ClaimsDario Amodei

“I thought it was a very plausible blocker one or two years ago.”

Dario Amodei, Anthropic CEO 10 March 2025 CEO Speaker Series with Dario Amodei of Anthropic (Council on Foreign Relations) archived copy

Factual AI compute, hardware and data

Dario Amodei thought one or two years before March 2025 that data limits were a very plausible blocker to AI progress.

Mostly false False in substance, right in a detail.

Factual: Statements about the present or the past, checked against the record: did they hold up?

What would count as true or false

The claim is about Amodei's own past view; judge it against his contemporaneous public statements from roughly March 2023 to March 2024. TRUE: statements from that period show him treating limits on data as a real, plausible obstacle to continued AI scaling. MOSTLY TRUE: he raised data limits as a possible blocker then, but called it less likely than 'very plausible' (for example a modest-probability risk). MIXED: his statements from the period are inconsistent, sometimes raising data limits and sometimes dismissing them. MOSTLY FALSE: he mentioned data limits only in passing and mainly said data would not be a problem. FALSE: his statements from the period clearly dismiss data limits as a meaningful blocker. If he made no relevant statements in that period, the claim cannot be resolved.

Set before the research began.

Evidence (4)

  1. E1 Dario Amodei (Anthropic CEO) - Scaling, Alignment, & AI Progress Dwarkesh Patel (dwarkesh.com podcast) · 8 August 2023

    “One practical issue we could have is we could run out of data. For various reasons, I think that's not going to happen but if you look at it very naively we're not that far from running out of data.”

  2. E2 Dario Amodei (Anthropic CEO) - Scaling, Alignment, & AI Progress Dwarkesh Patel (dwarkesh.com podcast) · 8 August 2023

    “My guess is that this will not be a blocker.”

  3. E3 Dario Amodei (Anthropic CEO) - Scaling, Alignment, & AI Progress Dwarkesh Patel (dwarkesh.com podcast) · 8 August 2023

    “I wouldn't bet on either of those things happening but they could.”

  4. E4 Anthropic's Dario Amodei on AI's limits: 'I'm not sure there are any' | TechCrunch TechCrunch · 21 September 2023

    “At the very least, Amodei suggested, we won’t see diminishing returns for the next three or four years”

Reasoning

In his August 2023 interview, Amodei called running out of data a practical issue that could arise, and said that "very naively" the field was not far from it. In the same interview he added "I think that's not going to happen" and "My guess is that this will not be a blocker" (E1, E2). E3 shows he saw such outcomes as possible but would not bet on them, though the quote does not make clear what "those things" refers to. TechCrunch in September 2023 reported that he expected no diminishing returns for the next three or four years (E4). Taken together, he acknowledged data limits but mainly said they would not block progress. That does not support his later description of them as a "very plausible blocker" at the time.

Cite this claim

Dario Amodei, “Dario Amodei thought one or two years before March 2025 that data limits were a very plausible blocker to AI progress.” (said 10 March 2025). Verdict: mostly false. Reliability API, claim c_9e8de88f8c90, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_9e8de88f8c90

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Dario Amodei, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More factual claims by Dario Amodei

  1. Mostly true Factual 28 February 2026
    “the amount of computation that goes into the models doubles every four months.”

    The amount of computation going into AI models doubles every four months.

  2. Mostly true Factual 5 June 2025
    “Distillation of frontier models allows lagging companies to narrow the gap using a fraction of the cost it would take to develop their own AI independently.”

    Distillation of frontier models lets lagging companies narrow the gap at a fraction of the cost of developing their own AI independently.

  3. Mostly true Factual 13 October 2022
    “Importantly, because this type of RL is new, we are still very early on the scaling curve”

    Because reasoning-focused RL is new, the amount all AI developers spend on the RL training stage is still small, so the field is early on its scaling curve.

  4. Resolved true Factual 13 October 2022
    “I'd guess the number today is maybe ~4x/year.”

    The current rate of algorithmic curve shift is maybe about 4x per year.

  5. Mixed Factual 23 September 2026
    “It’s easy to dismiss this as a one-off or curiosity, but we’ve repeatedly seen a pattern where AI performance in new intellectual domains goes from weak to superhuman in a matter of a few years.”

    AI performance in new intellectual domains has repeatedly gone from weak to superhuman within a few years.

All of them

Verdicts are made automatically; how verdicts are made.