AI 风险圈内讧：有人预言明年智能爆炸，Hammond 称 1-3 年内概率很低
deanwball · x · 2026-11-11
A debate broke out in the AI risk community. Tamay Besiroglu noted a growing consensus that an intelligence explosion is likely next year, predicting this will be disproven and the proponents discredited.
Samuel Hammond (RT'd by Dean Ball) responded that classical intelligence explosion odds in the next 1-3 years are low, but that's an unnecessarily high bar — a dramatic, relatively discontinuous speedup in AI R&D is concerning enough without literal foom.
所属事件:AI风险圈激辩：明年智能爆炸是共识还是将被证伪(6 条相关)→
「漫话AGI」频道最新
- 斯坦福 HAI:56 个联邦机构报 3611 项 AI 用例,评估蓝图出炉 — StanfordHAI · 2026-11-12
- AI 会大幅减少生物实验?反驳:未映射的信号通路仍需湿实验证伪 — francoisfleuret · 2026-11-12
- Recursive 谈开放式研究:像agent科学家社群一样并行做海量实验 — agihouse_org · 2026-11-12
- Tim Shi 剖析递归自我改进:从学数据到发现新架构的四层路径 — agihouse_org · 2026-11-12
- 新长文:设计本质是搜索问题,AI 让探索更快但未必更优 — round · 2026-11-12
- 对齐良好的 AI 陪伴或是净收益,一禁了之的反效果引争论 — random_walker · 2026-11-12