Research library
Physics and scientific discovery
View the four-date completion graphs and reconstruction notes.
Promise and evidence
AI could help discover compact explanatory models, select informative experiments, and accelerate simulation. AI Feynman demonstrates symbolic regression on physical equations. Recovering a known equation is a capability benchmark, not discovery of a new law of nature.
Apply the shared measurement protocol. Targets are proposed.
Dated baseline coverage
All linked measurable-outcome files below contain an evidence baseline reviewed November 7, 2026, with source dates, findings, limitations and follow-up measurements. These distinguish measured results from proxies and unknown values; they do not assert that proposed milestones are achieved. See the baseline coverage notes.
Progress estimates
Judgment-based estimates as of November 7, 2026, using the same 0–100% scale and five-part rubric as mathematics. 0% anchors the period before ChatGPT's public release in 2022; 100% requires AI to manage this field's continuing advancement without human participation. Confidence is low. These are evidence-based judgments, not empirical completion measurements, forecasts or percentages of jobs replaced.
| Measurable outcome | Estimated progress | Execution / breadth / agenda / validation / improvement (each out of 20) |
|---|---|---|
| Equation recovery | 0% | 0 / 0 / 0 / 0 / 0 |
| Experiment efficiency | 25% | 5 / 5 / 5 / 5 / 5 |
| Experimental confirmation | 0% | 0 / 0 / 0 / 0 / 0 |
| Novel prediction | 20% | 5 / 5 / 5 / 5 / 0 |
| Simulation efficiency | 40% | 15 / 5 / 5 / 10 / 5 |
| Theoretical novelty | 20% | 5 / 5 / 5 / 5 / 0 |
Each linked outcome explains its score, cites its dated evidence baseline and identifies the next measurement. Zero indicates no qualifying post-anchor progress established by the reviewed evidence for that outcome, not proof of no relevant capability. Do not average these overlapping indicators into a field-wide percentage or interpret their differences as a precise ranking of fields. All outcomes reaching 100% would still require a separate field-wide demonstration of sustained autonomous advancement.
Measurable features
| Feature | Measurement and acceptance rule |
|---|---|
| Equation recovery | Correct models / all held-out benchmark systems, with equivalent forms recognized and noise, missing variables, and budgets specified. |
| Novel prediction | Quantitative predictions registered before data collection that distinguish the proposed model from the strongest established alternatives. |
| Experimental confirmation | Predicted effects confirmed / all preregistered effects tested, with effect sizes, uncertainty, controls, and independent replication. |
| Simulation efficiency | Time, energy, and dollars per solution at a fixed independently checked error tolerance, including training and setup costs. |
| Experiment efficiency | Experiments, facility time, and total cost needed to resolve a prespecified scientific question. |
| Theoretical novelty | Independent expert classification of genuinely new explanatory content versus interpolation, reparameterization, or rediscovery. |
Evidence requirements
Release equations, assumptions, units, code, and data needed to reproduce the result. Verify dimensional consistency and applicable physical constraints. Test outside the fitted range when scientifically appropriate. Use blind analysis and correction for multiple hypothesis searches. Mathematical derivations can use formal proof where practical, but formal correctness does not substitute for empirical validation of physical assumptions.
Proposed milestones
Recover models on at least 80 of 100 held-out noisy systems at a frozen budget, including comparisons with established symbolic methods.
Produce one substantively new, expert-reviewed model that predicts a previously unmeasured effect and survives a preregistered experimental test.
Replicate the distinguishing prediction with an independent team or instrument and quantify where the model fails.
Across three research campaigns, reduce total time or experimental cost by at least 25% without reducing evidential standards. Repeatedly improving the research process is a stronger acceleration signal than a single striking result.