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
42 claims
“Hasn’t worked so far. That’s for sure.”
Gary Marcus, speaking in conversation 2025 Dr. Ben Goertzel's Conversation with Dr. Gary Marcus at the World AI Summit 2025 (SingularityNET)
Bolting extra components onto an LLM and scaling it up has so far not produced human-level AGI.
Resolved true The record shows it happened or was so.
What would count as true or false
Judge as of 2025, the date the claim was made. True: no system built by adding components to LLMs (tools, retrieval, reasoning modules) and scaling them had achieved human-level AGI by then, as reflected in a broad expert and benchmark consensus, with no credible recognition of AGI achieved. Mostly true: no consensus AGI existed, but some credible voices claimed AGI-level performance in a way that left the matter somewhat contested. Mixed: a substantial share of experts held that such systems had reached human-level general intelligence. Mostly false: widely recognized evidence showed such systems had largely reached human-level AGI. False: there was broad consensus that LLM-based systems had achieved human-level AGI.
Evidence
- More-powerful AI is coming. Academia and industry must oversee it — together Nature · 6 December 2024
“There was a consensus among researchers who spoke to Nature for a News Feature published this week (see Nature 636, 22–25; 2024) that large language models (LLMs), such as o1, Google’s Gemini and Claude, made by Anthropic, based in San Francisco, have not yet delivered AGI.”
- More-powerful AI is coming. Academia and industry must oversee it — together Nature · 6 December 2024
“François Chollet, a software engineer formerly at Google, based in Mountain View, California, and Subbarao Kambhampati, a computer scientist at Arizona State University in Tempe, have tested o1’s performance on tasks that require abstract reasoning and planning, and found that it comes up short of AGI.”
- Beyond big models: Why AI needs more than just scale to reach AGI | IBM IBM Think · 17 March 2025
“Most surveyed AI researchers believe that deep learning alone isn’t enough to reach AGI.”
- The shine began to wear off AI in 2024 as advances slowed down | New Scientist New Scientist · 11 December 2024
“Some say AI models that were released this year and can apparently reason more effectively than their predecessors show that the lofty goal of artificial general intelligence (AGI) is still on track, but not everyone is convinced.”
Reasoning
I resolved this claim as true. In December 2024, Nature reported that the researchers it spoke to agreed LLMs, including reasoning models like o1, had not yet delivered AGI [E1]. It also reported that outside researchers tested o1 and found it fell short of AGI [E2]. In March 2025, IBM reported that most surveyed AI researchers believe deep learning alone is not enough to reach AGI [E3]. New Scientist noted that some people saw the goal of AGI as still "on track," which is not a claim that it had been reached, and that others disagreed [E4]. None of the evidence shows credible recognition that AGI had been achieved.
Cite this claim
Gary Marcus, “Bolting extra components onto an LLM and scaling it up has so far not produced human-level AGI.” (said 2025). Verdict: resolved true. Reliability API, claim c_edbabc55660a, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_edbabc55660a
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.