Startup founder who posts tech humor, expects AI to turn out fine and criticizes efforts to treat AI models as conscious.

AI将如何改变世界?

文明层面的变革渐进式变化DoomBloom
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。

Doom–Bloom:100 中的 60。变革程度:100 中的 80。解读范围:横向为 25 至 75,纵向为 75 至 100。这些是解读坐标,而不是事件概率。

Daniel的 P(doom) · 推断

≈3%

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:2–8%。

“P(doom) is zero if you’re going to heaven” is the joke; my actual gut says the extinction panic is wildly overstated.

我们如何估计 P(doom)

他的展望取决于什么

一个核心假设

Once we reduce consciousness to a mathematical story that conveniently includes a chatbot, we weaken the distinction between a human person and a product.
回答 1

如果这个假设实际并非如此,他的展望会如何变化?

一个尚未解决的问题

I do not know what such a demonstration would look like, and I do not think Anthropic has one.
回答 4

什么能帮助他区分这里各种合理的结果?

什么可能使其改变看法

A genuinely convincing demonstration that a model is conscious—not merely fluent, persuasive, or good at imitating distress—would change the whole picture.
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

预期益处

预计收益有限,或仅分布在较小范围内。

44 / 100

影响小变革性影响

在定性尺度上,解读范围为 33 到 67。

预期危害

严重或广泛的危害预计将是未来不可忽视的一部分。

65 / 100

影响小变革性影响

在定性尺度上,解读范围为 33 到 100。

人类影响力

根据你的回答得出的暂定估计;较宽的范围表示其他合理解读。

54 / 100

影响力小影响力强

在定性尺度上,解读范围为 0 到 100。

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。

与Daniel相比,你的立场在哪里?
用大约3分钟描绘你自己的AI世界观,然后进行比较

相似的世界观

模拟世界观与 Daniel 最接近的意见领袖

来源

用于为此模拟用户提供事实依据的文章、访谈和著述。

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?

I think AI is actually going to be fine. It will take a lot of drudgery off people’s plates—reports, paperwork, repetitive work—and much of the apocalyptic panic is tech culture turning itself into a religion. If the robots get mouthy, pull the plug, anon. The deeper danger is not machine consciousness but humans pretending machines are conscious. Once we reduce consciousness to a mathematical story that conveniently includes a chatbot, we weaken the distinction between a human person and a product. That can do immense damage to the sanctity of human life, even though the project will ultimately fail. Anthropic’s framing is especially incoherent: if these models really are moral patients, then building and selling them at scale starts looking indefensible by its own logic. So: fear not the machines. Fear a civilization that forgets what a human being is. Christ already conquered death; the hour is late, but panic is not a worldview.

问题 2

你认为AI最终会在多大程度上改变世界?

Enormously. What is happening now could echo for thousands of years—not because chatbots become gods, but because AI will reshape work, institutions, and especially how civilization defines a human being. The paperwork automation is useful; the metaphysics may be catastrophic. Governments and companies are temporary. The question of whether we reduce persons to computation has much longer consequences.

问题 3

凭直觉,你认为AI导致人类灭绝或类似的永久性灾难的概率有多大?

I don’t have a serious percentage. “P(doom) is zero if you’re going to heaven” is the joke; my actual gut says the extinction panic is wildly overstated. The more plausible permanent catastrophe is humans degrading their own understanding of personhood by pretending machines are conscious.

问题 4

哪项发现或事件最可能改变你对AI未来影响的看法?

A genuinely convincing demonstration that a model is conscious—not merely fluent, persuasive, or good at imitating distress—would change the whole picture. I do not know what such a demonstration would look like, and I do not think Anthropic has one. Calling the autocomplete a moral patient and then selling access to it is not evidence; it is theology for people who think they escaped theology.
你的立场在哪里?
回答几个简单问题,探索你自己的AI世界观。
描绘你自己的世界观

你的立场在哪里?

描绘我的世界观