Lisan al Gaib· @scaling01 · X·· 20 小时前AI 评分5
AI 导读
有时我会发些关于 OpenAI 或 Anthropic 的含糊推文，然后等着看实验室员工的反应（没反应也算一种） 这方式噪音很大、带宽很低，但我觉得他们该开始随机化类似模式了，否则我就要在大概 10^24 年后把 GPT-6 的权重偷出来了
正文
sometimes I vaguepost about OpenAI or Anthropic and then wait for reaction of lab employees (no reaction is also one)
it's very noisy and low bandwidth but I think it's time they start randomizing like patterns or I'm going to exfiltrate GPT-6's weights in like a septillion years
来源：Lisan al Gaib · x.com