10 Matching Annotations
  1. Last 7 days
    1. If public models can already do useful work inside that kind of workflow, then the story is not 'Anthropic has a magical cyber artifact.' The story is that serious AI-assisted vulnerability research is no longer confined to a single frontier lab.

      这一发现挑战了Anthropic试图构建的叙事:即高级AI安全研究需要受限访问。研究表明,公共模型已经能够复制关键的安全发现,这意味着真正的'护城河'不是模型访问,而是验证、优先排序和操作化的能力。这打破了'只有前沿实验室才能进行高级AI安全研究'的神话。

  2. Jun 2024
    1. this is a serious problem because all they need to do is automate AI research 00:41:53 build super intelligence and any lead that the US had would vanish the power dynamics would shift immediately

      for - AI - security risk - once automated AI research is known, bad actors can easily build superintelligence

      AI - security risk - once automated AI research is known, bad actors can easily build superintelligence - Any lead that the US had would immediately vanish.

  3. Jan 2022
  4. Oct 2021
  5. Jul 2020
  6. Jun 2020
  7. May 2020
  8. Apr 2020
  9. Nov 2015
    1. Every three years, the Librarian of Congress issues new rules on Digital Millennium Copyright Act exemptions. Acting Librarian David Mao, in an order (PDF) released Tuesday, authorized the public to tinker with software in vehicles for "good faith security research" and for "lawful modification." The decision comes in the wake of the Volkswagen scandal, in which the German automaker baked bogus code into its software that enabled the automaker's diesel vehicles to reduce pollutants below acceptable levels during emissions tests.