“No private startup can actually afford to build and scale a Large Language Model without the help of Google, Amazon or Microsoft — and thus they would, assuming they actually became useful, effectively control every single product involving this technology.”
Our reading, the sentence that was checked:
No private startup can afford to build and scale an LLM without the help of Google, Amazon or Microsoft.
Assessment: Contradicted Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Monopoly Money, paragraph 34.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-monopoly-money-a4a448-a2
We recorded 11 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 2024-08-12. 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)
Conditions
- without help of Google, Amazon or Microsoft (qualifier, stated by the speaker)
From the speaker
- No private startup
- afford to build and scale a Large Language Model
- without the help of Google, Amazon or Microsoft
Our interpretation
- The ordinary reading is a universal claim of necessary financial or computing dependence on at least one of the three named companies.
- A privately financed startup qualifies even if its founder or parent business is wealthy. The source does not exclude such startups.
- Building and scaling means training a substantial competitive model and deploying inference, not merely announcing a model or fine-tuning someone else's small model.
- Help means material funding or computing infrastructure necessary for development and deployment. Incidental use of ordinary software or public research is not effective dependence.
- For the ongoing state, use the latest completed calendar quarter before publication: April through June 2024.
- No particular company's financial amount or accounting measure is asserted, so metric_scope is null.
Counts as supported if: Comprehensive evidence shows that private startups developing and deploying substantial LLMs necessarily depended on funding or compute from Google, Amazon, or Microsoft.
Counts as contradicted if: At least one private startup trained and deployed a substantial competitive LLM using independently identified financing and infrastructure outside the three named companies.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Contradicted
The dated evidence goes against what the claim committed to, under the criteria frozen before grading.
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 May 7, 2024 DeepSeek-V2 paper documents substantial model training and deployed inference on H800 hardware within the April–June 2024 target period [e1]. November 2024 reporting identifies High-Flyer funding and computing resources, while testimony published April 8, 2025 traces DeepSeek's resources to its High-Flyer origins in 2023 [e2, e3]. Together, these sources establish an independent financing and computing route for a private startup's substantial LLM, contradicting the universal claim of necessary dependence on Google, Amazon, or Microsoft [e1, e2, e3].
Limitations
- The financing and infrastructure finding combines a contemporary technical paper with later reporting and retrospective testimony.
- No audited target-quarter financing record or complete supplier inventory is provided.
- The evidence identifies an independent resource route; it does not certify the absence of every incidental relationship with the three companies.
- The evidence does not establish that building a competitive model was cheap or easy.
Evidence (3 items)
Goes against the claim: DeepSeek-AI, arXiv,
NVIDIA H800 GPUs
Why it matters here: The paper documents DeepSeek-V2's 236-billion-parameter model, substantial pretraining, and deployed inference on H800 hardware. Together with the independently identified High-Flyer resources in e2 and e3, this supplies a counterexample already operating before the post.
- Publisher
- DeepSeek-AI, arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: ChinaTalk,
fully funded by High-Flyer
Why it matters here: This identifies a financing source outside the named three firms and describes access to High-Flyer's compute. It is later contextual corroboration, not a target-period financial disclosure; no audited primary financing record was located.
- Publisher
- ChinaTalk
- Kind of source
- News report
- Describes events of
- Public from
Context only: Gregory C. Allen, CSIS; testimony hosted by the US House of Representatives,
DeepSeek
Why it matters here: The surrounding historical account traces DeepSeek's infrastructure and research funding to High-Flyer from its 2023 founding. This supports continuity of an independently financed route into the May 2024 model release.
- Publisher
- Gregory C. Allen, CSIS; testimony hosted by the US House of Representatives
- Kind of source
- Secondary summary
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Monopoly Money
No private startup can actually afford to build and scale a Large Language Model without the help of Google, Amazon or Microsoft — and thus they would, assuming they actually became useful, effectively control every single product involving this technology.
- Where
- paragraph 34
- Published
- 2024-08-12
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-monopoly-money-a4a448-a2: Current. Outcome: Contradicted. 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.