5 Matching Annotations
  1. Last 7 days
    1. The future we seek requires materials no one yet knows how to make. Our labs are learning how.

      【非共识】作者提出了一个大胆的前瞻性声明,暗示他们的实验室正在学习制造目前未知材料的方法。这一观点挑战了传统材料科学的研究边界,暗示AI可能发现人类科学家尚未考虑的材料合成路径。然而,这一声明缺乏具体案例或证据支持,更像是一种愿景而非已实现的技术突破。

  2. Jun 2026
    1. That was the moment that I felt like, okay, these models have now come to a point where they really, truly understand.

      大多数人认为AI模型只是基于模式识别的统计工具,无法真正'理解'科学概念。然而,作者声称GPT-5能够预测未发表实验的结果,并产生'真正理解'的洞察力,这挑战了人们对AI本质和认知能力的传统认知,暗示AI可能已达到某种形式的理解能力。

  3. Apr 2026
    1. Discovery should focus on trust boundaries, authentication flows, parsers, shared services, and legacy code that still sits on critical paths.

      这一建议挑战了传统安全扫描的广度优先方法,转而强调深度优先的特定领域。这表明AI安全研究应该更关注那些传统方法难以发现的复杂逻辑问题,而不是简单地扫描所有代码。这种转变可能带来更有效的安全投资回报。

    1. Add llms.txt metadata and root/package LICENSE files - Add website llms.txt support and move LICENSE to root - Fix llms.txt serving and restore package LICENSE

      令人惊讶的是:这个项目支持llms.txt元数据格式,这是一种新兴的AI可发现性标准,使AI模型能够更好地理解项目文档和代码结构。这种关注AI可发现性的做法表明项目开发者不仅关注当前功能,还前瞻性地考虑了AI与代码库的交互方式。

  4. Oct 2024
    1. https://web.archive.org/web/20241007071434/https://www.dbreunig.com/2024/10/03/we-need-help-with-discovery-more-than-generation.html

      Author says generation isn't a problem to solve for AI, there's enough 'content' as it is. Posits discovery as a bigger problem to solve. The issue there is, that's way more personal and less suited for VC funded efforts to create a generic tool that they can scale from the center. Discovery is not a thing, it's an individual act. It requires local stuff, tuned to my interests, networks etc. Curation is a personal thing, providing intent to discovery. Same why [[Algemene event discovery is moeilijk 20150926120836]], as [[Event discovery is sociale onderhandeling 20150926120120]] Still it's doable, but more agent like than central tool.