科学家称 OpenAI GPT 加上 辉达/英伟达 CUDA 不可能通往 AGI之路？人类害怕 ASI？公开信要求全球合作ban AI 直到证明安全

aidj

📜 Superintelligence is in Trouble for Global Ban
We just got another flashpoint in the AI safety war: a new open letter backed by top thinkers is calling on governments to ban the development of artificial superintelligence (ASI).
The Future of Life Institute released the letter, urging world govs to prohibit ASI development at least until we’re 100% sure it’s controllable and society says yes.

The signers argue unchecked ASI could lead to:
Human obsolescence in the economy
Loss of civil liberties and autonomy
Complete extinction, in worst-case scenarios
Do you know who signed it? → Yoshua Bengio and Geoffrey Hinton (Godfathers of AI) + Steve Wozniak (Apple co-founder) + Leo Gao (OpenAI staffer) & more.

But notably absent leadership we noticed from the biggest AI labs like: OpenAI / Google DeepMind / Anthropic / Meta / xAI (all big names right?)

Why it matters:
Just 5% want companies to build it without oversight. But there’s no clear definition of what “superintelligence” even is. And without cooperation from OpenAI, Google, or Meta, this could end up more symbolic than effective.

aidj

另一面好像是在说 中国AI 更危险的信号

(相关课题如果出现在中国网站，会酱负责任的客套话等）

挡 AI 巨头们和 台湾供应商的财路 “可恶”团体。

（禁止AI发展，没有什么晶片需求就没有什么芯片供应）

https://youtu.be/zwiui7SvX7U

aidj

结论：

当 通用人工智能（AGI）到达的时候
通过 recursive self-improving 进化
最长只需要 12月的时间 upgrade 到
超级人工智慧（ASI) 来不及
out of Human control
偏偏这个时候  govt 投入 全国资源
明知道自己 出了 safety 问题, 侥幸心理
也不会 打掉重炼 导致 Human extinction

https://youtu.be/UrPORCdfwy8?t=109

aidj

aidj 发表于 9-11-2025 07:22 AM
结论：

当 通用人工智能（AGI）到达的时候

https://youtu.be/7Fy5R4XljdE

aidj

vs

https://youtu.be/3df6rcsd2eY

aidj

Sam Altman 是不是搞到人很乱？
一边要 AI bailout 一边说 要 slow down AI development 因为存在危险



截图来自https://youtu.be/1Bud__Er-j4