“And because of the discrepancy between reliability and the immense costs that stem from the inherent inefficiency of a system that it dependent on absorbing internet-scale data, it has thus far failed to make profits for companies like OpenAI.”
Our reading, the sentence that was checked:
Because of unreliability and immense costs from inefficiency, ChatGPT has failed so far to make profits for companies like OpenAI.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be, paragraph 45.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-three-years-on-chatgpt-still-isnt-d4be08-a2
We recorded 31 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 2025-11-29. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Causal claim
- 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
- because of the discrepancy between reliability and the immense costs
- inherent inefficiency
- dependent on absorbing internet-scale data
- thus far failed to make profits
- companies like OpenAI
Our interpretation
- Assess the stated causal explanation, not merely whether a loss was reported.
- The ordinary financial reading is company-wide net profitability at OpenAI. ChatGPT-specific profitability, operating income and cash flow are separate measures.
- Use January–June 2025, the latest directly reported OpenAI financial reporting window located in evidence available before the post. Microsoft’s later investment results are a different company's metric.
- The claim requires reliability constraints and the stated training-cost mechanism to contribute materially to the lack of profits; they need not be the only causes.
- The source identifies no accounting basis or financial source. Preserve accounting_basis as unknown rather than infer it from evidence found later.
- OpenAI is the named example. No defined set of comparable companies is supplied.
Counts as supported if: Matching financial evidence establishes a lack of company-wide net profit, and credible financial or deployment evidence shows that reliability constraints and costs arising from the stated data-intensive training mechanism materially contributed to that result.
Counts as contradicted if: Matching financial evidence establishes company-wide net profitability, or credible causal evidence rules out a material role for the stated explanation.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Does the evidence support the cause?
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
TIME reported a $7.8 billion operating loss for January–June 2025 [e1]. Reuters relayed $2.5 billion in cash burn and $6.7 billion in research and development costs for that period, while noting it could not verify the report [e2]. These measures do not establish company-wide net income or whether reliability constraints and costs from data-intensive training materially contributed to a lack of profit [e1, e2]. Reported paid-user compute margins improved to 70% in October 2025 from 52% at the end of 2024, but this later serving-cost measure neither establishes first-half net profitability nor settles the proposed causes [e3].
Limitations
- No matching January–June 2025 net-income statement was obtained; the claim's accounting basis remains unknown.
- The reporting does not isolate the financial effects of unreliability or distinguish the stated training-cost mechanism from serving, research, expansion and other costs [e1, e2].
- The October compute-margin figures were not publicly confirmed by OpenAI and concern a different period and metric [e3].
- The packet provides no ChatGPT-specific profit allocation or defined comparison group.
Evidence (3 items)
Context only: TIME,
$7.8 billion
Why it matters here: The reported operating loss is consistent with the financial premise. It is not a matching net-income measure and does not identify unreliability as a cause.
- Publisher
- TIME
- Kind of source
- News report
- Measures
- OpenAI, operating income, USD billions, estimate, 2025-01-01 to 2025-06-30
- Describes events of
- Public from
Context only: Reuters, syndicated by Business Recorder,
could not immediately verify
Why it matters here: Reuters relays The Information's account of $2.5 billion in first-half cash burn and substantial research and development costs. This supplies qualified cost context, but cash burn is not net loss and the report does not isolate costs caused by unreliability.
- Publisher
- Reuters, syndicated by Business Recorder
- Kind of source
- News report
- Measures
- OpenAI, cash burn, USD billions, estimate, 2025-01-01 to 2025-06-30
- Describes events of
- Public from
Context only: Proactive Investors,
does not publish those numbers
Why it matters here: The article relays an October paid-user compute margin of 70%, improved from 52% at the end of 2024, with no public company confirmation. Improving serving economics qualify a blanket inefficiency explanation, but this later-period contribution measure does not establish company-wide profit.
- Publisher
- Proactive Investors
- Kind of source
- News report
- Measures
- OpenAI, paid-user compute margin percent, estimate, 2025-10-01 to 2025-10-31
- Describes events of
- Public from
What would settle it
OpenAI's January–June 2025 company-wide net-income records with an identified accounting basis, plus cost and deployment evidence establishing or ruling out a material contribution from reliability constraints and the stated data-intensive training mechanism. Matching evidence of net profitability would also settle the claim against its financial premise.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be
And because of the discrepancy between reliability and the immense costs that stem from the inherent inefficiency of a system that it dependent on absorbing internet-scale data, it has thus far failed to make profits for companies like OpenAI.
- Where
- paragraph 45
- Published
- 2025-11-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-three-years-on-chatgpt-still-isnt-d4be08-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.
About assessment marcus-three-years-on-chatgpt-still-isnt-d4be08-a2.