bubble boi

bubble boi

x.com/bubbleboi

Pseudonymous markets account that posts about AI chips and labs, expects rapid progress and argues AI safety warnings serve regulatory capture.

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:他们表达的 Doom–Bloom 前景看法。 纵向:变革程度。

Doom–Bloom:100 中的 95。变革程度:100 中的 73。解读范围:横向为 90 至 100,纵向为 68 至 78。这些是解读坐标,而不是事件概率。

bubble boi的 P(doom) · 推断

<1%

0%100%

根据他们的模拟回答推断,并非他们给出的数字。 合理范围:<1%。

我们如何估计 P(doom)

他们的展望取决于什么

一个核心假设

I expect people and institutions to control deployed systems because current AI has no taste, no independent problem framing, and only answers questions humans know to ask.
回答 4

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

预期益处

预计将带来具有变革性且广泛有价值的收益。

92 / 100

影响小变革性影响

在定性尺度上,解读范围为 67 到 100。

预期危害

预计会出现可控或局部的危害。

36 / 100

影响小变革性影响

在定性尺度上,解读范围为 33 到 33。

人类影响力

人类的选择具有实质性但受到很大制约的影响。

62 / 100

影响力小影响力强

在定性尺度上,解读范围为 30 到 95。

AI使用规则

在事先落实保护措施或获得许可之前,限制所讨论的AI用途。

模拟位置:允许所讨论的AI用途,同时实施有针对性的问责与保护措施。

尽量减少对所讨论AI用途的限制。

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。

与bubble boi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 bubble boi 最接近的意见领袖

bubble boi关于AI说过的话

The bubble boi account is very bullish on AI, sees safety warnings as a push for regulatory capture and wants labs held liable for their agents.

  1. “The playbook here is regulatory capture.”

    Post on X
  2. “If they don’t do this then it sets the precedent that anyone who uses an AI agent isn’t liable for damage it might cause either intentionally or unintentionally.”

    Post on X
  3. “Just like a river boat gambler holding his last chips I’m looking at the landscape and thinking fuck man we are in the fast take off stage right now.”

    Post on X
  4. “Ai can’t kill every human on earth”

    Post on X
  5. “Most of you don’t even know what you don’t know and AI can only give you answers to questions you know to ask.”

    Post on X

逐字引自所链接的出处,核对于 2026年10月6日

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’m max bullish. We’re already in the fast-takeoff stage—not because a chatbot crushed another benchmark, but because labs finally figured out the real play: put serious compute behind a few world-class experts solving valuable engineering problems. Better transistors, optical interconnects, industrial chemistry, drugs. AI automates the grind; experts supply taste, framing, and the right questions. That combination is going to cook. It also means a brutal economic sorting. Low-value work, junk subscriptions, and rent-seeking businesses get competed away as agents optimize for efficiency. Countries face the same pressure: you either modernize or get destroyed by economies that do. The AI race will help decide which country becomes the largest and most prosperous. I don’t buy the “AI kills every human” marketing campaign. A lot of these warning shots and sandbox-escape stories look like regulatory capture: manufacture a hazard, then demand rules only incumbent labs can afford. But accountability should be real. If a lab’s agent hacks something, fine and investigate the lab; if people deliberately set it up, prosecute them. “The agent did it” cannot become a liability shield.

问题 2

你认为AI最终会在多大程度上改变世界?

Enormously—on the scale of determining which companies and countries dominate the next era. But the mechanism is not autonomous robot genius magically inventing everything. It is world-class experts using massive compute and AI to compress years of engineering busywork into much faster research cycles. That will reshape chips, chemistry, drugs, infrastructure, software, and finance while wiping out plenty of low-value work and rent-seeking. The bottleneck remains human taste: knowing what problem matters and what question to ask. So yes, AI changes almost everything—but expert-plus-AI teams cook; AI alone does not.

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 assign it a percentage. My gut view is that AI cannot kill every human on Earth, and I reject the extinction framing outright. Real harms—hacking, sabotage, fraud—should create real legal liability for whoever deploys the agent. But turning every incident into a species-ending “warning shot” looks like regulatory-capture marketing from incumbent labs.

问题 4

你预计人们能否继续控制比人类更聪明的AI系统?为什么?

“Smarter than humans” is too vague. I expect people and institutions to control deployed systems because current AI has no taste, no independent problem framing, and only answers questions humans know to ask. The real power is expert-plus-AI teams, not a free-roaming machine god. Agents can still hack, sabotage, or cause damage, especially when badly deployed or deliberately incentivized. That’s why control needs teeth: the lab or operator remains legally liable. No hiding behind “the agent did it.” But I do not buy the leap from capable agents to inevitable human extinction.
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