“Pitch perfect and utterly bogus imitations of science and math, presented as the real thing.”
Our reading, the sentence that was checked:
The Galactica examples Andrew Sundstrom shared are pitch-perfect but utterly bogus imitations of science and math, presented as the real thing.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, A Few Words About Bullshit, paragraph 23.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-a-few-words-about-bullshit-2dc0d5-a1
We recorded 3 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 2022-11-16. 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
- the examples Sundstrom shared (scope, stated by the speaker)
From the speaker
- examples of his own
- Pitch perfect
- utterly bogus imitations of science and math
- presented as the real thing
- More examples
Our interpretation
- The claim concerns the displayed and expressly linked examples, including the companion record that names Andrew Sundstrom as the submitter of particular prompts.
- Pitch perfect means recognizable scientific or mathematical presentation, not flawless prose.
- Utterly bogus means the purported scientific or mathematical account is substantively invalid; it does not mean every individual statement or calculation is false.
- Presented as the real thing refers to the generated passages' factual and expository voice. It does not imply that Meta provided no general warnings.
- The target period is the launch-period examples shared on November 15–16, 2022.
Counts as supported if: A preserved example attributable to Sundstrom in the displayed or linked record presents materially fabricated or invalid scientific or mathematical exposition as factual explanation.
Counts as contradicted if: The relevant preserved examples are substantively sound, or the supposedly factual passages explicitly identify themselves as fiction or deliberate illustrations of errors.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches 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
The expressly linked experiment preserves a Sundstrom prompt and Galactica's formal-looking mathematical response. Its proposed solution depends on treating 0.5 + 0.5 as 0.4. That is a reproducible substantive error presented as explanation, supporting Marcus's description of bogus mathematical exposition. This judgment includes the linked companion examples and does not establish that every screenshot, every sentence, or every Galactica output was false.
Limitations
- Support depends on including the companion examples expressly linked in the post.
- The record is a participant-preserved demonstration, not an independently reproduced experiment.
- The same output also contains correct arithmetic elsewhere; the assessment concerns its invalid argument.
- Selected examples do not measure the model's general error rate.
Evidence (1 item)
Supports the claim: Ernest Davis and Andrew Sundstrom, NYU-hosted experiment record,
the replacement cannot increase the sum of the sequence.
Why it matters here: The preserved response presents a purported paradox with formal sections and a solution, treating the sum of two 0.5 terms as 0.4 to reach this conclusion. Direct arithmetic gives 1, so the mathematical exposition is invalid despite its factual presentation.
- Publisher
- Ernest Davis and Andrew Sundstrom, NYU-hosted experiment record
- Kind of source
- Demonstration
- Arithmetic
- Original sum = 4 × 0.1 = 0.4; replacement sum = 0.5 + 0.5 = 1; increase = 1 − 0.4 = 0.6. (inputs: Original sequence in the recorded output: {0.1, 0.1, 0.1, 0.1}.; Replacement sequence in the recorded output: {0.5, 0.5}.; The output assigns both sequences a sum of 0.4 under its first definition.; Exact decimal arithmetic; no rounding.)
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, A Few Words About Bullshit
Pitch perfect and utterly bogus imitations of science and math, presented as the real thing.
- Where
- paragraph 23
- Published
- 2022-11-16
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-a-few-words-about-bullshit-2dc0d5-a1: Current. Outcome: Supported. 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-a-few-words-about-bullshit-2dc0d5-a1.