“Or, put another way, each additional step becomes increasingly (and exponentially) more expensive to take.”
Our reading, the sentence that was checked:
Each additional step of AI model improvement becomes exponentially more expensive.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Pop Culture, paragraph 14.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-pop-culture-cf098f-a1
We recorded 27 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 2024-07-08. 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
- each additional step
- increasingly (and exponentially) more expensive
- linear improvement in model performance
- obtaining the data
- the compute required to process it
Our interpretation
- Read a step as an equal improvement on a fixed performance measure, and expensive as the financial cost of obtaining data and training the model.
- Use April–June 2024, the latest completed quarter before publication, for the ongoing-state claim. The surrounding argument concerns research and development conditions then.
- Rising spending across successive model generations alone does not establish exponentially rising cost per equal performance gain.
- This is an industry-level scaling claim, not a financial amount for a named company; metric_scope is null.
Counts as supported if: Comparable results show that equal performance gains require exponentially increasing total development expenditure, with the performance measure and cost basis held consistent.
Counts as contradicted if: Comparable results in the claimed setting show equal performance gains without the asserted exponential cost increase.
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
The linked research supports exponentially greater data needs for a particular zero-shot performance relationship. It does not establish exponentially increasing financial cost for each model improvement. Other research shows that methodological changes can improve performance and computational efficiency together, but the available comparison does not directly test this claim in its target period.
Limitations
- Data volume, compute use and financial cost are different measures.
- Results across concepts cannot automatically be generalized to successive model generations.
- The finding is unresolved, not evidence that improvement is inexpensive.
Evidence (2 items)
Context only: arXiv — Vishaal Udandarao and colleagues,
require exponentially more data
Why it matters here: This supports the narrower data-scaling premise for tested multimodal models and concepts. It does not measure financial expenditure or establish the claimed relationship for every successive model improvement.
- Publisher
- arXiv — Vishaal Udandarao and colleagues
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: arXiv — Jordan Hoffmann and colleagues,
substantially less compute
Why it matters here: The study reports improved model performance alongside lower downstream compute requirements, cautioning against a universal cost rule. These are different cost categories and an earlier period, so this is not decisive contradiction.
- Where
- Abstract
- Publisher
- arXiv — Jordan Hoffmann and colleagues
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
Comparable experiments measuring total development costs against equal gains on a specified performance metric, with enough observations to distinguish exponential growth from other relationships.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Pop Culture
Or, put another way, each additional step becomes increasingly (and exponentially) more expensive to take.
- Where
- paragraph 14
- Published
- 2024-07-08
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-pop-culture-cf098f-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.