“Remember how in October and in March I told you that vibe coding — in the sense of amateurs using large language models to write code to “build products that would have previously required teams of engineers” — would never be remotely reliable?”
Our reading, the sentence that was checked:
Vibe coding by amateurs using LLMs to build products that previously required engineering teams will never be remotely reliable.
Assessment: Unresolved No deadline stated on 2026-10-29 Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Is vibe coding dying?, paragraph 1.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-is-vibe-coding-dying-b1629b-a2
We recorded 2 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 2025-10-22. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Forecast
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No deadline stated
- Deadline state on 2026-10-29
- No deadline stated (the release’s reference date)
Conditions
- amateurs using LLMs to write code for products that previously required engineering teams (qualifier, stated by the speaker)
From the speaker
- vibe coding
- amateurs using large language models to write code
- build products that would have previously required teams of engineers
- would never be remotely reliable
Our interpretation
- Read as a forecast with no deadline: amateurs relying on LLMs will not reach even rough reliability when building and running products that once needed engineering teams.
- 'Never' cannot be confirmed at any date. It can only be refuted, by clear evidence that amateurs reliably build and run team-scale products this way.
- Professional developers using AI assistants fall outside the claim.
Counts as supported if: Cannot be fully confirmed because there is no deadline. Continued reports of failure, and the absence of reliable amateur-built team-scale products, are consistent with it but not decisive.
Counts as contradicted if: Credible documented evidence, such as independent reliability studies or several well-documented products, shows that amateurs without engineering teams routinely build, run and maintain team-scale production products reliably with LLM tools.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
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 deadline stated
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
This forecast has no deadline. It says vibe coding by amateurs will never be remotely reliable. Under the contract, it cannot be confirmed. It can only be contradicted early. Some evidence fits the claim but does not settle it. In a TechCrunch report dated 2025-11-10, Lovable's CEO said users building sensitive apps, such as banking apps, must still hire security experts (e1). A SaaStr article from August 2025 discusses a Replit database deletion incident. It concludes that vibe coding by non-technical users may remain limited to specific use cases (e3). On the other side, the same CEO said a duo earns $700,000 a year from a startup they launched on Lovable (e2). The report does not say whether they are amateurs, how complex the product is, or how reliable it is. The packet has no independent study and no well-documented cases of amateurs reliably building and running team-scale products. So the contradiction condition is not met.
Limitations
- Most of the evidence is anecdotal or comes from a vendor's CEO.
- The packet has no independent study of how reliable amateur-built products are in production.
- 'Remotely reliable' has no threshold, and 'amateurs' and 'team-scale products' are not defined.
- The SaaStr item is dated only to August 2025.
Evidence (3 items)
Supports the claim: TechCrunch,
the platform still requires users building sensitive applications — banking apps, for instance — to hire security experts, just as they would with traditional development.
Why it matters here: A leading vibe-coding vendor says amateurs still need experts for sensitive apps. This fits the claim but does not settle a 'never' forecast.
- Where
- paragraph 9
- Publisher
- TechCrunch
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: SaaStr (Jason Lemkin),
true “vibe coding”—where non-technical users can independently build and deploy production applications—may remain limited to specific use cases
Why it matters here: A heavy vibe-coding user concluded in 2025 that non-technical production use remains limited. This is consistent with the claim but is not a final test.
- Publisher
- SaaStr (Jason Lemkin)
- Kind of source
- News report
- Describes events of
- Public from
Goes against the claim: TechCrunch,
a Swedish duo is making $700,000 annually from a startup they launched seven months ago on the platform.
Why it matters here: The CEO describes a revenue-earning product built on the platform. Its reliability and complexity are not documented, so this does not refute the claim.
- Where
- paragraph 7
- Publisher
- TechCrunch
- Kind of source
- News report
- Describes events of
- Public from
What would settle it
An independent study, or several well-documented cases, showing that non-programmers reliably build, run and maintain team-scale production products with LLM tools would contradict the claim. No evidence can confirm 'never'.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Is vibe coding dying?
Remember how in October and in March I told you that vibe coding — in the sense of amateurs using large language models to write code to “build products that would have previously required teams of engineers” — would never be remotely reliable?
- Where
- paragraph 1
- Published
- 2025-10-22
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-is-vibe-coding-dying-b1629b-a2: 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.