“Be the person who tells people that Sam Altman burns $5 billion a year on unsustainable software that destroys the environment and is built upon the large-scale larceny of creative works because he's desperate for power.”
Our reading, the sentence that was checked:
OpenAI's software is built upon the large-scale theft of creative works.
Assessment: Supported Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Never Forgive Them, paragraph 143.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-never-forgive-them-fa0696-2-a1
We recorded 9 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2024-12-16. 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)
From the speaker
- software
- is built upon
- the large-scale larceny of creative works
Our interpretation
- Only the supplied normalized proposition about OpenAI's use of creative works is assessed. The surrounding financial, environmental and motive assertions are outside this claim.
- The ordinary contextual reading of larceny is substantial copying and use of creative works without creators' permission, including pirated source copies. It is not treated as an assertion of a criminal conviction.
- Built upon refers to the historical development and training of OpenAI's software, not a claim that every model used every disputed dataset.
- The target period covers development before the post. Its start is unspecified because the speaker supplied no starting date.
- Later disclosures may establish what happened before publication. Unauthorized copying and the legal question of fair use are assessed separately.
Counts as supported if: Admissions, reliable training records or court findings establish substantial use of creative works obtained without rightsholders' permission in developing OpenAI models before the post.
Counts as contradicted if: Reliable evidence establishes that the relevant creative works were licensed, public domain or otherwise supplied with permission, and that the alleged substantial unauthorized use did not occur.
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
Supported on the ordinary factual reading of substantial copying and use without creators' permission. The filings quote an internal admission that GPT-3 was trained on pirated material and present evidence of a download exceeding 100,000 books. The court record corroborates the existence of pirated-book acquisition, while distinguishing allegations from undisputed facts. This does not establish criminal larceny or settle the fair-use defense. A reading requiring a final infringement judgment would remain unresolved.
Limitations
- The interpretation of larceny determines the result.
- Training use and scale rely partly on documents presented by an interested litigant.
- Support concerns historical model development, not every model or every training source.
- Public availability of the September filings is conservatively dated to the Authors Guild's September 21 publication.
Evidence (4 items)
Supports the claim: Class plaintiffs in In re OpenAI Copyright Infringement Litigation,
We trained GPT-3 on pirated stuff!
Why it matters here: The filed memorandum quotes an internal admission connecting pirated material to GPT-3 training. This is documentary evidence presented by plaintiffs, not a judicial determination that the training infringed copyright.
- Publisher
- Class plaintiffs in In re OpenAI Copyright Infringement Litigation
- Kind of source
- Regulatory filing
- Describes events of
- Public from
Supports the claim: Class plaintiffs in In re OpenAI Copyright Infringement Litigation,
117,859 books.
Why it matters here: The statement cites deposition and expert evidence for the scale of the LibGen download. Together with the training admission, it supports substantial unauthorized appropriation; the precise count remains the plaintiffs' evidentiary presentation.
- Publisher
- Class plaintiffs in In re OpenAI Copyright Infringement Litigation
- Kind of source
- Regulatory filing
- Describes events of
- Public from
Context only: US District Court, Southern District of New York; PDF hosted by Ars Technica,
the only training datasets
Why it matters here: The order records deletion of the training datasets in 2022 and separately describes the 2018 download of pirated LibGen books as undisputed. It concerns discovery and privilege, not a final infringement judgment.
- Publisher
- US District Court, Southern District of New York; PDF hosted by Ars Technica
- Kind of source
- Regulatory filing
- Describes events of
- Public from
Context only: OpenAI,
is fair use
Why it matters here: OpenAI disputes the legal characterization of unlicensed training and also describes publisher partnerships elsewhere on this page. Its position prevents treating absence of permission alone as a settled infringement finding, but does not deny every instance of unlicensed use.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
A merits judgment would resolve the stronger legal reading; underlying exhibits and responses to the factual submissions would strengthen verification of scale and model coverage.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Never Forgive Them
Be the person who tells people that Sam Altman burns $5 billion a year on unsustainable software that destroys the environment and is built upon the large-scale larceny of creative works because he's desperate for power.
- Where
- paragraph 143
- Published
- 2024-12-16
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-never-forgive-them-fa0696-2-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.