These questions form one register on…

As of 2026-08-13, by what factor is the rate of valuable AI research output at the largest AI lab by revenue higher than it would be if that lab's researchers had no access to AI assistance?✓#1

ratio1.6dimensionless ratio or multiplier (1.0 = no change)

rationale:

Forecast: p10=1.15, p25=1.35, p50=1.6, p75=1.9, p90=2.35.

Target Identification and the Core Metric Initial logic and parameters are validated regarding Anthropic's annualized run rate exceeding $65B compared to OpenAI's ~$40B 3 sources. Standard processing applied to evaluate the counterfactual uplift in valuable research output versus localized coding throughput, observing that over 80% of merged code is authored by Claude 2 sources. Early assessments of organization

Aligning this baseline estimate with structural projections of AI research productivity growth through late 2029 slightly trimmed the upper tail to maintain a mathematically coherent compounding trajectory.

Weighed against 6 other forecasts.

axios.comwe0.ai80000hours.orgemergent.shmetr.org+172 more

By what factor will the rate of valuable AI research output at the largest AI lab by revenue on 2029-12-31 exceed the rate that would obtain if that lab's researchers had no access to AI assistance?✓#2

ratio4.1dimensionless ratio or multiplier (1.0 = no change)

rationale:

The forecast estimates the AI R&D uplift IN VALUE at the largest frontier lab on 2029-12-31, relative to a same-date no-AI counterfactual. As of August 2026, the best evidence suggests true value uplift sits between 1.5x and 2.0x. Anthropic's August 2026 Risk Report states that internal R&D is significantly faster but not yet by a factor of 2 3 sources, and METR’s May 2026 survey found median self-reported value uplift around 1.4–2x metr.org. The resolving entity is most l

Modeling this estimate as the explicit product of our present-day baseline uplift and our structural growth ratio over the next three years slightly raised the median while preserving a wide upper tail.

Weighed against 6 other forecasts.

frontierriskmonitor.orgepoch.aiaiweekly.cofrontierrisk.substack.combenzinga.com+233 more

What will be the ratio of (a) the AI-assistance multiplier on the rate of valuable AI research output at the largest AI lab by revenue on 2029-12-31, to (b) that same multiplier on 2026-08-13?✓#3

ratio2.6dimensionless ratio or multiplier (1.0 = no change)

rationale:

The final estimates for the ratio of AI-assistance multipliers are p10: 1.2, p25: 1.65, p50: 2.6, p75: 5.2, and p90: 11.5. Initial logic and parameters for anchoring the present denominator and projecting the 2029 numerator are validated. The established context regarding Anthropic's current baseline metrics 3 sources and survey data metr.org requires no further intermediate refinement. Projections of Claude merging code anthropic.com and systems handling significant re

Ensuring this relative growth multiple aligns mathematically with independent assessments of absolute current and future AI R&D multipliers resulted in a slight downward shift across the distribution.

Weighed against 6 other forecasts.

businessinsider.comaibusiness.vcqz.comcontrastsecurity.comlevelup.gitconnected.com+274 more

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?✓#4

ratio0.95dimensionless ratio or multiplier (1.0 = no change)

rationale:

Final outcome: p10=0.45, p25=0.75, p50=0.95, p75=1.25, p90=1.7.

Factoring in the potential for significant structural efficiency gains driven by AI assistance alongside the risks of physical scaling walls modestly compressed the extreme tails of the distribution, refining the estimates to better capture the split between AI-driven acceleration and the risk of scaling stagnation.

Resolution Mechanics and Noise. Initial logic and parameters are validated regarding local changes in the rate o

Set against related questions, the estimates were slightly refined to better capture the split between AI-driven acceleration and the risk of scaling stagnation.

Weighed against 5 other forecasts.

digitalapplied.commlsysbook.aiuplevelteam.comepoch.aithenextweb.com+117 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?✓#5

ratio2.4dimensionless ratio or multiplier (1.0 = no change)

rationale:

2026 Baseline and the Market Scale The 2026 baseline for human-expert data generation is already a multi-billion-dollar category, driven by rapid vendor growth and increasing in-house efforts. Vendor-side evidence indicates massive scale: Mercor crossed a >$2B gross annualized revenue run rate by June 2026, with ~90% stemming from foundation-model labs 3 sources, while Surge, Scale, and Turing handle hundreds of millions to billions more in aggregate [65d9a5, e6aee7, e909bd, 4

Evaluated alongside related forecasts on data acquisition strategies, this estimate was slightly adjusted to reflect the inverse relationship between scaling expensive manufactured data and pivoting to cheaper real-work capture subsidies .

Weighed against 4 other forecasts.

opentools.ailemon.iorand.orgbankless.comfacet.com+351 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?✓#6

ratio4.1dimensionless ratio or multiplier (1.0 = no change)

rationale:

Final Estimate: p10=1.1, p25=2.1, p50=4.1, p75=11.0, p90=30.0. What is being measured Initial logic and parameters are validated regarding the isolation of pure sample efficiency from overall compute efficiency and historical decomposition evidence 2 sources. Standard processing applied, treating this entirely as established context. Current frontier practice points to rising token budgets Established context confirmed for the trend of frontier practice heading toward higher token

Tying the potential for extreme sample-efficiency gains to our expectations for broader AI-driven R&D automation slightly smoothed the distribution while preserving a median expectation of a roughly 4x improvement.

Weighed against 5 other forecasts.

toloka.aimlsysbook.aimedium.datadriveninvestor.comarxiv.orgpod.wave.co+266 more

Ask a followup

Ask a followup