What Gary Marcus and Ed Zitron said about AI, and what happened.
So, were they right?
Judged only on the claims the evidence could settle, as of 2026-10-29. Two AI reviewers from different companies checked each one on its own. How the answer is worked out.
Gary Marcus
Mostly right.
Of the 600 claims the evidence could settle:
- 88%supported530 claims
- 5%mixed30 claims
- 7%contradicted40 claims
Not counted: 803 unresolved, 2,353 that can’t be checked because they are opinion, rhetoric, or too open to test, and 1,684 held until they pass every check.
Most supported claims were about reliability (206) and capabilities (198).
Most contradicted claims were about capabilities (13) and regulation policy (12).
The share supported held steady from year to year, between 85% and 92%.
Ed Zitron
Too early to say.
Only 4 of Ed Zitron’s claims have been settled by the evidence so far: 1 supported, 2 mixed, and 1 contradicted.
Not counted: 25 unresolved, 1,290 that can’t be checked because they are opinion, rhetoric, or too open to test, and 589 held until they pass every check.
Look up a claim
Every claim checked against evidence All of Gary Marcus’s claims All of Ed Zitron’s claims
Predictions that came due
Calls that named a deadline and that the evidence has now settled, newest first.
Supported
“In my Politico “Black Swan” prediction for 2026 year, from January I made the seemingly wild prediction that,“By the end of 2026, President Trump will have begun to distance himself from the aggressively pro-AI industry policies that characterized his AI strategy in 2025””
Supported
“Examples of concealed bias will be discovered within days or months.”
Supported
“Bullshit. Hallucinations are still here, and aren’t going away, anytime soon.”
Supported
“To be sure, most or all trained coders will (if they are not already) soon use AI as a part of their workflow.”
Contradicted
“All that we will have this year are demos.”
Supported
“They will be a major focus of attention in the field in 2025.”
Supported
“AI “Agents” will be endlessly hyped throughout 2025 but far from reliable, except possibly in very narrow use cases.”
Supported
“Copyright lawsuits over generative AI will continue throughout the year.”