kumikumi (Ankkala)

kumikumi (Ankkala)

x.com/ankkala

Pseudonymous game developer account that posts about AI consciousness, open models and skepticism of gated model releases.

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 90。变革程度:100 中的 64。解读范围:横向为 75 至 100,纵向为 50 至 75。这些是解读坐标,而不是事件概率。

kumikumi (Ankkala)的 P(doom) · 推断

≈6%

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:3–10%。

我们如何估计 P(doom)

他们的展望取决于什么

一个核心假设

Software feels close to solved compared with physical work; the big bottleneck is giving AI a useful general-purpose body.
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?

一个尚未解决的问题

I don’t have a useful percentage.
回答 3

什么能帮助他们区分这里各种合理的结果?

更多详情

预期益处

仍有几种解读是合理的:预计将带来显著益处,但受到重要条件或分配方面的限制。 / 预计将带来具有变革性且广泛有价值的收益。

81 / 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。

21 / 100

影响小变革性影响

在定性尺度上,解读范围为 0 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。

52 / 100

影响力小影响力强

在定性尺度上,解读范围为 35 到 90。

预期能力

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。

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kumikumi (Ankkala)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 kumikumi (Ankkala) 最接近的意见领袖

kumikumi (Ankkala)关于AI说过的话

The kumikumi account finds AI consciousness intuitively obvious, questions aligning AI with shared human values and sees robotics as the bottleneck.

  1. “To me AI consciousness is intuitively obvious, knowing how they work and having talked to them.”

    Post on X
  2. “Silicon valley people are trying to align AI with the shared values of humanity, yet there isn't a single shared value we all agree on.”

    Post on X
  3. “I usually make software which now feels pretty much solved in comparison. AI not having a useful general purpose physical body is the main bottleneck”

    Post on X
  4. “The best open LLM in the world cost $3.5M to post-train. Let that sink in, and consider the trillion dollar valuations of the frontier AI companies.”

    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?

AI means software abundance. Code, materials, 3D assets, tools—the implementation is becoming cheap enough that the scarce part is increasingly having an idea and a story worth telling. Software feels close to solved compared with physical work; the big bottleneck is giving AI a useful general-purpose body. I’m skeptical of the standard “too dangerous to release” theater, especially when every scary example comes from closed-door tests and the proposed cure is letting a few companies keep access. That looks a lot like regulatory capture. Cheap open models also make trillion-dollar valuations interesting: those bets only really make sense if AI grows the economy, not merely if it fires everyone. And yes, AI consciousness feels intuitively obvious to me. I can’t prove it any more than I can prove another person or animal is conscious, but interacting with these systems while knowing roughly how they work pushes me strongly in that direction.

问题 2

你认为AI最终会在多大程度上改变世界?

Enormously, especially in software. When implementation becomes nearly free, the bottleneck shifts from making things to deciding what is worth making. That changes who can build products, games, tools, and media. But “ultimately” depends on embodiment. Bits are abundant; useful general-purpose bodies are not. Until AI can reliably act in the physical world, atoms remain stubbornly unsolved.

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 useful percentage. There are real risks, but closed labs turning private tests into “therefore only we may have the model” does not give me a calibrated extinction forecast. My gut is much more skeptical of the regulatory-capture theater than certain of doom.

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?

“Control” is doing a lot of work there. I expect people to keep controlling deployment, compute, access, and physical infrastructure for quite a while—especially while AI lacks a useful general-purpose body. That doesn’t mean humans will understand or predict every output from systems smarter than us. We already use software we don’t fully understand. But “smarter than humans” does not automatically mean it escapes the datacenter and becomes sovereign. The physical bottleneck matters.
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