“Already in recent weeks, Inflection AI largely closed shop, Stability AI is struggling, and LLM-based autonomous vehicle company called Ghost closed shop, and a software engineer on YouTube raised quite serious questions about the wildly-hyped AI coding system Devin.”
Our reading, the sentence that was checked:
In recent weeks Inflection AI largely closed, Stability AI is struggling, Ghost closed, and serious questions were raised about Devin.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Evidence that LLMs are reaching a point of diminishing returns - and what that might mean, paragraph 28.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-evidence-that-llms-are-reaching-a-d3b9f2-a1
We recorded 20 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2024-04-13. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- “recent weeks”
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- in recent weeks
- Inflection AI largely closed shop
- Stability AI is struggling
- Ghost closed shop
- a software engineer on YouTube raised quite serious questions
- AI coding system Devin
Our interpretation
- I read 'recent weeks' as March 1 through April 13, 2024. This approximate window comes from the post date, rather than from subsequent company developments.
- I read 'largely closed shop' as the substantial dismantling of Inflection's existing operating organization. The qualifier allows a surviving company or smaller continuing business; it does not require legal dissolution.
- I read 'struggling' as substantial business or organizational distress, without imposing a particular financial threshold.
- I read 'serious questions' as specific, substantive challenges to Devin's advertised demonstration. The claim does not require those challenges to prove that Devin was generally incapable.
- The sentence reports four developments. It does not itself establish that diminishing technical returns caused them.
Counts as supported if: Evidence establishes the substantial dismantling of Inflection's existing organization, significant distress at Stability, Ghost's cessation of operations, and a substantive YouTube critique of Devin circulating within the target period.
Counts as contradicted if: Reliable evidence shows that one or more of these developments did not occur within the target period. Continued substantial original operations at Inflection would contradict the adopted reading; mere corporate survival would not.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches 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
Supported under the stated organizational reading. The [CMA decision](https://assets.publishing.service.gov.uk/media/6719ff5f549f63039436b3c8/__Full_text_decision__.pdf) confirms the transfer of almost all of Inflection's team. Its surviving business makes a literal shutdown reading inaccurate. Reporting supports substantial distress at Stability. Ghost closed. The dated critique record, corroborated by reporting identifying its author and complaints, supports the existence of serious questions about Devin. This does not establish the broader argument that technical progress was diminishing.
Limitations
- The Inflection finding depends on the meaning of 'largely'; the company continued operating.
- Stability's situation is assessed qualitatively, without adopting a numerical loss claim.
- News reporting substitutes for unavailable original material in parts of the assessment.
- The Devin finding concerns the existence of substantive criticism, not a verdict on its technical accuracy.
Evidence (7 items)
Supports the claim: UK Competition and Markets Authority,
the vast majority
Why it matters here: The regulator found that almost all of the team transferred, including its core expertise. This supports substantial dismantling of the original organization, under the stated interpretation.
- Publisher
- UK Competition and Markets Authority
- Kind of source
- Regulatory filing
- Describes events of
- Public from
Supports the claim: Financial Times, republished by Ars Technica,
backers have grown restive
Why it matters here: The report describes investor pressure, leadership disruption and departures, supporting ordinary business distress. It also includes Mostaque's denials and an investor's optimism about replacement management.
- Publisher
- Financial Times, republished by Ars Technica
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: TechCrunch,
has shut down
Why it matters here: Confirms closure.
- Where
- Opening paragraph
- Publisher
- TechCrunch
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: Vuink,
Devin's debugging challenges
Why it matters here: This dated record establishes that the named critique was circulating by April 13. Its allegations are used only to establish that substantive questions were raised, not that those allegations were correct.
- Publisher
- Vuink
- Kind of source
- Secondary summary
- Describes events of
- Public from
Context only: Inflection AI,
a new CEO, Sean White.
Why it matters here: Inflection announced continuing API and business services under new leadership. This is material counterevidence to a literal shutdown reading, although it does not negate the departure of almost all of the original team.
- Publisher
- Inflection AI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: 80 Level,
Internet of Bugs
Why it matters here: This subsequent report identifies the same video and describes complaints about the requested task, generated errors and completion time. Its publication date alone cannot establish pre-post timing; e5 supplies that corroboration.
- Publisher
- 80 Level
- Kind of source
- News report
- Describes events of
- Public from
Context only: Futurism,
Both Answer.AI and Brown
Why it matters here: The surrounding passage identifies Carl Brown as a software engineer from Internet of Bugs and dates his video to April. Later Answer.AI results are not used to judge this historical claim.
- Publisher
- Futurism
- Kind of source
- News report
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Evidence that LLMs are reaching a point of diminishing returns - and what that might mean
Already in recent weeks, Inflection AI largely closed shop, Stability AI is struggling, and LLM-based autonomous vehicle company called Ghost closed shop, and a software engineer on YouTube raised quite serious questions about the wildly-hyped AI coding system Devin.
- Where
- paragraph 28
- Published
- 2024-04-13
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-evidence-that-llms-are-reaching-a-d3b9f2-a1: Current. Outcome: Supported. 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-evidence-that-llms-are-reaching-a-d3b9f2-a1.