“It’s also weird that people discussing agents don’t seem to want to discuss that OpenAI’s Operator Agent does not work, that AI browsers are fundamentally broken, or that agentic AI does not do anything that people discuss.”
Our reading, the sentence that was checked:
OpenAI's Operator does not work, AI browsers are fundamentally broken, and agentic AI doesn't do what people say.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, AI Is Really Weird, paragraph 236.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-ai-is-really-weird-e829bf-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
- OpenAI’s Operator Agent does not work
- AI browsers are fundamentally broken
- agentic AI does not do anything that people discuss
Our interpretation
- The ordinary reading concerns dependable completion of publicly promoted tasks, rather than an inability to complete even one task.
- Use the latest completed calendar quarter before publication, 2026 Q1, for this ongoing technological claim. Earlier tests and later product improvements cannot establish performance in that period.
- Assess Operator, AI browsers, and the broader agent claim separately. Success by another browser agent does not establish Operator's performance.
- The source supplies neither a task population nor an acceptable failure rate. Do not invent a numerical reliability threshold.
Counts as supported if: Representative, reproducible tests during the target period show persistent failure to complete the ordinary tasks promoted for the named products, with substantial intervention required.
Counts as contradicted if: Representative tests during the target period show dependable completion of those tasks across the named products and relevant agent applications.
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 testing supports a real reliability concern, while the benchmark shows substantial capability in a different browser agent. Neither establishes the full three-part claim. Treating occasional success as sufficient disproof would impose an unusually literal reading; treating frequent intervention in one product as proof about all agents would overgeneralize.
Limitations
- One independent evaluation is small and informal.
- The vendor benchmark is not an Operator test.
- The verdict depends partly on what counts as working reliably.
Evidence (2 items)
Supports the claim: Ars Technica,
needed nudging
Why it matters here: Firsthand testing supports the concern about dependable autonomous operation. The small, informal task selection cannot establish failure across all browsers or Operator.
- Where
- Final results
- Publisher
- Ars Technica
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: Browser Use,
78%
Why it matters here: The retrospective account reports substantial task completion, which cautions against reading the claim as total incapacity. It concerns a different product, uses vendor-selected tasks and an automated judge, and does not establish reliability for Operator.
- Publisher
- Browser Use
- Kind of source
- Benchmark result
- Describes events of
- Public from
What would settle it
Archived Q1 2026 evaluations of Operator and a representative set of AI browsers, with repeated promoted tasks, explicit success criteria, and intervention logs.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, AI Is Really Weird
It’s also weird that people discussing agents don’t seem to want to discuss that OpenAI’s Operator Agent does not work, that AI browsers are fundamentally broken, or that agentic AI does not do anything that people discuss.
- Where
- paragraph 236
- Published
- 2026-04-08
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-ai-is-really-weird-e829bf-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.