Research library
Robotics
View the four-date completion graphs and reconstruction notes.
Promise and evidence
Robots could extend AI assistance into manufacturing, maintenance, household work, and laboratories. BEHAVIOR provides a benchmark grounded in household activities. Simulation performance is a useful development measure but must be distinguished from dependable physical operation.
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) |
|---|---|---|
| Complete task success | 20% | 5 / 5 / 5 / 5 / 0 |
| Generalization | 35% | 10 / 10 / 5 / 5 / 5 |
| Intervention rate | 30% | 10 / 5 / 5 / 5 / 5 |
| Reliability | 30% | 10 / 5 / 5 / 5 / 5 |
| Skill acquisition | 30% | 10 / 5 / 5 / 5 / 5 |
| Useful productivity | 30% | 10 / 5 / 5 / 5 / 5 |
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 |
|---|---|
| Complete task success | Fully completed tasks / all assigned trials, judged against observable end states rather than partial motion success. |
| Intervention rate | Human rescues, remote-control episodes, and instruction corrections per 100 tasks, with labor minutes. |
| Generalization | Success on unseen buildings, object variants, layouts, and disturbances without retraining for the individual test. |
| Reliability | Operating hours between failures, recovery success, damage events, and task abandonment rate. |
| Useful productivity | Accepted physical outputs per hour and fully loaded dollars per output, including setup, maintenance, supervision, and depreciation. |
| Skill acquisition | Demonstrations, robot-hours, and human hours required to learn a new task to a prespecified reliability target. |
Evidence requirements
Use randomized task selection and complete trial logs or video, not selected demonstrations. Separate simulation from physical tests, and supervised operation from autonomy. Define what counts as damage or a near miss. Zero incidents in a small trial is not proof of a low underlying incident rate; report exposure and uncertainty.
Proposed milestones
Complete at least 90% of 200 randomized physical trials across 20 tasks and five unfamiliar environments, with all interventions disclosed.
Sustain 1,000 operating hours in a real workflow with fewer than one human rescue per 100 tasks and no serious incident; treat this as pilot evidence, not universal safety certification.
Reduce total cost per accepted output by at least 20% against the matched existing workflow, without worsening preregistered quality or incident limits.
Use laboratory robots to shorten independently reproduced experiment cycles across three research campaigns. Physical productivity supports a singularity pathway most directly when it accelerates the creation and testing of new technology.