Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
80 74 86 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
Insufficient record: 4 predictions resolved, 5 needed for a score.
-
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
85 74 94 -
Guidance
A company's own figures for its results over a period, checked against what it then reported.
Insufficient record: 0 guidance claims resolved, 5 needed for a score.
All types combined: 81 (likely 76 to 86, high confidence), from 86 claims: 70 factual claims, 16 promises. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Mark Zuckerberg's claims
42 claims
“Watermelon is basically, that’s shipping soon.”
Mark Zuckerberg, Meta CEO, interview 15 September 2026 Transcript: Mark Zuckerberg Interview on Meta's Muse AI Agent (Sources Podcast) (The Singju Post)
Meta will ship its Watermelon model soon.
Delivered on time The record shows it happened or was so.
Said for an organisation they lead or work for; it counts on both records.
What would count as true or false
'Soon' is treated as within about three months of 15 September 2026, that is, by 15 December 2026. The test is whether Meta publicly shipped the model it calls 'Watermelon'. Shipping means a release to users or developers through Meta AI apps, an API or open weights. A release under a different public name counts if credible reporting or Meta itself identifies it as Watermelon. True: Watermelon was publicly released by 15 December 2026. Mostly true: it was released on time but only in a limited public form (for example a limited preview or restricted regions). Mostly false: it was released only after 15 December 2026, or only internally or to a closed partner group. False: dated reports show it was delayed beyond the horizon, cancelled or folded into another project without release.
Evidence
- Meta Muse Wonder: benchmarks, pricing, Wonder Full sakutto.ai · 14 October 2026
“Meta released Wonder and the larger Wonder Full together on October 12, 2026.”
- Meta Muse Wonder: benchmarks, pricing, Wonder Full sakutto.ai · 14 October 2026
“Meta says it started rebuilding its AI work from the ground up 14 months ago. According to CNBC, the project's codename was Watermelon.”
- Meta Muse Wonder: benchmarks, pricing, Wonder Full sakutto.ai · 14 October 2026
“Developers can call Muse Wonder through the Meta Model API today under the model name muse-wonder-1.0, but Muse Wonder Full is not offered there.”
- Meta's Watermelon released by October 31, 2026? agenticlearning.ai · 4 October 2026
“No source shows Watermelon released or publicly accessible as of Oct 4”
Reasoning
An independent article dated 14 October 2026 reports that Meta released Muse Wonder and Muse Wonder Full on 12 October 2026 (E1). Citing CNBC, it says the project's codename was Watermelon (E2). It also says developers could call Muse Wonder through the Meta Model API as muse-wonder-1.0 (E3), and API release counts as shipping under the criteria. An earlier scan found no release as of 4 October (E4), which fits an October 12 release well before the 15 December horizon.
Cite this claim
Mark Zuckerberg, “Meta will ship its Watermelon model soon.” (said 15 September 2026). Verdict: delivered on time. Reliability API, claim c_43afaab1cad3, method 0.1.0+3f84d14d. https://epistacklabs.com/claims/c_43afaab1cad3
Claims without a verdict
- Unresolvable 144
- Not yet due 17
- Own records only 41
- Not researched 223
- Not checkable 1,060
Is a verdict here wrong, or do you speak for Mark Zuckerberg? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.