9 Matching Annotations
  1. Apr 2026
    1. Because these benchmarks are human-authored, they can only test for risks we have already conceptualized and learned to measure.

      这句话揭示了当前 AI 安全评测体系的致命盲区:所有 benchmark 都是人类提前想好的问题,而真正危险的「未知的未知」(unknown unknowns)根本无法被预设题目捕捉。这意味着我们现有的模型安全认证,本质上是一场对已知风险的自我测试。

  2. Oct 2023
    1. they're not wrong, but they don't teach what you don't know you don't know, whereas the one I link to makes this critical unknown unknown become a known unknown and then a known known. I didn't know you had a 1) local branch, 2) locally-stored remote-tracking branch, and 3) remote branch until I read that answer. Prior to that I thought there was only a local branch and remote branch. The locally-stored remote-tracking branch was an unknown unknown. Making it go from that to a known known is what makes that answer the best.
  3. Jan 2023
    1. the entanglement of human health and global climate change is rapidly manifesting itself in a raft of “known unknowns”—future disruptions we can reasonably forecast without yet knowing the precise form they will take or the extent of their consequences. When extended for a prolonged period of social time, such certain uncertainty breeds fear and a narrowing of intellectual and emotional attention.

      !- Known unknowns : interesting perspective

  4. Aug 2022
  5. Apr 2021
  6. Oct 2020
  7. Feb 2017
    1. The project plan is at best an educated guess. An educated guess that can be improved upon by capitalizing on the expertise of others that have been previously involved in projects of this nature. And yet there are many projects for which there are exists no similar requisite expertise. As such, the project plan remains very much simply ones conscientious best guess at what needs to be done.

      I believe here is where fear creeps in, since nobody in the team has encountered a situation with that problem set before. They literally have nothing to fall back on, no past knowledge or experience, they have to create knowledge.