = no change)

rationale:

My median estimate sits just below 1.0 (0.96), reflecting an expectation of roughly constant or slightly accelerating efficiency doublings. Evaluated alongside related indicators, the projected 4.6x improvement in sample efficiency supports the view of consistently sustained, rather than radically accelerated, algorithmic progress. A major factor anchoring and slightly tightening the distribution near 1.0 is the likelihood of a methodological artifact: i

Evaluated alongside related indicators, the estimate held steady as the projected 4.6x improvement in sample efficiency supports the view of consistently sustained, rather than radically accelerated, algorithmic progress.

Weighed against 58 other forecasts.

alphaxiv.orgscholar.google.co.inepoch.aiepochai.substack.comlesswrong.com+39 more

What will be the ratio of (a) frontier-lab spending on human-expert data generation in calendar 2029, to (b) frontier-lab spending on human-expert data generation in calendar 2026?✓#2

ratio2.8dimensionless ratio or multiplier (1.0 = no change)

rationale:

Establishing the 2026 baseline relies heavily on vendor revenue disclosures and run rates, as labs do not publish exact human-data spending. The top five frontier developers likely spend an aggregate of $5–10 billion in calendar 2026. This is built from major vendor footprints: Mercor reached a ~$2 billion gross run rate by mid-2026 with ~90% tied to foundation-model labs Mercor revenue, valuation & funding | Sacra[Mercor revenue, valuation & funding | Sacra](https

Set against related questions, the distribution was adjusted slightly to account for the dampening effect that a projected 4.6x improvement in sample efficiency may have on human-data budgets.

Weighed against 64 other forecasts.

fastcompany.comtechbuzz.aisapphireventures.comgetlatka.comtomshardware.com+121 more

By what factor will the training-token count required for a frontier model to reach a fixed capability level fall between 2026-01-01 and 2031-01-01?✓#3

ratio4.6dimensionless ratio or multiplier (1.0 = no change)

rationale:

Forecast Summary & Core Dynamics I estimate the token count required to reach a fixed capability level will fall by a median factor of 4.6x between 2026 and 2031, with an interquartile range of 2.4x to 9.5x and a wide 80% confidence interval ranging from 1.3x to 22.0x. Evaluating this question alongside related estimates of compute efficiency doublings and structural data constraints reinforces the median while reflecting, particularly in the interquartile range, a fundamental tension betwee

Set against related questions on training bottlenecks, the distribution was slightly compressed because parallel expectations of a 2.8x increase in human data spending suggest labs will rely somewhat more on brute-force data than pure algorithmic efficiency.

Weighed against 55 other forecasts.

useluminix.comepoch.aiarxiv.orglesswrong.comfuturetech.mit.edu+68 more

Ask a followup

Ask a followup