“We are not even close to machines that exceed the capacity of typical humans to cope with unusual circumstances while driving”
Our reading, the sentence that was checked:
Machines are not close to exceeding typical humans at coping with unusual circumstances while driving.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, No, Virginia, AGI is not imminent, paragraph 17.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-no-virginia-agi-is-not-imminent-43e8b7-a1
We recorded 20 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2023-08-20. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
Conditions
- unusual circumstances (qualifier, stated by the speaker)
From the speaker
- We are not even close
- machines that exceed the capacity of typical humans
- cope with unusual circumstances while driving
Our interpretation
- This describes driving capability at publication, rather than predicting a date for future capability.
- The ordinary reading is that a substantial capability gap remained before automated driving broadly exceeded ordinary human drivers in unexpected or uncommon road situations.
- Unusual circumstances include unexpected road hazards, emergency scenes, and crash-imminent situations. No single category is treated as representative of all unusual driving.
- The target period is the current calendar year through August 20, 2023. The claim concerns the contemporary technology state and identifies no reporting cycle.
- The phrase 'unusual circumstances' specifies the comparison's scope; it is not a conditional forecast antecedent.
Counts as supported if: Representative, comparable evaluations of unusual driving situations show a substantial remaining capability shortfall relative to typical human drivers.
Counts as contradicted if: Representative, comparable evaluations show that automated systems were already near or above typical human performance across a broad range of unusual driving situations.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
Official records document difficulty around emergency scenes, but they do not measure a gap against typical humans. Counterevidence includes strong performance in selected crash simulations and lower aggregate insurance claim rates. Neither establishes broad performance across unusual driving situations. The available evidence therefore does not settle the stronger claim that machines were 'not even close'.
Limitations
- Official complaints document incidents rather than controlled comparisons.
- Waymo's simulation findings are company-reported and scenario-specific.
- The insurance comparison was published after the post but concerns data evaluated before it.
- Later technological improvements cannot determine the truth of this historical observation.
Evidence (3 items)
Context only: San Francisco Municipal Transportation Agency, San Francisco County Transportation Authority, and San Francisco Planning,
emergency response
Why it matters here: The filing describes continuing difficulties, including a driverless Waymo blocking a fire truck on April 25. It supplies no matched human exposure or performance baseline, so it cannot establish how far machines were from typical humans.
- Publisher
- San Francisco Municipal Transportation Agency, San Francisco County Transportation Authority, and San Francisco Planning
- Kind of source
- Regulatory filing
- Describes events of
- Public from
Context only: Waymo,
consistently exceeded that high benchmark.
Why it matters here: Waymo reported outperforming an attentive-human reference model in reconstructed fatal-crash simulations. This challenges a sweeping inference of poor unusual-situation capability, but the selected simulations predate the target period and do not represent all unusual driving.
- Publisher
- Waymo
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: arXiv; Luigi Di Lillo and colleagues,
This study compares the safety of autonomous- and human drivers.
Why it matters here: The study reports lower bodily-injury and property-damage claim rates for Waymo against mileage- and location-calibrated human baselines. Its data were evaluated as of August 1, but aggregate insurance outcomes do not separately measure unusual-situation capability.
- Where
- Abstract
- Publisher
- arXiv; Luigi Di Lillo and colleagues
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
A representative evaluation of systems available by August 20, 2023 against ordinary human drivers on matched unusual road situations, with enough coverage and uncertainty estimates to establish whether a substantial capability gap remained.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, No, Virginia, AGI is not imminent
We are not even close to machines that exceed the capacity of typical humans to cope with unusual circumstances while driving
- Where
- paragraph 17
- Published
- 2023-08-20
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-no-virginia-agi-is-not-imminent-43e8b7-a1: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-no-virginia-agi-is-not-imminent-43e8b7-a1.