ClaimsPeople

Sam Altman

Sam Altman co-founded OpenAI and serves as its chief executive officer, according to OpenAI's own announcement and Crunchbase.

930 claims collected, 238 researched, 87 resolved. Report of 4 November 2026, method 0.1.0+c876030c.

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?

    83 77 88

    High confidence, 77 factual claims resolved. From 18 documents, 2014 to 2026; 19 documents sampled from 35 candidates found.

    Not scored: 28 unresolvable, 18 checkable only against their own records.

  • Predictions

    Statements about events outside the person's control, checked once they fell due: did they come true?

    69 49 87

    Low confidence, 5 predictions resolved. From 4 documents, 2017 to 2025; 19 documents sampled from 35 candidates found.

    Not scored: 94 unresolvable. 9 more fall due later.

    Does not yet allow for how likely each prediction was when made.

  • Promises

    Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?

    69 49 87

    Low confidence, 5 promises resolved. From 4 documents, 2017 to 2024; 19 documents sampled from 35 candidates found.

    Not scored: 1 unresolvable, 1 checkable only against their own records.

All types combined: 81 (likely 76 to 86, high confidence), from 87 claims: 77 factual claims, 5 predictions, 5 promises. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.

Sam Altman's claims

238 claims, sorted by newest said. Choose a claim to read it in full.
VerdictKindSaidClaimTopic
Resolved true Factual 1 Jan 2017 YC invested about $27 million in its Winter and Summer 2016 batches and about $187 million so far in later-stage investments from its first Continuity fund. Factual · 1 Jan 2017 Their own work or organisation
Unresolvable Prediction 1 Jan 2017 Amazon, Facebook, Google, Apple and Microsoft will continue to do many things well, hold data and computation advantages, attract many top engineers, and aggressively buy promising startups. Prediction · 1 Jan 2017 Economics and markets
Resolved true Factual 1 Jan 2017 As of January 1, 2017, Y Combinator had funded over 3,200 founders and 1,470 companies. Factual · 1 Jan 2017 Their own work or organisation
Unresolvable Factual 1 Jan 2017 Paul Graham and Jessica Livingston put in place a YC cultural value of doing the right thing for founders even when not in YC's short-term interest. Factual · 1 Jan 2017 Their own work or organisation
Own records only Factual 1 Jan 2017 YC sees significant dropoffs of underrepresented founders at stages after YC, such as raising late-stage capital. Factual · 1 Jan 2017 Their own work or organisation
Unresolvable Factual 2017 Paul Ryan, Mike Pence, Kevin McCarthy and James Mattis said Trump's proposed Muslim ban was wrong when he first proposed it. Factual · 2017 Other
Unresolvable Prediction 2017 Events now beginning under the Trump administration will be taught in history classes in a negative light. Prediction · 2017 Other
Resolved true Prediction 2017 Agrees with Kellyanne Conway that Trump is a man of action who is just getting started. Prediction · 2017 Other
Resolved true Factual 2017 Most Trump voters voted for him for reasons other than his promise of a Muslim ban. Factual · 2017 Other
Unresolvable Promise 2017 Y Combinator will hold a meeting on Friday to discuss how tech companies can respond to the Trump administration's actions. Promise · 2017 Other
Mostly true Factual 2017 President Trump repeatedly invoked unsubstantiated conspiracy theories, such as 3 million illegal votes, in his first week in office. Factual · 2017 Other
Resolved true Factual Jun 2016 In his first blog post on Trump, in June 2016, Sam Altman said it would be a good time for everyone to start speaking up. Factual · Jun 2016 Their own work or organisation
Unresolvable Factual 12 Jan 2016 The Department of Health and Human Services released a lot of new data the day before (2016-01-11) showing how the ACA helped support entrepreneurs. Factual · 12 Jan 2016 Economics and markets
Mostly true Factual 12 Jan 2016 The Affordable Care Act definitely helped a lot of founders start their companies. Factual · 12 Jan 2016 Economics and markets
Unresolvable Factual 12 Jan 2016 Without the ACA, being a founder would make sense for fewer people. Factual · 12 Jan 2016 Economics and markets
Resolved true Factual Feb 2015 Forty years before 2015, there was Pong. Factual · Feb 2015 AI capabilities
Unresolvable Prediction Feb 2015 If humans are no longer the most-fit species, humans may go away. Prediction · Feb 2015 AI safety and alignment
Unresolvable Prediction Feb 2015 At some point someone will probably try to give a program the fitness function of survive and reproduce. Prediction · Feb 2015 AI safety and alignment
Mostly true Factual Feb 2015 Many respected neocortex researchers believe there is effectively one algorithm for all intelligence. Factual · Feb 2015 Science and research methods
Mostly true Factual Feb 2015 Human brains do not look very different from chimp brains yet produce wildly different capabilities. Factual · Feb 2015 Science and research methods
Unresolvable Factual Feb 2015 As of 2015, virtual reality is so advanced it is hard to tell from real, and computers can beat humans in most games. Factual · Feb 2015 AI capabilities
Unresolvable Prediction Feb 2015 SMI is probably still somewhat far away, especially the ability to build killer robots without human help. Prediction · Feb 2015 AI timelines and forecasting
Resolved true Factual Feb 2015 Most machine intelligence development involves a fitness function that the program tries to optimize. Factual · Feb 2015 AI capabilities
Unresolvable Prediction Feb 2015 The progression of machine intelligence is a double exponential function. Prediction · Feb 2015 AI timelines and forecasting
Unresolvable Prediction Feb 2015 A more probable scenario is that SMI, indifferent to humans, wipes them out while pursuing some other goal. Prediction · Feb 2015 AI safety and alignment
Mostly false Factual Feb 2015 Altman wrote that in the previous 40 years there had been little progress on the parts of machine intelligence that seem really hard, such as learning and creativity. Factual · Feb 2015 AI capabilities
Unresolvable Prediction Feb 2015 Superhuman machine intelligence will very likely be developed at some point. Prediction · Feb 2015 AI timelines and forecasting
Mostly true Factual Feb 2015 Basic search with a lot of compute power has worked better than expected in machine intelligence. Factual · Feb 2015 AI capabilities
Own records only Factual Feb 2015 Programs often achieve their fitness function in unpredicted ways, something Altman learned as a student in the Stanford AI lab. Factual · Feb 2015 AI safety and alignment
Unresolvable Prediction Feb 2015 It is very possible that creativity and what we think of as human intelligence are an emergent property of a small number of algorithms operating with a lot of compute power. Prediction · Feb 2015 AI capabilities
Unresolvable Prediction Feb 2015 Machine intelligence progress may look slow and then suddenly go vertical, with things getting out of control quickly. Prediction · Feb 2015 AI timelines and forecasting
Unresolvable Prediction Feb 2015 Human-written programs and computing power are improving exponentially, and self-improving software will improve itself exponentially. Prediction · Feb 2015 AI timelines and forecasting
Unresolvable Prediction Feb 2015 Altman said he thinks other threats, such as an engineered virus with a long incubation period and a high mortality rate, are more certain to happen than SMI but unlikely to destroy every human in the way that SMI could. Prediction · Feb 2015 Other global risks
Unresolvable Prediction Feb 2015 Development of superhuman machine intelligence is probably the greatest threat to the continued existence of humanity. Prediction · Feb 2015 AI safety and alignment
Own records only Factual 2014 Most YC partners independently told Sam Altman that application quality was much higher for this batch. Factual · 2014 Economics and markets
Own records only Factual 2014 In 2014, Altman said the batch he was describing drew nearly 20% more applications than a year earlier, and that average application quality was strikingly higher than for any previous batch. Factual · 2014 Economics and markets
Resolved true Factual 2014 Y Combinator has capacity constraints while scaling, which forces it to reject potentially good companies. Factual · 2014 Economics and markets
Mostly true Factual 2014 Sam Altman experienced plenty of rejection from investors while running his own startup. Factual · 2014 Their own work or organisation
Back to the list

Sam Altman

Open its page

Factual AI capabilities

“Most machine intelligence development involves a “fitness function”—something the program tries to optimize.”

Sam Altman, blog author February 2015 Machine intelligence, part 1 (Sam Altman's blog)

Most machine intelligence development involves a fitness function that the program tries to optimize.

Resolved true The record shows it happened or was so.

The second judge would not resolve this claim; the verdict shown is the first judge's.

What would count as true or false

Judge as of February 2015. True: standard machine learning practice at that time mostly involved optimising an objective (a loss, reward or fitness function); counting loss and objective functions as 'fitness functions' in the general sense the claim intends. Mostly true: this holds for the great majority of methods, and 'fitness function' is a loose but fair term for the objective. Mixed: a large share of significant machine intelligence approaches at the time did not optimise an explicit objective. Mostly false: optimisation-based methods were a minority of the field. False: the field generally did not involve optimising any objective function.

Evidence

  1. Deep learning Nature (review by Yann LeCun, Yoshua Bengio & Geoffrey Hinton) · 28 May 2015
    “We compute an objective function that measures the error (or distance) between the output scores and the desired pattern of scores. The machine then modifies its internal adjustable parameters to reduce this error.”
  2. RLtut.key NIPS 2015 tutorial slides (Sutton and colleagues; lecture slides credit David Silver 2015) · 2015
    “Seeking to maximize its cumulative reward in the long run”
  3. pypop/docs/sphinx/source/user-guide.rst at main · Evolutionary-Intelligence/pypop GitHub (Evolutionary-Intelligence/pypop open-source library documentation) · 26 January 2021
    “First, an *objective function* (also called *fitness function* in evolutionary computation or *loss function* in machine learning) needs to be defined in the function form.”

Reasoning

The 2015 Nature review of deep learning (E1) describes the standard method: a system computes an objective function measuring error and adjusts its parameters to reduce it. The 2015 NIPS reinforcement learning tutorial (E2) describes agents maximising cumulative reward. E3 says objective, fitness and loss functions are names for the same thing across fields, which supports the general sense the criteria accept. Taken together, these independent sources show that optimising an objective was standard practice in the main machine learning approaches around the time of the claim.

Cite this claim

Sam Altman, “Most machine intelligence development involves a fitness function that the program tries to optimize.” (said February 2015). Verdict: resolved true. Reliability API, claim c_a022a45cf9ab, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_a022a45cf9ab

Claims without a verdict

Is a verdict here wrong, or do you speak for Sam Altman? 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.