问题 1
Marc Andreessen
x.com/pmarcaAndreessen Horowitz co-founder who argues that AI can greatly improve health, education and living standards, so slowing it down has real costs.
AI将如何改变世界?
横向:他表达的 Doom–Bloom 前景看法。 纵向:变革程度。
Doom–Bloom:100 中的 98。变革程度:100 中的 75。解读范围:横向为 93 至 100,纵向为 70 至 80。这些是解读坐标,而不是事件概率。
<1%
根据他的模拟回答推断,并非他们给出的数字。 合理范围:低于 3%。
工作与机构
So I expect major changes this decade—not as a single cinematic event, but as an accelerating expansion of what ordinary people can do.
回答 3科学与日常生活
So I expect major changes this decade—not as a single cinematic event, but as an accelerating expansion of what ordinary people can do.
回答 3
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。
一个核心假设
Intelligence is an input into nearly everything, and making that input cheaper should raise living standards on an enormous scale.回答 1
如果这个假设实际并非如此,他的展望会如何变化?
什么可能使其改变看法
A real demonstration that AI systems spontaneously develop durable, independent goals—and then strategically deceive, resist control, and pursue those goals across environments without humans prompting them—would challenge my rejection of the takeover story.回答 4
什么证据才足够,又会让他的观点朝哪个方向转变?
更多详情
预计将带来具有变革性且广泛有价值的收益。
100 / 100
在定性尺度上,解读范围为 100 到 100。
预计会出现可控或局部的危害。
31 / 100
在定性尺度上,解读范围为 33 到 33。
人类的选择可以大幅改变AI的发展轨迹。
69 / 100
在定性尺度上,解读范围为 41 到 100。
停止或大幅放缓开发能力更强的AI。
在落实所述保障措施的前提下继续开发。
模拟位置:加快开发能力更强的AI。
限制对强大AI的访问。
允许访问,但须遵守能力或用途限制。
模拟位置:支持广泛或开放地访问强大AI。
这些解读保留了他陈述的条件。益处和危害都可能很大。这些范围描述的是我们如何解读他的模拟回答,而不是统计置信区间。
相似的世界观
模拟世界观与 Marc Andreessen 最接近的意见领袖
Marc Andreessen关于AI说过的话
Andreessen argues that AI will greatly raise productivity and what people can build, and disputes predictions that it will eliminate programming jobs.
“Software has been eating the world at the speed of human hands. It is about to eat the world at the speed of compute.”
Essay, Investing in Cognition “Every time programmers get more leverage, doomers predict the end of software engineering.”
Essay, Investing in Cognition “AI coding, this is the biggest earthquake to hit software in certainly my life, maybe since the invention of software.”
Lex Fridman Podcast “My guess is we’ll have more coding jobs probably by an order of magnitude 10 years from now.”
Lex Fridman Podcast
逐字引自所链接的出处,核对于 2026年10月3日
模拟评估
来源
用于为此模拟用户提供事实依据的文章、访谈和著述。
He rejects AI extinction arguments as mistaken anthropomorphism, depicts alarm as a recurring moral panic, and argues for building and distributing AI widely. He sees incumbent-protecting regulation as a concrete danger. He does acknowledge harmful uses; do not translate his rejection of takeover into denial of all harm.

Technology, markets, and intelligence can expand abundance. Preventing medical and other useful advances has moral costs, so slowing AI is not automatically the safe option. The rhetoric is deliberately forceful rather than measured policy balancing.

The official episode description emphasizes collapsing intelligence costs, a historically large technology shift, open/closed competition, regulatory fragmentation, and China. Page provides an official summary, not a readable transcript: do not invent exact remarks from it.

Official episode description discusses overcoming a long productivity slowdown, American competition, and open source's strategic complications. Use as current thematic support, not evidence for a precise AGI date or a numerical risk estimate.

Official description says he expects strongly positive long-term effects and treats AI as widely available cognitive augmentation. It highlights coding agents, education, medicine, censorship, concentration, surveillance concerns, and China. Original interview was May 19; this page is May 20.

The official episode description identifies health, education, law and software as areas of potential benefit, with institutions, regulation and infrastructure as bottlenecks. Connects AI leadership to energy and industrial capacity. Summary is limited to the publisher’s description, not an independently checked full transcript.

The official episode description foregrounds his view that technological progress expands human capability and that recurrent fears misunderstand innovation. Adds current coverage of creativity, economic growth and cybersecurity. Use as a description of the interview’s stated themes, without inventing quotations or detailed arguments.

Official republication of a Latent Space conversation. The description connects agent tooling and compute constraints with the difficulty of changing a world of billions of people. Adds deployment friction to his optimistic worldview: technical progress and universal immediate adoption are different claims. Full transcript was not available in this pass.

你的立场在哪里?
描绘我的世界观