“Well, no. Augmentation is way easier, because you don’t have to solve the whole job.”
Our reading, the sentence that was checked:
Using AI to augment workers is much easier than replacing them, because augmentation does not require solving the whole job.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Three baffling claims about AI and machine learning in four days, statistical errors in top journals, and claims from Yann LeCun that you should not believe., paragraph 23.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-too-much-benefit-of-the-doubt-cd49a0-a1
We recorded 13 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 2022-10-16. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Causal claim
- 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)
From the speaker
- Augmentation is way easier
- because you don’t have to solve the whole job
Our interpretation
- The plain reading compares useful assistance with autonomous performance of the same occupation's full duties.
- Easier means materially lower capability and engineering requirements at a comparable standard of useful, reliable service.
- The proposed mechanism is that people retain duties the machine cannot perform. Mere adoption of assistance does not establish this mechanism or its size.
- The claim concerns the technical situation at publication, rather than an unconditional law about every occupation.
Counts as supported if: Comparisons or task analyses show that retaining difficult duties with people materially reduces the capabilities and engineering effort needed for useful assistance relative to full replacement.
Counts as contradicted if: Comparable evidence shows that effective assistance generally requires essentially the same capabilities and effort as full replacement, or that human-machine coordination removes the claimed advantage.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Does the evidence support the cause?
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
The proposed mechanism is plausible, and a [randomized study](https://pubmed.ncbi.nlm.nih.gov/35703413/) shows that useful assistance can leave interpretation with radiologists. That does not establish how much easier the system was to build than a replacement. The [ESR survey](https://pubmed.ncbi.nlm.nih.gov/35727355/) also shows that assistance often brought no reported workload reduction. Neither source directly tests the asserted causal comparison.
Limitations
- Successful assistance is not itself a measurement of development difficulty.
- Lack of workload reduction does not prove assistance is as difficult as replacement.
- The evidence covers a narrow set of medical applications.
Evidence (2 items)
Context only: American Journal of Roentgenology; abstract hosted by PubMed,
improved radiologist efficiency
Why it matters here: The randomized study demonstrates useful assistance on chest CT examinations while humans continue interpreting them. It measures reading time, not the relative effort of developing assistance and full replacement.
- Publisher
- American Journal of Roentgenology; abstract hosted by PubMed
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
Context only: Insights into Imaging / European Society of Radiology; abstract hosted by PubMed,
no such effect
Why it matters here: The survey reports limited workload benefits for many users of diagnostic assistance. This qualifies an inference from tool availability to practical ease, but it does not compare development requirements with full replacement; precise survey fieldwork dates were not identified.
- Publisher
- Insights into Imaging / European Society of Radiology; abstract hosted by PubMed
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
What would settle it
A documented comparison of assistance and full-job automation requirements, showing which difficult capabilities remain with people and the resulting difference in development or deployment effort.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Three baffling claims about AI and machine learning in four days, statistical errors in top journals, and claims from Yann LeCun that you should not believe.
Well, no. Augmentation is way easier, because you don’t have to solve the whole job.
- Where
- paragraph 23
- Published
- 2022-10-16
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-too-much-benefit-of-the-doubt-cd49a0-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-too-much-benefit-of-the-doubt-cd49a0-a1.