Some recent news articles discuss updates to our AI timelines since AI 2027, most notably our new timelines and takeoff model, the AI Futures Model (see blog post announcement).[1] While we’re glad to see broader discussion of AI timelines, these articles make substantial errors in their reporting. Please don’t assume...
We’ve significantly upgraded our timelines and takeoff model! It predicts when AIs will reach key capability milestones: for example, Automated Coder / AC (full automation of coding) and superintelligence / ASI (much better than the best humans at virtually all cognitive tasks). This post will briefly explain how the model...
Some recent news articles discuss updates to our AI timelines since AI 2027, most notably our new timelines and takeoff model, the AI Futures Model (see blog post announcement).[1] While we’re glad to see broader discussion of AI timelines, these articles make substantial errors in their reporting. Please don’t assume that their contents accurately represent things we’ve written or believe! This post aims to clarify our past and current views.[2]
The articles in question include:
the method of estimating hardware vs software share seems biased in the direction of exaggerating hardware share. (because training compute and software efficiency both tend to increase over time, so doing a univariate regression on training compute will include some of the effect from software improvements. so subtracting off that coefficient from the total will underestimate the effect of software improvements / “algorithmic progress”.)
From the paper:
Comparing these two estimates allows us to decompose the total gain into hardware and software components. By subtracting the compute effect (0.048) from the total effect (0.083), we isolate a residual of 0.035. This residual represents algorithmic progress, an economic catch-all for improvements in model architecture, software optimization, and user learning—effectively the Solow residual of AI production. In percentage terms, this decomposition suggests that compute scaling drives approximately 56% of the total reduction in time, while algorithmic advancements account for the remaining 44%.
Yep! Here's an example where the 50% horizon and 80% horizon can be lower for an agent whose success profile dominates another agent (i.e. higher success rate at all task lengths), even for
(1) monotone nonincreasing success rates (i.e. longer tasks are harder)
(2) success rate of 1 at minimum task length
(3) success rate of 0 at maximum task length
before points are
[(0,1), (1, 1/15), (2, 0), (3,0)]
after points are
[(0,1), (1, 0.1), (2, 0.1), (3, 1/15)]
https://www.desmos.com/calculator/nqwn6ofmzq
Progress probably has sped up in the past couple of years. And training compute scaling has, if anything, slowed down (it hasn't accelerated, anyway). So yes, I think "software progress" probably has sped up in the past couple of years.
I haven't looked into whether you can see the algorithmic progress speedup in the ECI data using the methodology I was describing. The data would be very sparse if you e.g. tried to restrict to pre-2024 models for greater alignment with the Algorithmic Progress in Language Models paper, which is where the 3x per year number comes from.
Also, that 3x per year number is only measuring pre-training improvements. Post-training (1) didn't really exist before 2022 and (2) was notably accelerated in 2024 by the introduction of RLVR. I wouldn't be confident in whether pre-training algorithmic progress alone is much faster than 3x per year today. (as rumor would have it, there's substantial divergence between the different AGI companies on the rate of pretraining progress.)
Thanks!
There are a lot of parameters. Maybe this is necessary but it's a bit overwhelming and requires me to trust whoever estimated the parameters, as well as the modeling choices.
Yep. If you haven't seen already, we have a basic sensitivity analysis here. Some of the parameters matter much more than others. The five main sliders on the webpage correspond to the most impactful ones (for timelines as well as takeoff.) There are also some checkboxes in the Advanced Parameters section that let you disable certain features to simplify the model.
Regarding form factor / conciseness: thanks for the feedback! Seems like people have widely varying opinions here. Would you prefer a form factor... (read 786 more words →)
I think this is an interesting analysis. But it seems like you're updating more strongly on it than I am. Here are some thoughts.
The Forethought SIE model doesn't seem to apply well to this data:
In the Forethought model (without the ceiling add-on), the growth rate of "cognitive labor" is assumed to equal the growth rate of "software efficiency", which is in turn assumed to be proportional to the growth rate of cumulative "cognitive labor" (no matter the fixed level of compute). Whether this fooms or fizzles is determined only by whether this constant of proportionality is above or below 1.
In this framework, and with fixed experiment compute, the game is to find a trend... (read 506 more words →)
We’ve significantly upgraded our timelines and takeoff model! It predicts when AIs will reach key capability milestones: for example, Automated Coder / AC (full automation of coding) and superintelligence / ASI (much better than the best humans at virtually all cognitive tasks). This post will briefly explain how the model works, present our timelines and takeoff forecasts, and compare it to our previous (AI 2027) models (spoiler: the AI Futures Model predicts longer timelines to full coding automation than our previous model by about 3-5 years, in significant part due to being less bullish on pre-full-automation AI R&D speedups). Added Jan 2026: see here for clarifications regarding how our forecasts have changed... (read 7337 more words →)
Registering that I don't expect GPT-5 to be "the biggest model release of the year," for various reasons. I would guess (based on the cost and speed) that the model is GPT-4.1-sized. Conditional on this, the total training compute is likely to be below the state of the art.
For what it's worth, the model showcased in December (then called o3) seems to be completely different from the model that METR benchmarked (now called o3).
This is such a cool method! I am really curious about applying this method to Anthropic's models. Would you mind sharing the script / data you used?