“Here's the thing, I get that that might be what they're pushing toward, but generative AI isn't even automation at this point.”
Our reading, the sentence that was checked:
As of mid-2024, generative AI does not actually work as automation.
Assessment: Contradicted Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Rot Economics - An Interview With MIT's Daron Acemoglu, paragraph 28.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-rot-economics-an-interview-with-mits-daron-acemoglu-ab54c5-a1
We recorded 8 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 2024-07-24. 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)
Conditions
- at this point (qualifier, stated by the speaker)
From the speaker
- generative AI
- isn't even automation
- at this point
Our interpretation
- The ordinary reading is that generative AI was not successfully performing real work tasks without a person executing each task.
- Automation can cover a bounded task, including answering customer questions. It need not replace an entire occupation or work without errors.
- For this ongoing observation, use April–June 2024, the latest complete quarter before publication.
Counts as supported if: Representative deployment evidence establishes that purported generative AI automation still required humans to perform the tasks, with no credible working counterexample.
Counts as contradicted if: A documented deployment in the target period successfully performs a real work task using generative AI without human execution of each instance.
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
Under the ordinary task-level meaning of automation, the categorical statement is contradicted. Intercom reported deployed Fin systems resolving customer questions in April 2024, and its technical description identifies GPT-4 as part of the system. Customer feedback exposes problems with counting assumed resolutions, so 46% should not be treated as a verified success rate. Those problems do not erase the evidence that some questions were answered successfully without a human agent.
Limitations
- The main quantitative evidence comes from the vendor.
- The customer account is anecdotal, not a controlled evaluation.
- This does not establish reliable automation of whole jobs or complex workflows.
Evidence (3 items)
Goes against the claim: Intercom,
46%
Why it matters here: Intercom reported actual customer deployments resolving queries directly through Fin. This supports the existence of bounded task automation, although the reported percentage is not an independently verified success rate.
- Publisher
- Intercom
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: Intercom,
OpenAI’s new GPT-4
Why it matters here: Establishes the generative component of the system measured in e1. This launch announcement alone would not establish successful deployment in the target quarter.
- Publisher
- Intercom
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: Intercom Community; customer Henrik Lenerius,
assumed resolved
Why it matters here: A customer's firsthand account raises false-positive resolution concerns, including incorrect answers followed by abandonment. It limits confidence in the reported rate but also describes successful answers; it does not establish that no tasks were automated.
- Publisher
- Intercom Community; customer Henrik Lenerius
- 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, Rot Economics - An Interview With MIT's Daron Acemoglu
Here's the thing, I get that that might be what they're pushing toward, but generative AI isn't even automation at this point.
- Where
- paragraph 28
- Published
- 2024-07-24
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-rot-economics-an-interview-with-mits-daron-acemoglu-ab54c5-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.
About assessment wyea-rot-economics-an-interview-with-mits-daron-acemoglu-ab54c5-a1.