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?
67 61 73 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
60 44 76 -
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
Insufficient record: 0 promises resolved, 5 needed for a score.
All types combined: 66 (likely 61 to 72, high confidence), from 105 claims: 94 factual claims, 11 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Marc Andreessen's claims
20 claims
“this is not, strictly speaking, human intelligence.”
Marc Andreessen, venture capitalist, interviewee 12 June 2023 Marc Andreessen on Why AI Will Save the World (EconTalk)
The AI being discussed is not, strictly speaking, human intelligence.
Resolved true The record shows it happened or was so.
What would count as true or false
Judged as of 2023-06-12. True: the AI systems being discussed (e.g. large language models) are machine systems, not human intelligence, which is the consensus of computer science and cognitive science. Mostly true: this holds, with a minor quibble (e.g. the systems are trained on human-generated data). Mixed or mostly false would apply only if credible expert consensus held these systems to be human intelligence in some substantive sense. False: the evidence shows the AI discussed is in fact human intelligence (e.g. humans secretly doing the work).
Evidence
- How Can the Current State of AI Guide Future Conversations of General Intelligence? PMC (National Center for Biotechnology Information) · 20 March 2024
“Current LLMs are trained by next-token prediction with giant models on very large amounts of data.”
- What is Artificial General Intelligence (AGI)? | IBM IBM · 17 September 2024
“Artificial general intelligence (AGI) is a hypothetical stage in the development of machine learning (ML) in which an artificial intelligence (AI) system can match or exceed the cognitive abilities of human beings across any task.”
- How Can the Current State of AI Guide Future Conversations of General Intelligence? PMC (National Center for Biotechnology Information) · 20 March 2024
“Having vastly more storage, computing power, and speed is often the enabler, allowing for methodologies that humans may not consider true intelligence.”
- Behind ChatGPT’s Success: Investigating Worker Exploitation: My Interview in ‘La Via Libera’ (June 30, 2023) | Antonio A. Casilli Antonio A. Casilli (personal academic blog) · 30 June 2023
“Even ChatGPT, OpenAI’s celebrated chatbot, owes its capabilities to human intervention.”
Reasoning
Independent technical descriptions say current LLMs are large models trained by next-token prediction on very large datasets (E1). They also say these systems rely on storage and compute advantages, using methods that differ from human intelligence (E3). That makes them machine systems, not human intelligence. E4 says ChatGPT's capabilities depend on human intervention, but it does not show that the AI itself is human intelligence. So E4 is at most the kind of minor quibble the criteria anticipate, and it does not change the verdict.
Cite this claim
Marc Andreessen, “The AI being discussed is not, strictly speaking, human intelligence.” (said 12 June 2023). Verdict: resolved true. Reliability API, claim c_f8090c1884ab, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_f8090c1884ab
Claims without a verdict
Is a verdict here wrong, or do you speak for Marc Andreessen? 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.