“My concern is that I believe we’re in the midst of a subprime AI crisis, where thousands of companies have integrated generative AI at prices that are far from stable, and even further from profitable.”
Our reading, the sentence that was checked:
We are in a subprime AI crisis in which thousands of companies have integrated generative AI at unstable, unprofitable prices.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, The Subprime AI Crisis, paragraph 71.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-subprimeai-0c9a66-a2
We recorded 50 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 2024-09-16. 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
- we’re in the midst
- thousands of companies
- have integrated generative AI
- prices that are far from stable
- even further from profitable
Our interpretation
- I assess the factual description underlying the crisis metaphor.
- Thousands means at least 2,000 companies with business integrations.
- The surrounding discussion makes profitability refer to the model providers' prices, rather than the customers' profitability.
- The ordinary reading requires materially changing or unsustainable prices and losses attributable to supplying the integrated services. Company-wide losses alone do not establish this.
- For this ongoing observation, I use April through June 2024, the latest completed calendar quarter before publication.
Counts as supported if: Evidence identifies thousands of business integrations and establishes that their relevant provider prices were materially unstable and loss-making during the target period.
Counts as contradicted if: Representative evidence establishes that the claimed population was below thousands, or that the relevant prices were stable and profitable on a clearly specified, comparable cost basis.
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
On April 23, 2024, AWS reported tens of thousands of customers using Bedrock to build and scale applications; this does not count completed business integrations [e1]. An analysis published August 27 estimated a 75% gross margin for OpenAI's direct GPT-4-class API traffic in June 2024, excluding training costs and employee salaries [e2]. In its August 7, 2026 retrospective, FutureSearch said its margin estimates remained neither confirmed nor refuted [e3]. These items establish neither thousands of completed integrations at materially unstable, loss-making prices during April–June 2024 nor representative evidence of stable, profitable prices on a comparable cost basis [e1] [e2] [e3].
Limitations
- AWS's customer count does not identify how many customers had completed business integrations [e1].
- The margin estimate covers only direct GPT-4-class API traffic, excludes training and salaries, and remains unverified [e2] [e3].
- The packet lacks price histories and attributable service costs linked to the claimed population during April–June 2024.
Evidence (3 items)
Context only: Amazon Web Services,
tens of thousands of customers
Why it matters here: The adoption scale is consistent with thousands of business users. It does not establish how many completed integrations depended on unstable, loss-making prices.
- Where
- Opening paragraph
- Publisher
- Amazon Web Services
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: FutureSearch researchers, published on LessWrong,
was 75%.
Why it matters here: This estimate challenges the inference that company losses prove API calls were sold below their serving costs. Its narrower cost basis and uncertain inputs prevent a decisive contradiction.
- Publisher
- FutureSearch researchers, published on LessWrong
- Kind of source
- Independent measurement
- Measures
- OpenAI, Estimated direct GPT-4-class API gross margin excluding training and employee salaries percent, estimate, 2024-06-01 to 2024-06-30
- Describes events of
- Public from
Context only: FutureSearch,
not confirmed and not refuted.
Why it matters here: The analysts' retrospective explicitly limits the evidentiary weight of their earlier margin estimates. This later review is context, not an outcome observation within the target quarter.
- Publisher
- FutureSearch
- Kind of source
- Secondary summary
- Describes events of
- Public from
What would settle it
Records for April–June 2024 linking completed business integrations to comparable provider price histories, service revenue and attributable costs, with an explicit accounting basis. These would establish the population, price stability and service profitability.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, The Subprime AI Crisis
My concern is that I believe we’re in the midst of a subprime AI crisis, where thousands of companies have integrated generative AI at prices that are far from stable, and even further from profitable.
- Where
- paragraph 71
- Published
- 2024-09-16
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-subprimeai-0c9a66-a2: 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.
HeldBy this site
The final-text check failed: citations_correct.
About assessment wyea-subprimeai-0c9a66-a1, which was not published.