The Decoder · 2026/10/22 01:10:46
Hassabis 将 AGI 预期缩短至两三年
Google DeepMind 创始人 Demis Hassabis 将其对通用人工智能（AGI）到来的预测从三至五年大幅缩短至两至三年。他提出以“爱因斯坦测试”作为衡量标准，即系统能否仅凭1901年前的知识复现爱因斯坦1905年的论文。此外，Hassabis 呼吁建立由政府监管的行业监督委员会，在模型发布前进行审查，并警告糟糕的监管比没有监管更糟。
报道全文原始报道全文
Oct 21, 2026
Demis Hassabis, Google Deepmind founder and Nobel laureate in chemistry, now believes AGI could arrive within two to three years. He defines AGI as a system matching the full cognitive range of the human brain.
That timeline keeps shrinking. Earlier this year, he said three to five years. "It's going to happen soon. My estimate keeps moving closer to the present," he told the Wall Street Journal. His litmus test is the "Einstein test": Could a system trained only on pre-1901 knowledge reproduce Einstein's 1905 papers?
Hassabis has compared AGI's impact to "10 times the industrial revolution at 10 times the speed." Since stepping down from Google DeepMind, he's focused on Google's drug discovery AI spin-off Isomorphic Labs. On safety, he recently pitched an industry-funded oversight board under government control to vet models before release. Bad regulation, he warns, is worse than none.