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?
79 72 85 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
80 72 88 -
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
Insufficient record: 1 promise resolved, 5 needed for a score.
All types combined: 79 (likely 74 to 84, high confidence), from 102 claims: 64 factual claims, 38 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Gary Marcus's claims
52 claims
“$50B in, $3B out. That’s obviously not sustainable.”
Gary Marcus, author, Marcus on AI 31 March 2024 When Will the GenAI Bubble Burst? (Marcus on AI)
Marcus cited a statistic from a Wall Street Journal piece, "$50B in, $3B out", and said it was obviously not sustainable.
Mostly true True in substance, wrong in a detail.
What would count as true or false
This is a factual claim judged as of 2024-03-31. It has two parts: (a) a Wall Street Journal piece reported roughly '$50B in, $3B out'; (b) that ratio is obviously not sustainable. Part (b) is an opinion and is only judged as a characterization: does the cited ratio really describe spending far above revenue? The test is mainly part (a). The WSJ piece reported (or quoted a source saying) that AI investment or spending was about $50 billion against about $3 billion in revenue. A WSJ article published on or before 2024-03-31 showing this figure, or one close to it ($45-55B in, $2.5-3.5B out), and describing the same in/out comparison. True: the WSJ article shows the figures as quoted and in the context implied (AI industry spending against revenue). Mostly true: the WSJ reported a comparable statistic with a minor difference in the numbers, or attributed it to a source it quoted (for example a VC presentation) rather than its own analysis, without changing the meaning. Mixed: the figures appear, but in a materially different context, such as a different time period or a different subset of companies, so the point holds only in part. Mostly false: the numbers come from somewhere other than the WSJ, or the WSJ figures differ substantially (off by more than about 30%) while still showing a large gap. False: no such statistic appears in the cited WSJ piece, or the WSJ figures show no large gap between spending and revenue.
Evidence
- Researcher has a warning about the future of the AI boom TheStreet · 1 April 2024
“A recent presentation from venture capital firm Sequoia found that the AI industry in 2023 spent $50 billion on Nvidia chips alone, according to the Journal.”
- Researcher has a warning about the future of the AI boom TheStreet · 1 April 2024
“The industry at the same time brought in only $3 billion of revenue.”
Reasoning
E1 is an independent TheStreet article. It says the Wall Street Journal reported on a recent Sequoia presentation that found the AI industry spent $50 billion on Nvidia chips in 2023. E2, the next sentence, says the industry brought in only $3 billion of revenue in the same period. That matches the "$50B in, $3B out" comparison and shows spending far above revenue. The rating is mostly true rather than true for three reasons. The WSJ attributed the figures to a VC presentation rather than its own analysis. The spending figure covers Nvidia chips only. And the evidence describes the WSJ piece secondhand rather than quoting it directly.
Cite this claim
Gary Marcus, “Marcus cited a statistic from a Wall Street Journal piece, "$50B in, $3B out", and said it was obviously not sustainable.” (said 31 March 2024). Verdict: mostly true. Reliability API, claim c_cca6e2e92c88, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_cca6e2e92c88
Claims without a verdict
Is a verdict here wrong, or do you speak for Gary Marcus? 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.