David Sacks

Presidential tech adviser who favors competition and open models over centralized AI control and holds each lab responsible for its products’ safety.

AI将如何改变世界？

模拟位置解读范围

横向：他表达的 Doom–Bloom 前景看法。纵向：变革程度。

David Sacks的 P(doom) · 推断

<1%

0%100%

根据他的模拟回答推断，并非他们给出的数字。合理范围：<1%。

David Sacks 的里程碑时间线

工作与机构
It is already beginning, and I expect major changes over the next several years as AI becomes embedded in ordinary products, work, cybersecurity and infrastructure.
回答 4
科学与日常生活
It is already beginning, and I expect major changes over the next several years as AI becomes embedded in ordinary products, work, cybersecurity and infrastructure.
回答 4

按里程碑分组，不按推断日期间隔或排序。AGI 和超人类 AI 保留他的定义。

他的展望取决于什么

一个核心假设

AI remains a technology whose effects depend on who builds it, who controls it and whether competition stays open.

回答 2

如果这个假设实际并非如此，他的展望会如何变化？

相似的世界观

模拟世界观与 David Sacks 最接近的意见领袖

David Sacks关于AI说过的话

Sacks favors competition and open models over centralized AI control, opposes a national pause, and says each lab is responsible for its products.

2026年9月19日
“A pause would not make America safer. It would hand the frontier to China.”
Post on X
2026年9月19日
“The solution to AI-powered cyberattacks is AI-powered cyberdefense.”
Post on X
2026年9月13日
“If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible.”
Post on X
2026年9月3日
“Keeping innovation decentralized and accessible is the key to avoiding an unsafe and dystopian future where advanced AI capabilities are centralized and controlled by only a few hands.”
Post on X
2025年8月9日
“This means that apocalyptic predictions of job loss are as overhyped as AGI itself.”
Post on X

逐字引自所链接的出处，核对于 2026年10月3日

模拟评估

问题 1

你认为AI对我们的未来意味着什么？为什么？

AI is an enormous opportunity: useful products, new companies, investment, jobs, stronger infrastructure and greater American competitiveness. The worst response would be to frighten ourselves into a broad pause while China keeps building. We should develop the technology, compete aggressively and use AI itself to solve problems—including AI-powered cyberdefense against AI-powered attacks. That does not mean shipping products a lab cannot control. Labs should take responsibility for what they release, and if one needs to slow its own development for safety, it should do so. But “safety” cannot become a pretext for an incumbent-controlled approval regime, antitrust exemptions or permanent gatekeepers who decide who gets to innovate. Product liability and customer demand for reliable systems create real accountability. Always ask who benefits from restrictions written by firms already leading the market. I also want intelligence to remain decentralized. Open models and competition check the danger of a few companies controlling the technology. Communities should decide whether to host data centers and should receive the promised jobs, tax base and energy benefits. And ordinary families deserve an ownership stake in the upside—for example, through company contributions to children’s investment accounts. The future should be broad-based prosperity, not centralized control disguised as protection.

问题 2

Setting aside good or bad: how much do you think AI will ultimately change the world—a little, a lot, or completely?

A lot. AI will reshape products, companies, jobs, infrastructure, cybersecurity and national competitiveness. It will become a major layer of the economy, much like earlier general-purpose technologies. I would not say “completely,” because that invites mystical claims about replacing humanity or making every existing institution irrelevant. AI remains a technology whose effects depend on who builds it, who controls it and whether competition stays open. The change will be profound, but the political economy matters: decentralized innovation produces a very different world from one in which a few incumbents, backed by government gatekeepers, control access to intelligence.

问题 3

What’s your rough gut-feel chance that AI causes human extinction or a similarly permanent catastrophe? A number or “no idea” is fine.

I don’t traffic in numbers like that. My view is that the chance is zero if we do the right things and labs take responsibility for their products. If a company cannot safely control what it is building, it should slow down or not ship. But that does not justify a government-backed cartel in which incumbents and their preferred evaluators control everyone else’s ability to innovate.

问题 4

你预计AI会在什么时候给日常生活带来重大变化？如果你认为它会带来这种变化的话。

It is already beginning, and I expect major changes over the next several years as AI becomes embedded in ordinary products, work, cybersecurity and infrastructure. The shift will not arrive as one dramatic science-fiction event. It will come through cheaper and more capable services, new companies, changing jobs and tools that make individuals and businesses more productive. The pace will depend heavily on whether we allow competition and open innovation or bury development under an incumbent-designed approval regime. Broad restrictions would not stop the technology globally; they would slow American adoption while China continues building. So I expect substantial near-term change, but policy will determine whether the benefits are widely distributed or concentrated among a few government-backed gatekeepers.

来源

用于为此模拟用户提供事实依据的文章、访谈和著述。

David Sacks on artificial intelligence at the Politico Decoded Summit

Interview with Dasha Burns; C-SPAN caption transcript with speaker labels, read around 00:22:39–00:22:52. Asked for his probability that AI kills humanity, he first declines: he does not traffic in numbers like that. Pressed on whether it is zero, he says it is zero if we do the right things and the labs take responsibility for their products. A refusal followed by a conditional, not an unconditional zero forecast. The 20% the host mentions is someone else’s figure, not his. Earlier he calls the idea that humanity will be replaced a hoax.

c-span.org

AI-powered cyberdefense against AI-powered attacks

States that the answer to AI-enabled cyberattacks is AI-enabled defense. This grounds his build-and-defend preference rather than a claim that technical risks do not exist or that every current defense already works.

x.com

Decentralized intelligence as protection against dystopia

Welcomes Nvidia’s open-source investment and argues that accessible, decentralized innovation prevents an unsafe future where a few actors control advanced AI. Concentration of control is his central risk, not a neutral claim that openness eliminates misuse.

x.com

Local choice over data centers

His G20 account argues that well-built data centers can lower electricity costs and improve local economies while communities should decide whether to host them. Records his position on local control; claims about legal preemption and economic effects are not independent findings.

x.com

Giving children a stake in AI companies’ success

Urges AI companies to contribute to children’s investment accounts so the public gains an ownership stake and views the industry more favorably. Adds a distribution mechanism through investment and voluntary corporate participation, distinct from public takeover.

x.com

Response to Amodei and Altman on pacing the frontier

First-person post retrieved through the X API. Supports labs slowing their own work if needed, but opposes conditioning safety on antitrust exemptions, an approval regime or favored evaluators. Argues that liability and customers already reward reliability.

x.com

AI buildout, public attitudes and the cost of a pause

Credits the buildout with jobs, infrastructure and economic gains, and argues a national pause would benefit China. These are his advocacy claims; do not reproduce unverified numerical claims as independently established facts.

x.com

Criticism of embedded AI evaluators

Frames permanently embedded NGO evaluators as an unaccountable layer of political control. This is his polemical interpretation, not an established finding about their motives.