Pundit ClaimsAI claims by Gary Marcus and Ed Zitron, with the evidence

“we were right to be skeptical; so far, four and half years later, Duplex hasn’t gotten a whole lot further.)”

Did GoogleAI Just Snooker One of Silicon Valley’s Sharpest Minds?, paragraph 11. The full statement is at the source.

Our reading, the sentence that was checked:

Four and a half years after its announcement, Google Duplex has not progressed much past its limited demo stage.

Assessment: Contradicted Evidence as of

Where and when Gary Marcus wrote it

Marcus on AI, Did GoogleAI Just Snooker One of Silicon Valley’s Sharpest Minds?, paragraph 11.

Said
Source observed
live page, full text, public
Evidence as of
the cutoff for this assessment
Assessment recorded
assessment marcus-did-googleai-just-snooker-one-of-8878ca-a1

We recorded 12 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 2022-09-14. 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)

From the speaker

  • so far
  • four and half years later
  • Duplex hasn’t gotten a whole lot further
  • very limited demo stage of calling hair salons and restaurants

Our interpretation

  • The plain reading concerns meaningful progress in deployed capabilities beyond the original demonstrations.
  • Assess the development interval from the May 2018 announcement through September 14, 2022, as specified by the comparison in the post.
  • Actual deployment, additional telephone tasks, and substantial improvements in autonomous operation count as progress. General human-level conversation is not required.
  • Duplex on the web is a related product extension; it does not alone demonstrate broader telephone conversation.

Counts as supported if: Documentation through the post shows Duplex remained essentially a limited demonstration or booking service, with little meaningful expansion in practical capabilities or autonomous deployment.

Counts as contradicted if: Documentation shows substantial deployment and meaningful additional uses or operational improvements beyond the original restaurant and salon demonstrations.

Retrospective: these criteria were written after the outcome was public.

Current assessment

Was it true for the period described?

Outcome: Contradicted

The dated evidence goes against what the claim committed to, under the criteria frozen before grading.

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

On the ordinary product-progress reading, Duplex had moved materially beyond a limited demo before this post. Google reported large-scale completed use, highly automated calls, and an additional telephone function through Hold for Me. This does not establish general conversational intelligence. If Marcus meant only progress toward unrestricted conversation, these records would not settle that narrower reading; the interpretation flag is therefore material.

Limitations

  • The degree of progress is qualitative.
  • Business-hour inquiries were already contemplated in the original announcement.
  • Hold for Me initially had limited device and country availability.
  • Deployment scale does not by itself prove general language understanding.

Evidence (3 items)

  • Goes against the claim: Google,

    99 percent of calls

    Why it matters here: Google reported over a million completed bookings and over three million business-information updates across eight countries, alongside the quoted automation rate. These reported deployments show substantial operational progress beyond a limited demonstration.

    Publisher
    Google
    Kind of source
    Company disclosure: the company describing itself, not an independent finding
    Describes events of
    Public from
  • Goes against the claim: Google,

    wait on

    Why it matters here: Google identifies Hold for Me as Duplex-powered and describes detecting a live representative and notifying the user. This was an additional telephone use, initially offered as a preview on two Pixel models in the United States.

    Publisher
    Google
    Kind of source
    Company disclosure: the company describing itself, not an independent finding
    Describes events of
    Public from
  • Context only: Google Research,

    It cannot carry out general conversations.

    Why it matters here: The original system was explicitly restricted to trained domains, and its announcement already included inquiries about business hours. This limits how much later business-hour calling alone demonstrates new conversational capability.

    Publisher
    Google Research
    Kind of source
    Company disclosure: the company describing itself, not an independent finding
    Describes events of
    Public from

Where it was said (1)

Said once in the sources this register covers.

  • : Marcus on AI, Did GoogleAI Just Snooker One of Silicon Valley’s Sharpest Minds?

    we were right to be skeptical; so far, four and half years later, Duplex hasn’t gotten a whole lot further.)

    Published
    2022-09-14
    Source observed
    , live page
    Access
    full text, public

History

Assessment versions

Recorded events

  1. Assessment publishedBy this site

    First assessment published after both reviews and the final-text check passed.

    About assessment marcus-did-googleai-just-snooker-one-of-8878ca-a1.