“It requires endless pre-training, post-training and script-based MacGuffins to do tasks with mathematically-guaranteed hallucinations that burn more tokens, raising costs on customers who are already in mutiny less than a quarter into being forced to pay a cost that is already unprofitable.”
Our reading, the sentence that was checked:
Generative AI requires endless pre-training, post-training and script-based workarounds to do tasks with mathematically guaranteed hallucinations that burn more tokens, raising costs on customers who are already in mutiny less than a quarter into paying an already-unprofitable price.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, AI's Brokenomics, paragraph 87.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-brokenomics-5636f9-a1
We recorded 41 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 2026-06-15. 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)
Conditions
- less than a quarter into being forced to pay (scope, stated by the speaker)
From the speaker
- requires endless pre-training, post-training and script-based MacGuffins
- mathematically-guaranteed hallucinations
- burn more tokens
- raising costs on customers
- already in mutiny less than a quarter
- a cost that is already unprofitable
Our interpretation
- The ordinary reading is a causal chain connecting continuing model development and error handling to additional tokens, higher customer bills, customer resistance, and loss-making service.
- I read endless as an ongoing requirement, rather than literally infinite computation.
- Mathematically guaranteed hallucinations means errors cannot be eliminated across the relevant tasks. It does not mean every answer must be wrong.
- The preceding discussion suggests modeled gross margins on OpenAI and Anthropic customer usage. It does not identify an observed company-wide net margin.
- The three months preceding publication bound the claimed recent pricing experience. They do not establish when any particular customer began paying.
- Mutiny means documented customer resistance, cancellations, spending restrictions, or migration prompted by charges.
Counts as supported if: Technical and operational evidence establishes the claimed necessities, and matched customer and provider records connect them to extra tokens, higher bills, resistance within three months of the relevant pricing change, and negative service gross margins.
Counts as contradicted if: Evidence directly disproves a necessary link in that chain for the stated scope. A cheaper isolated task, a positive margin on another product, or an error-free individual answer would not suffice.
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
The sources establish narrower facts: formal limits on general-purpose accuracy, incentives to guess, a usage-billing change, and reported customer spending controls. They do not connect these facts into the claimed causal chain. In particular, a hallucination theorem does not establish extra billing or loss-making service. The evidence therefore neither supports nor contradicts the full statement.
Limitations
- The claim combines technical necessity, customer behavior, and financial results without identifying a common measurement scope.
- The customer spending report is secondary and the underlying memo was not independently available.
- Modeled subscription margins cannot be treated as observed company profitability.
Evidence (4 items)
Context only: arXiv,
impossible to eliminate hallucination
Why it matters here: The formal result concerns computable models used as general problem solvers across computable ground-truth functions. It does not establish additional billed tokens, continual retraining requirements, or negative margins.
- Where
- Abstract
- Publisher
- arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: Nature,
reward guessing over admitting uncertainty
Why it matters here: The paper distinguishes statistical error pressure from incentives to answer rather than abstain. This limits a blanket interpretation of inevitability, but does not test the claimed billing and margin consequences.
- Where
- Abstract
- Publisher
- Nature
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
Context only: GitHub,
GitHub AI Credits consumed
Why it matters here: This establishes a relevant recent billing change. It concerns GitHub and does not establish the claimed OpenAI or Anthropic margins, customer backlash, or a causal connection to hallucinations.
- Publisher
- GitHub
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: The Decoder,
budgets, allocations, and dedicated tools
Why it matters here: Reporting based on The Information describes planned customer spending controls. It does not show that hallucinations caused the spending, identify the relevant first payment date, or establish provider losses.
- Publisher
- The Decoder
- Kind of source
- Secondary summary
- Describes events of
- Public from
What would settle it
Matched task, token, billing, customer-response, and provider-margin records, together with technical evidence establishing or disproving the asserted necessities.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, AI's Brokenomics
It requires endless pre-training, post-training and script-based MacGuffins to do tasks with mathematically-guaranteed hallucinations that burn more tokens, raising costs on customers who are already in mutiny less than a quarter into being forced to pay a cost that is already unprofitable.
- Where
- paragraph 87
- Published
- 2026-06-15
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-brokenomics-5636f9-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.