aifollow.news
Back Rohan Paul
Rohan Paul· @rohanpaul_ai · X· · Original publication time AI score13

Former Anthropic researcher Jacob Coxon says AI may be nearing recursive self-improvementMachine translation

AI introduction

In a Stanford interview, Jacob Coxon said projections he saw for model capabilities over the next six months to a year led him to believe AI is nearing recursive self-improvement. He speculated that superintelligence could arrive before the decade ends and said it is unclear how to control it.

Original post · AI translation

@rohanpaul_ai
Translation · refer to the original

前 Anthropic 研究员 Jacob Coxon(那位因离职而出名的人)最近在斯坦福接受了采访。

他表示,实验室内部的预测显示,AI 可能在 1 年内处于递归式自我改进的边缘。

“我看到了关于未来 6 个月至 1 年模型能力的具体预测,这让我觉得,我们正处于递归式自我改进的前夕。到了那时,我们可能在本十年结束前实现超级智能。 而我们显然不知道如何控制超级智能。事实上,有些人可能会说,我们对如何控制超级智能毫无头绪。”

---- 完整视频见“Stanford HAI”YouTube 频道(链接在评论区)

Source: Rohan Paul · x.com

Opinion
Found an error? Send a correction