“If stopping training was all that it’d take to turn these businesses profitable, they’d have done it already, because inference would be a money-printer rather than a cursed object eating away at Altman and Amodei’s souls.”
Our reading, the sentence that was checked:
Stopping model training would not make OpenAI and Anthropic profitable; inference is not a money-printer.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, AI's Brokenomics, paragraph 88.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-brokenomics-2906d8-a1
We recorded 41 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 2026-06-15. 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
- stopping training
- turn these businesses profitable
- they’d have done it already
- inference would be a money-printer
Our interpretation
- The testable financial reading is that removing training expenses alone would leave OpenAI and Anthropic unprofitable.
- Hold contemporaneous revenue and non-training expenses constant. This does not assume that future revenue would survive indefinitely without new models.
- Profitable refers to the whole business after remaining expenses, not merely a positive inference gross margin.
- Use January through March 2026, the latest completed calendar quarter that evidence could report before the June 15 post. The post names no accounting period.
- The claim does not specify an accounting basis. It remains unknown; strong financial scoring is therefore unavailable under the stated rules.
- The assertion about what management would already have chosen is an argument, not independent evidence of the counterfactual financial result.
Counts as supported if: Comparable records for each company show a loss after removing only training expenses while retaining all other relevant expenses.
Counts as contradicted if: Comparable records show that removing training expenses alone would produce a profit for either named company, with any difference between the companies reported explicitly.
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
The required comparison is company profit after removing only training expenses. The available sources do not provide a matched Q1 reconciliation for both companies. Reported positive OpenAI gross margins weaken a broader suggestion that serving customers necessarily loses money, but do not settle the whole-business claim. Anthropic's expense disclosures and earlier estimates also do not supply the required comparison. The unspecified accounting basis independently prevents a supported or contradicted financial score.
Limitations
- The target quarter is an explicit interpretation of an undated observation.
- OpenAI evidence is secondary reporting about private records.
- Research-and-development expense, training expense, cash burn, and net loss are different measures.
Evidence (3 items)
Context only: The Decoder,
Gross margin did climb from 33 to 39 percent.
Why it matters here: The reported positive gross margin cautions against equating company losses with loss-making inference. It does not establish profit after all non-training expenses; no primary OpenAI financial statement was obtained.
- Publisher
- The Decoder
- Kind of source
- Secondary summary
- Measures
- OpenAI, Gross profit margin percent, estimate, 2026-01-01 to 2026-03-31
- Describes events of
- Public from
Context only: Anthropic, PBC / U.S. Securities and Exchange Commission,
personnel-related costs and allocated overhead costs
Why it matters here: The filing's functional expense notes distinguish customer-serving costs from broader research expenses. The cited presentation covers annual and half-year periods and does not supply the required Q1 training-only counterfactual.
- Publisher
- Anthropic, PBC / U.S. Securities and Exchange Commission
- Kind of source
- Regulatory filing
- Measures
- Anthropic, Definition of cost of revenue in functional expense presentation, USD millions, company reported, 2023-01-01 to 2026-06-30
- Describes events of
- Public from
Context only: Samaritan Research,
$4.1B
Why it matters here: This supplies an explicitly interpreted training-cost estimate from reporting on 2025. It is a different period and cannot establish either company's Q1 2026 profit after training is removed.
- Publisher
- Samaritan Research
- Kind of source
- Secondary summary
- Measures
- Anthropic, Estimated model-training compute expense, USD billions, estimate, 2025-01-01 to 2025-12-31
- Describes events of
- Public from
What would settle it
A stated accounting basis and comparable Q1 financial reconciliations for both companies showing revenue and every remaining expense after removing training alone.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, AI's Brokenomics
If stopping training was all that it’d take to turn these businesses profitable, they’d have done it already, because inference would be a money-printer rather than a cursed object eating away at Altman and Amodei’s souls.
- Where
- paragraph 88
- Published
- 2026-06-15
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-brokenomics-2906d8-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.