ClaimsPeople

Marc Andreessen

Marc Andreessen co-created the Mosaic browser, cofounded Netscape, and is cofounder and general partner of the venture capital firm Andreessen Horowitz.

1,009 claims collected, 216 researched, 105 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?

    67 61 73

    High confidence, 94 factual claims resolved. From 11 documents, 1995 to 2026; 19 documents sampled from 30 candidates found.

    Not scored: 44 unresolvable, 17 checkable only against their own records.

  • Predictions

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

    60 44 76

    Low confidence, 11 predictions resolved. From 6 documents, 1995 to 2025; 19 documents sampled from 30 candidates found.

    Not scored: 41 unresolvable, 1 checkable only against their own records. 4 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?

    Insufficient record: 0 promises resolved, 5 needed for a score.

    Not scored: 3 unresolvable.

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

20 claims, sorted by newest said. Choose a claim to read it in full.
VerdictKindSaidClaimTopic
Resolved true Factual 7 Jan 2026 Within 6 to 12 months of a frontier model's release, a small model with the same capabilities appears. Factual · 7 Jan 2026 AI capabilities
Resolved true Factual 7 Jan 2026 Consumer AI products commonly have monthly price tiers of $200 or $300. Factual · 7 Jan 2026 AI, the economy and work
Resolved true Factual 7 Jan 2026 There are 5 billion people on Earth using some version of mobile broadband Internet. Factual · 7 Jan 2026 Technology
Resolved true Factual 7 Jan 2026 Cursor, though it started by calling large models, is building its own AI model. Factual · 7 Jan 2026 AI, the economy and work
Resolved true Prediction 26 Jan 2025 Andreessen said energy independence and large net energy exports are a matter of choice, and that he thinks the new Trump administration will choose to pursue them. Prediction · 26 Jan 2025 AI policy and governance
Resolved true Prediction 2 Sep 2023 There will be Chinese AI systems that embody Communist ideology. Prediction · 2 Sep 2023 AI policy and governance
Resolved true Factual 12 Jun 2023 The Transformer algorithm was invented in 2017. Factual · 12 Jun 2023 AI capabilities
Resolved true Factual 12 Jun 2023 The AI being discussed is not, strictly speaking, human intelligence. Factual · 12 Jun 2023 AI capabilities
Resolved true Factual 18 Apr 2020 As of April 2020 there were no therapies or vaccine for the coronavirus, despite years of warnings about bat-borne coronaviruses. Factual · 18 Apr 2020 Biosecurity
Resolved true Factual 18 Apr 2020 As of April 2020 the US lacked enough coronavirus tests and test materials, including cotton swabs and common reagents. Factual · 18 Apr 2020 Other global risks
Resolved true Prediction 18 Apr 2020 The essay "It's Time to Build" will be the target of criticism. Prediction · 18 Apr 2020 Their own work or organisation
Resolved true Factual 1 Jan 2018 Andreessen Horowitz has since reversed course almost completely and is now deeply invested in healthcare. Factual · 1 Jan 2018 Their own work or organisation
Resolved true Factual 1 Jan 2018 Supersonic passenger flight with the Concorde existed previously but no longer exists. Factual · 1 Jan 2018 Technology
Resolved true Factual 1 Jan 2018 In the American economy, the rate of job creation and destruction is falling rather than rising. Factual · 1 Jan 2018 AI, the economy and work
Resolved true Factual 1 Jan 2018 Warren Buffett invested in Heinz ketchup. Factual · 1 Jan 2018 Economics and markets
Resolved true Factual 1 Jan 2018 Andreessen said "we" repeatedly see people fail because they price too low ("too hungry to eat") and so cannot afford the sales and marketing needed to get anybody to buy the product. Factual · 1 Jan 2018 Economics and markets
Resolved true Factual 20 Aug 2011 Marc Andreessen said Disney had to buy Pixar, a software company, to remain relevant in animated movies. Factual · 20 Aug 2011 Economics and markets
Resolved true Factual 20 Aug 2011 Videogame makers are today's fastest growing entertainment companies, with the industry growing to $60 billion from $30 billion five years earlier. Factual · 20 Aug 2011 Economics and markets
Resolved true Factual Jun 1995 After the company shipped product in December, its revenues went from zero to a level where it was pretty much able to fund the bulk of its work and development. Factual · Jun 1995 Economics and markets
Resolved true Factual Jun 1995 Yahoo was started by Ph.D. students at Stanford. Factual · Jun 1995 Technology
Back to the list

Prediction AI policy and governance

“Look, you're going to have Chinese AI.”

Marc Andreessen, venture capitalist, interviewee 2 September 2023 Marc Andreessen on Why AI Will Save the World (EconTalk)

There will be Chinese AI systems that embody Communist ideology.

Resolved true The record shows it happened or was so.

What would count as true or false

The claim is that Chinese AI systems will exist that embody the ideology of the Chinese Communist Party. TRUE: Chinese-developed AI systems, such as large language models or chatbots from Chinese firms, were publicly released, and documentation shows they are built or required to follow CCP or state ideology. Qualifying documentation includes Chinese regulations requiring generative AI to uphold 'core socialist values', regulatory ideological vetting of models, or documented testing showing that models systematically give Party-line answers or refuse politically sensitive topics. MOSTLY TRUE: Chinese AI systems were deployed and censor politically sensitive topics in line with state requirements, but the evidence of positive ideological alignment beyond censorship is thin. MOSTLY FALSE: Chinese AI systems exist but show no meaningful ideological constraint beyond ordinary content moderation. FALSE: no significant Chinese AI systems were released, or credible testing shows they do not reflect state ideology.

Evidence

  1. Interim Measures for the Management of Generative Artificial Intelligence Services Lexis (lexiscn.com), English text of CAC and other ministries' Order No. 15 · 10 July 2023
    “Core socialist values shall be upheld, and it is prohibited to generate any content prohibited by laws or administrative regulations, such as content inciting subversion of national sovereignty or the overturn of the socialist system”
  2. Political censorship in large language models originating from China PubMed Central (National Library of Medicine) · 17 February 2026
    “we find substantially higher rates of refusal to respond, shorter responses, and inaccurate responses to a battery of 145 political questions in China-originating models.”
  3. On DeepSeek, you can watch AI navigate censorship in real time NBC News · 30 January 2025
    “Repeating a question sometimes generated different results, but in every instance, DeepSeek either declined to answer or produced an answer that took an explicitly pro-Chinese government stance, whereas ChatGPT’s responses appeared consistently more neutral or in line with non-Chinese sources.”
  4. Chinese AI Models Repeat State Doctrine on Sensitive Topics or Refuse to Answer – Aleph Alpha Documents i6eal.de (Ideal Syka), reporting on an Aleph Alpha analysis · 4 October 2026
    “Chinese AI models such as Alibaba's Qwen, DeepSeek, and Moonshot AI (Kimi) frequently respond along party lines on politically sensitive topics or refuse to answer.”

Reasoning

China's Interim Measures for generative AI services require these services to uphold 'core socialist values' and ban content that incites the overturn of the socialist system (E1). NBC News found that DeepSeek, a publicly released Chinese model, either declined to answer politically sensitive questions or answered with an explicitly pro-government stance (E3). An academic study found much higher refusal and inaccuracy rates on political questions for China-originating models (E2), and an Aleph Alpha analysis found that Qwen, DeepSeek and Kimi often answer along party lines or refuse to answer (E4). Together this meets the TRUE criteria: the models were released, regulation requires them to follow state ideology, and testing documents Party-line answers.

Cite this claim

Marc Andreessen, “There will be Chinese AI systems that embody Communist ideology.” (said 2 September 2023). Verdict: resolved true. Reliability API, claim c_6a0b141f732d, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_6a0b141f732d

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.