“Tech's largest cash cow since the cloud computing boom of the 2000s is based on a technology that is impossibly unreliable, a technology with a potent inverted Midas touch that burns far more money than it makes.”
Our reading, the sentence that was checked:
Generative AI is impossibly unreliable and burns far more money than it makes.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Subprime Intelligence, paragraph 22.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-sam-altman-fried-ee3327-a2
We recorded 26 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 2024-02-19. 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
- impossibly unreliable
- burns far more money than it makes
- OpenAI made around $1.6 billion in revenue in 2023
- competitor Anthropic
Our interpretation
- Read the reliability assertion as saying factual outputs require substantial checking because material errors are common, rather than as a literal claim that every output is wrong.
- Read the financial assertion as total development and operating costs substantially exceeding realized revenue across generative AI suppliers.
- Use calendar 2023, the latest completed year and the year explicitly used in the surrounding financial comparison.
- The surrounding passage relies on press estimates, so the financial accounting basis is estimate.
- Results for OpenAI or Anthropic alone cannot establish a sector-wide result.
Counts as supported if: Representative evaluations show widespread material unreliability, and comparable 2023 financial evidence shows aggregate generative AI supplier costs substantially exceeding realized revenue.
Counts as contradicted if: Representative evidence establishes dependable performance across the claimed factual uses, or comparable sector-wide financial evidence shows costs did not substantially exceed revenue.
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
FinanceBench, published November 20, 2023, found that GPT-4-Turbo with retrieval incorrectly answered or refused 81% of its tested financial questions; this supports concern about that configuration and task, but does not settle the broad reliability assertion [e2]. A May 5, 2023 report estimated OpenAI's 2022 losses at $540 million; it does not establish supplier-wide costs relative to realized revenue in calendar 2023 [e3]. Neither item establishes or contradicts the claim at the required scope [e2, e3].
Limitations
- The FinanceBench result combines incorrect answers with refusals and does not establish reliability across ordinary factual uses [e2].
- The financial evidence covers one company's estimated 2022 losses, with no precise loss definition or comparable sector-wide 2023 revenue and total-cost figures [e3].
Evidence (2 items)
Context only: arXiv,
incorrectly answered or refused to answer
Why it matters here: This original evaluation supports concern about financial question answering in a particular configuration. It combines incorrect answers with refusals and does not establish that generative AI is generally unusable or uniformly unreliable.
- Where
- Abstract
- Publisher
- arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: The Decoder,
lost $540 million
Why it matters here: This supplies historical context for OpenAI's losses, using a secondary report because the original was inaccessible. It covers 2022 and one company, so it cannot settle the 2023 sector-wide comparison.
- Publisher
- The Decoder
- Kind of source
- Secondary summary
- Measures
- OpenAI, Reported losses; precise loss definition unspecified, USD dollars, estimate, 2022-01-01 to 2022-12-31
- Describes events of
- Public from
What would settle it
Representative 2023 factual-reliability evaluations identifying models, tasks, configurations and error definitions, plus comparable calendar-2023 realized revenue and total development and operating costs across a defined population of generative AI suppliers.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Subprime Intelligence
Tech's largest cash cow since the cloud computing boom of the 2000s is based on a technology that is impossibly unreliable, a technology with a potent inverted Midas touch that burns far more money than it makes.
- Where
- paragraph 22
- Published
- 2024-02-19
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-sam-altman-fried-ee3327-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.