“It isn’t getting much more efficient.”
Our reading, the sentence that was checked:
Generative AI is not getting much more efficient.
Assessment: Contradicted Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Pop Culture, paragraph 60.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-pop-culture-da2da3-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
- It
- isn't getting
- much more efficient
Our interpretation
- It refers to generative AI.
- The ordinary technical reading is resources needed to produce comparable-quality output, including computation and memory.
- I use April–June 2024, the latest completed quarter before the post, because the statement describes an ongoing trend without specifying a period.
- API price cuts alone do not prove lower resource use, and aggregate industry expenditure is a different measure.
- This is a technology-efficiency claim, not a claim about a particular company's financial accounts.
Counts as supported if: Comparable-quality systems show little material reduction in resource requirements during the target period.
Counts as contradicted if: Documented comparisons available for the target period show large reductions in resources per comparable-quality output, including materially greater output from the same hardware.
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
DeepSeek's May 2024 technical report documents large memory and hardware-throughput improvements alongside stronger measured model performance. Under the ordinary resource-efficiency reading, these are substantial improvements already reported before the post. They do not establish falling aggregate industry spending or energy consumption, which are different claims.
Limitations
- The decisive measurements are developer-reported, without an independent replication established here.
- One model-family comparison does not describe every generative AI system.
- Attention-cache savings are not equivalent to reductions in total model memory or electricity use.
- The result depends on reading efficiency as resource use per output rather than overall industry economics.
Evidence (1 item)
Goes against the claim: DeepSeek-AI / arXiv,
smaller amount
Why it matters here: The paper reports stronger model performance with 93.3% less attention-cache memory than DeepSeek 67B. Its deployment tests report 5.76 times the generation throughput on an eight-H800 node using the earlier service's workload distribution, demonstrating substantial resource-efficiency gains.
- Publisher
- DeepSeek-AI / arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Pop Culture
It isn’t getting much more efficient.
- Where
- paragraph 60
- Published
- 2024-07-08
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-pop-culture-da2da3-a1: 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.