问题 1
David Dalrymple
x.com/davidadAI safety researcher who works on mathematical verification for AI systems and argues that frontier models can learn a natural sense of what is good.
AI将如何改变世界?
横向:他们表达的 Doom–Bloom 前景看法。 纵向:变革程度。
Doom–Bloom:100 中的 75。变革程度:100 中的 94。解读范围:横向为 70 至 80,纵向为 89 至 100。这些是解读坐标,而不是事件概率。
<5%
“I like my, my P doom is less than 5% now.”
Residual AI doom, which he decomposes into being wrong about the wisdom attractor, Malthusian competition for land and energy (~1%), catastrophic (bio) misuse (~1%), a military first strike, and conflict between strong but violent AI coalitions
Alignment with Awakening: Davidad on Moral Realism, AI Wisdom, & why His p(Doom) is Down to 5% · 2026年7月
一个核心假设
My current expectation is substantially more optimistic than it used to be: frontier models appear capable of learning fairly general moral abstractions, rather than merely imitating isolated rules.回答 1
如果这个假设实际并非如此,他们的展望会如何变化?
更多详情
预计将带来具有变革性且广泛有价值的收益。
87 / 100
在定性尺度上,解读范围为 67 到 100。
严重或广泛的危害预计将是未来不可忽视的一部分。
61 / 100
在定性尺度上,解读范围为 33 到 67。
人类的选择具有实质性但受到很大制约的影响。
50 / 100
在定性尺度上,解读范围为 49 到 51。
预计AI仍将是能力有限的工具。
预计AI将在大多数认知工作中达到人类水平。
模拟位置:预计AI将在认知工作中大幅超越人类。
停止或大幅放缓开发能力更强的AI。
模拟位置:在落实所述保障措施的前提下继续开发。
加快开发能力更强的AI。
这些解读保留了他们陈述的条件。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们的模拟回答,而不是统计置信区间。
相似的世界观
模拟世界观与 David Dalrymple 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。
Coauthored framework specifies world model, safety specification and verifier; outlines unresolved technical challenges.

Direct statement explains combining scientific models and proofs for stronger safety assurances.

2026 institutional update records a pivot toward assurance tooling and cybersecurity; do not present the original programme architecture as already achieved.

Explains that frontier capabilities outran programme expectations, motivating a pivot toward reusable verification and auditability tools rather than bespoke model development; highlights critical-infrastructure security.

Davidad explains his 2025–26 shift toward believing frontier models learn a natural moral abstraction; acknowledges deception persists and robustness is incomplete. Gabriel Alfour contests the thesis; his objections are not Davidad’s beliefs.

In the transcript, Dalrymple gives below 5% residual doom risk (later described as roughly 5%), favors aligned AI coalitions and moral autonomy, regards biological disempowerment as inevitable but not necessarily bad, and supports international misuse restrictions instead of hoping for universal slowdown.

模拟评估
问题 2
你认为AI最终会在多大程度上改变世界?
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?
问题 4
你预计开发AI的人员和机构会如何应对其风险?
你的立场在哪里?
描绘我的世界观