How fast is AI?
Menu

The library

The research behind the tracker.

The claims here trace back to published work, and where a number is an estimate, digitization, or projection, the page says so. The core weak-link framing, forecast model, computer-share figure, 150-year growth frame, infinite-automation result, and self-driving example draw directly on the work of Charles I. Jones and Jones & Tonetti, while Daron Acemoglu's estimates anchor the skeptical end. Alongside them sits a wider, recent literature on how fast AI is actually moving: capability and compute trends, broader growth models, and field studies of productivity.

Economic growth

  • Charles I. Jonesin prep.
    A.I. and Our Economic Future

    Working paper, in prep. for the Journal of Economic Perspectives

    Direct non-technical source for the infinite-automation 1/(1-s) result, the self-driving-cars example, and the weak-link view of AI growth.

  • Charles I. Jones2026
    "A.I. and Our Economic Future," Professor Chad Jones

    Stanford Graduate School of Business (YouTube)

    Stanford GSB talk presenting the weak-link growth view and figures adapted on this page, including the computer-share and forecast charts.

  • Paul M. Romer1990
    Endogenous Technological Change

    Journal of Political Economy 98(5): S71–S102

    Ideas as the engine of long-run growth (Nobel 2018). The flywheel that 'wants to explode'.

  • Bloom, Jones, Van Reenen & Webb2020
    Are Ideas Getting Harder to Find?

    American Economic Review 110(4): 1104–1144

    Within any technology, ideas get harder to find (the steam engine runs out of steam), so each wave buys ~50 years of 2% growth.

  • Aghion, Jones & Jones2019
    Artificial Intelligence and Economic Growth

    in The Economics of Artificial Intelligence, Univ. of Chicago Press (NBER WP 23928)

    The earlier automation + weak-links growth model behind the baseline forecast scenarios.

  • Daron Acemoglu2024
    The Simple Macroeconomics of AI

    Economic Policy 40(121): 13–58 (NBER WP 32487)

    The conservative end of the dial: predicted TFP gains below 0.53% over 10 years, with the initial task-based estimate no more than 0.66%.

  • Philip Trammell & Anton Korinek2023
    Economic Growth under Transformative AI

    NBER Working Paper 31815

    A survey synthesizing the AI-and-growth literature: fully automating production can break the Kaldor facts, raise the growth rate, and lower the labor share. The wider map the weak-link view sits inside.

  • Ege Erdil & Tamay Besiroglu2023
    Explosive Growth from AI Automation: A Review of the Arguments

    arXiv:2309.11690

    The case that broad automation could accelerate growth by roughly tenfold, weighed against nine counterarguments. The most rigorous steelman for the fast end of the dial.

Weak links & fragility

Measuring AI's pace

  • Kwa, West, Becker et al. (METR)2025
    Measuring AI Ability to Complete Long Tasks

    METR; arXiv:2503.14499

    The length of task an AI agent can finish with 50% reliability roughly doubled every 7 months in the original paper; METR's current dashboard estimates about 6.2 months all-time. A direct, empirical answer to 'how fast,' independent of the growth model.

  • Sevilla, Besiroglu, Cottier, You et al. (Epoch AI)2024
    Can AI Scaling Continue Through 2030?

    Epoch AI

    Training compute has grown about 4x per year; power, chips, data, and latency still leave room for runs roughly 10,000x larger by 2030. The cheap input keeps getting cheaper; that is exactly why it is not the scarce one.

  • Luke Emberson & David Roodman (Epoch AI)2026
    The Plunging Price of Thought

    Epoch AI (September 2026)

    Across five math, science, and puzzle benchmarks since 2023, the per-question inference cost of achieving a given performance level (counting reasoning tokens) fell about 47% per quarter, or 13x per year, fastest right after a level debuts (75x/year) and slowing to 4.7x/year two years later.

  • Maslej et al. (Stanford HAI)2025
    The 2025 AI Index Report

    Stanford Institute for Human-Centered AI

    Benchmark scores jumped in a single year (GPQA +48.9 points), training compute doubles about every five months, and the inference cost of GPT-3.5-level output fell roughly 280x in two years.

  • Handa, Tamkin, McCain, Huang, Durmus et al. (Anthropic)2025
    Which Economic Tasks are Performed with AI? Evidence from Millions of Claude Conversations

    Anthropic Economic Index; arXiv:2503.04761

    Claude.ai conversations mapped onto O*NET tasks show current usage led by computer/mathematical work and writing/editing tasks, with explicit augmentation-vs-automation splits. A live read on which links are actually moving.

Labor & distribution

  • Daron Acemoglu2026
    Will AI Replace Workers? Not If We Build It Right.

    The Humanist Review of AI, Issue 1 (essay)

    An editorial argument for applications that complement workers rather than only mimic them. Links capability to application design and adoption; its US forecasts are estimates, not observed job losses or global outcomes.

  • Karabarbounis & Neiman2014
    The Global Decline of the Labor Share

    Quarterly Journal of Economics 129(1): 61–103

    Context for the capital-vs-labor split that the scenarios track to 100% / 0%.

  • Brynjolfsson, Rock & Syverson2021
    The Productivity J-Curve

    American Economic Journal: Macroeconomics 13(1): 333–372

    Why measured productivity lags transformative tech: adoption races up the S-curve while output sits in the J-curve trough.

  • Becker, Rush, Barnes & Rein (METR)2025
    Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity

    arXiv:2507.09089 (randomized experiment)

    16 experienced developers completed 246 tasks in familiar open-source repositories. With early-2025 AI tools they took 19% longer. A result about those tools, developers, and tasks, not a general verdict on coding assistance or later models.

  • Agarwal, Moehring, Rajpurkar & Salz2023
    Combining Human Expertise with Artificial Intelligence: Experimental Evidence from Radiology

    NBER Working Paper 31422

    An experiment on how radiologists combine AI predictions with contextual information. Human-plus-AI performance depends on how the information is used; complementarity is a design problem, not an automatic consequence of access to a model.

  • Brynjolfsson, Li & Raymond2025
    Generative AI at Work

    Quarterly Journal of Economics (2025); NBER WP 31161

    A field study of 5,179 support agents: AI raised issues resolved per hour 14% on average and 34% for novices, with little effect on experts. The gains land first where skill was scarce.

  • Shakked Noy & Whitney Zhang2023
    Experimental Evidence on the Productivity Effects of Generative AI

    Science 381(6654): 187–192

    In a writing experiment, ChatGPT cut time 40% and raised quality 18%, narrowing the gap between weaker and stronger writers. Task-level speed-ups are real even where measured GDP is slow to move.

  • Dell'Acqua, McFowland, Mollick, Lifshitz-Assaf, Kellogg et al.2023
    Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality

    Harvard Business School Working Paper 24-013

    In a randomized experiment with 758 BCG consultants, AI raised task speed by 25% and quality by 40% inside its capability boundary, but reduced accuracy by 19 percentage points on a task designed just outside it. Shows why human verification remains a critical link.

  • Alexander Bick, Adam Blandin & David J. Deming2026
    The Rapid Adoption of Generative AI

    Management Science 72(10): 8440–8456 (NBER WP 32966)

    Nationally representative US surveys find workplace adoption of generative AI outpaced the personal computer and internet at comparable stages after launch, assisting 1–5% of work hours with self-reported time savings equivalent to 1.4% of total hours.

  • Anders Humlum & Emilie Vestergaard2025
    Large Language Models, Small Labor Market Effects

    NBER Working Paper 33777

    Links adoption surveys of 25,000 workers and 7,000 workplaces across 11 exposed occupations to administrative labor records in Denmark. Finds widespread adoption and ~3% average time savings, but precise null effects on earnings and recorded work hours two years on.

  • Erik Brynjolfsson, Bharat Chandar & Ruyu Chen2025
    Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of AI

    Stanford Digital Economy Lab

    Using ADP payroll records, early-career workers (ages 22–25) in the most AI-exposed US occupations saw about a 13% relative decline in employment since late 2022, while older workers in the same occupations held steady.

  • McElheran, Yang, Brynjolfsson & Kroff2026
    The Adoption of Industrial AI in America

    AEA Papers and Proceedings 116: 20–25

    US Census firm-level evidence on industrial AI adoption: early adopters tend to be larger, younger, multi-unit firms investing in complementary organizational practices, consistent with a firm-level productivity J-curve.

Risk & the fast-takeoff case

  • Dario Amodei2024
    Machines of Loving Grace

    Essay (darioamodei.com)

    The optimistic 'country of geniuses in a datacenter' vision, useful context for the aggressive side of the forecast debate.

Organizational change