Research note: A simpler AI timelines model predicts 99% AI R&D automation in ~2032

2026-02-12 · Source: AI Alignment Forum · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Advanced, long

Summary

A simplified AI forecasting model predicts over 99% automation of AI research and development (R&D) by late 2032, based on current compute growth and algorithmic progress rates. This model, a more robust and understandable version of the AI Futures Model (AIFM), uses 8 parameters compared to AIFM's 33. It projects a 1,000x to 10,000,000x increase in AI efficiency and 300x-3,000x research output by 2035. The model makes conservative assumptions, including no full automation (automation approaches but never reaches 100%) and no substitutability between tasks (following Amdahl's law). It defines AI development dynamics using Cobb-Douglas for research progress, a Jones model for software efficiency, and a sigmoid function for the fraction of automatable tasks based on effective compute.

Key takeaway

For AI Researchers and Research Scientists evaluating future AI development timelines, this model suggests that superhuman AI researchers are likely before 2036, even without aggressive assumptions like full automation or automated research taste. You should consider the implications of a median 99% AI R&D automation by mid-2032 on your long-term project planning and resource allocation, particularly regarding the projected 300x-3,000x increase in research output.

Key insights

Simplified AI forecasting suggests >99% AI R&D automation by 2032, even with conservative assumptions.

Principles

• Simpler models can be more robust.
• Automation velocity and diminishing returns are key uncertainties.
• Compute growth significantly impacts research output.

Method

The model assumes Cobb-Douglas research progress, Jones model for software efficiency, and a sigmoid for automation fraction based on log(effective compute), with human labor as the bottleneck.

In practice

• Focus on empirical data for automation velocity.
• Track uplift studies to refine automation fraction (f) estimates.

Topics

• AI Automation Forecasting
• AI R&D Automation
• Algorithmic Progress
• AI Futures Model
• Compute Growth

Code references

• tkwa/ai-takeoff-model

Best for: AI Researcher, AI Scientist, Research Scientist

Related on AIssential

• Experts give a 50% chance of AGI by 2050, 75% chance by the mid-2060s, and 95% chance by 2090 — Experts predict AGI has a 50% chance by 2050, rising to 95% by 2090.
• Forecasting Scientific Progress with Artificial Intelligence — Current AI systems cannot reliably forecast scientific progress or its timing, despite identifying plausible research directions.
• 🔮 Exponential View #573: Are the AI labs building for an intelligence explosion? — A frontier AI model training its successor by 2028 is a 60% possibility, shifting AI development to an industrial scaling problem.
• The next 36 months will be WILD — Expert consensus points to AGI/ASI by 2027-2028, driven by super-exponential AI progress and an irreversible industrial race.
• AI Futures Timelines and Takeoff Model: Dec 2025 Update — The AI Futures Model provides updated, transparent forecasts for AI milestones, emphasizing R&D automation and benchmark extrapolation.
• Leading AI expert delays timeline for its possible destruction of humanity — An AI expert has delayed his timeline for AI achieving autonomous coding and superintelligence, citing real-world complexities.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Alignment Forum.