OpenAI knew early on that using more compute made its AI systems smarter and more capable.
Own records only Only the person's or their organisation's own records speak to it, so it is not scored.
Factual: Statements about the present or the past, checked against the record: did they hold up?
Said for an organisation they lead or work for; it counts on both records.
Reasoning
What OpenAI knew internally and when is not independently verifiable beyond its own account.
Cite this claim
Sam Altman, “OpenAI knew early on that using more compute made its AI systems smarter and more capable.” (said 4 September 2026). Verdict: own records only. Reliability API, claim c_db1ac33e2617, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_db1ac33e2617
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Sam Altman
-
Mostly true
“This is hugely complex and expensive and we are still doing this in like a sort of bespoke one-off way, although it’s getting better.”
-
Resolved true
“It’s really hard to find a gigawatt of power available in short term.”
It is really hard to find a gigawatt of available power in the short term.
-
Mostly true
“The very end of 2015, we announced it, and then the first day of work was like January of 2016.”
-
Mostly true
“not until 2020 did we train GPT-3, which was the first model that I think actually was over some threshold of usability.”
OpenAI trained GPT-3 in 2020, which Altman considers the first model above a threshold of usability.
-
Mostly true
“Absolutely. To go back to the analogy of electricity, there’s this amazing chart of what it cost to have an hour of light at night.”
Verdicts are made automatically; how verdicts are made.