“To be clear, LLMs can absolutely write code, and can absolutely create software, but neither of those mean that the code is good, stable or secure, or that the same can be said of the software they create.”
Our reading, the sentence that was checked:
LLMs can write code and create software, but that doesn't mean the code is good, stable, or secure.
Assessment: Supported Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, AI Is Really Weird, paragraph 84.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-ai-is-really-weird-47e4e7-a1
We recorded 42 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 2026-04-08. 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
- LLMs can absolutely write code
- can absolutely create software
- neither of those mean that the code is good, stable or secure
- the same can be said of the software they create
Our interpretation
- Use 2026 Q1, the latest completed quarter before publication.
- The plain reading is that code-generation capability does not itself guarantee software quality, stability, or security.
- The sentence does not say that all generated code is defective or that it is necessarily worse than human-written code.
- A working generated program with documented material defects is enough to demonstrate the absence of the asserted guarantee; population-wide defect rates are unnecessary.
Counts as supported if: Public artifacts or evaluations show that LLMs produce functioning software while generated outputs can still contain material correctness, stability, or security defects.
Counts as contradicted if: Evidence establishes that generating functioning software itself assures the listed quality properties within the claim's scope.
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
A public generated compiler both performed substantial compilation work and exhibited regressions, crashes, and hangs. This supports the modest claim that generating code and software does not guarantee quality or stability. Successful outputs are consistent with the claim, which expressly acknowledges that capability.
Limitations
- This does not establish how often generated software is defective or compare it with human-written software.
- The decisive examples concern functionality and stability; they do not separately measure security across coding LLMs.
- The external bug reports were inspected as records, not independently reproduced.
Evidence (3 items)
Supports the claim: Anthropic,
not all
Why it matters here: Anthropic's generated compiler demonstrates substantive software-generation capability while retaining functional limitations. That directly fits the distinction in Zitron's sentence.
- Publisher
- Anthropic
- Kind of source
- Demonstration
- Describes events of
- Public from
Supports the claim: Anthropic,
broke existing functionality
Why it matters here: The developer reports regressions in functioning generated software. Generation capability therefore did not guarantee stability in this example.
- Publisher
- Anthropic
- Kind of source
- Demonstration
- Describes events of
- Public from
Supports the claim: bi6c, GitHub issue tracker,
10 unique bugs
Why it matters here: External testers identified compiler panics and hangs and supplied locations and reproduction materials. This independently supports the existence of material defects in the generated program.
- Publisher
- bi6c, GitHub issue tracker
- Kind of source
- Independent measurement
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, AI Is Really Weird
To be clear, LLMs can absolutely write code, and can absolutely create software, but neither of those mean that the code is good, stable or secure, or that the same can be said of the software they create.
- Where
- paragraph 84
- Published
- 2026-04-08
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-ai-is-really-weird-47e4e7-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.