“I absolutely, 100% believe that both OpenAI and Anthropic are doing the same, and that neither of them has some magical way of making inference cheap enough to justify letting people burn thousands of dollars on a $100-a-month or $200-a-month subscription.”
Our reading, the sentence that was checked:
OpenAI and Anthropic let users burn thousands of dollars of compute on $100–$200 subscriptions and have no way to make inference cheap enough to justify it.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Revenge of The Business Idiot, paragraph 146.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-the-revenge-of-the-business-idiot-f22af6-a1
We recorded 30 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 2026-05-26. 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
- both OpenAI and Anthropic
- letting people burn thousands of dollars
- $100-a-month or $200-a-month subscription
- neither of them has some magical way of making inference cheap enough
Our interpretation
- The ordinary reading concerns actual inference expenses incurred by each provider for heavy subscribers, rather than the retail API price of equivalent usage.
- The claim concerns some heavy subscribers, not necessarily the average subscriber or every subscription.
- Use April 2026, the latest complete monthly billing period before publication, for this ongoing monthly-cost observation.
- Interpret 'justify' as covering the inference expense from the subscription payment. Broader strategic benefits are not measurable from this wording.
- No accounting basis or identifiable financial source is supplied for this assertion. The contract therefore records an unknown basis.
Counts as supported if: Comparable subscriber-level records show that both companies incurred thousands of dollars in actual monthly inference expenses for users paying $100 or $200, after applicable serving efficiencies.
Counts as contradicted if: Comparable records for either company establish that the alleged thousands of dollars represented retail API value and that actual inference expenses for the relevant heavy usage were covered by the subscription payment.
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 records make subsidization plausible but do not establish this claim for both companies in the target period. Altman's admission concerns January 2025. The Anthropic estimate is earlier secondary reporting with acknowledged cost uncertainty. Later usage estimates use retail API prices. None supplies matched April serving costs and subscription receipts, and the claim's accounting basis is unspecified.
Limitations
- Private subscriber-level cost records were unavailable.
- The selected billing month and meaning of 'justify' require interpretation.
- An unknown accounting basis prevents a supported or contradicted money verdict.
Evidence (3 items)
Context only: The Decoder,
subsidize its coding services.
Why it matters here: The article reports estimates of substantial subsidization but acknowledges that Anthropic's actual serving costs are unknown. This is secondary reporting about an earlier period, not verified April subscriber accounting.
- Where
- Opening paragraph
- Publisher
- The Decoder
- Kind of source
- Secondary summary
- Measures
- Anthropic, Estimated maximum monthly compute consumption of a $200 Claude Code subscription, USD dollars per subscriber per month, estimate, 2026-03-01 to 2026-03-07
- Describes events of
- Public from
Context only: Sam Altman, OpenAI CEO,
we are currently losing money on openai pro subscriptions!
Why it matters here: This establishes a company admission of subscription losses at an earlier date. It gives neither thousands of dollars in per-user costs nor April 2026 economics.
- Publisher
- Sam Altman, OpenAI CEO
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Measures
- OpenAI, Pro subscription profitability, cost definition unspecified, USD unspecified, company reported, ? to 2025-01-06
- Describes events of
- Public from
Context only: Tom's Hardware,
$14,000 in API pricing.
Why it matters here: This later report values usage at API prices. It does not measure the provider's actual inference expense and cannot establish the April claim.
- Where
- Article subtitle
- Publisher
- Tom's Hardware
- Kind of source
- News report
- Measures
- OpenAI, Maximum subscription usage valued at retail API prices, USD dollars per subscriber per month, estimate, 2026-06-01 to 2026-06-13
- Describes events of
- Public from
What would settle it
Verified April subscriber-level usage, actual serving costs and subscription receipts for both companies, with a stated cost basis and evidence about the relevant heavy-user group.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Revenge of The Business Idiot
I absolutely, 100% believe that both OpenAI and Anthropic are doing the same, and that neither of them has some magical way of making inference cheap enough to justify letting people burn thousands of dollars on a $100-a-month or $200-a-month subscription.
- Where
- paragraph 146
- Published
- 2026-05-26
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-the-revenge-of-the-business-idiot-f22af6-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 wyea-the-revenge-of-the-business-idiot-f22af6-a1.