A learning system can continuously incorporate real-world data in a way that numerical solvers fundamentally cannot, capturing and compounding the knowledge that is currently trapped out there in the real world.
揭示了AI驱动设计的另一大优势：打通仿真与现实的闭环。传统求解器难以穷尽制造公差等现实复杂因素，而学习系统能持续吸收实测数据，形成越用越聪明的“数据飞轮”。将现实中散落的隐性知识固化为模型能力，这是传统工具无法企及的质变。
βテスト期間中のご利用は無料です。
Beta 期间完全免费——对于一个声称能替代 CSO 团队数周工作的产品来说，这个策略令人惊讶。背后的逻辑是：Sakana 需要真实的企业级研究任务作为训练数据和案例积累，而这些数据只有企业用户才能提供。「用免费换真实场景数据」是 AI 产品冷启动的经典策略，但在如此高端的 B2B 定位下使用，意味着 Sakana 对自己产品当前状态的坦诚：它还不够好到让企业为初版买单，但已经足够好到值得企业免费试用。
frontier AI companies can run more of the best AIs to speed up their own AI research, relative to their competitors. Right now these gains are maybe noticeable but not game-changing, but that'll probably change in the next few years.
这是整篇文章埋下的最深的炸弹：当顶尖 AI 公司开始用 AI 加速自身的 AI 研究，算力优势将产生复利效应——算力领先 → AI 研究更快 → 更好的模型 → 更快的研究 → 更大的算力领先。这个「飞轮」一旦转起来，计算差距将不再是线性的，而是指数级加速扩大。对所有「追赶者」而言，这是一个潜在的「逃逸临界点」。
@chrisaldrich, I appreciate your feedback. Indeed there is magic in making notes which comes not only from finding connections in the ZK but also from making connections in mind. Maybe I'm confused. A mindset that makes note-making fun is one way to recruit the body's dopamine mechanism. This creates a positive feedback loop. More mote-making turns to more dopamine which turns to more note-making. Maybe even some notes on dopamine. (I have 11 already!) My sense of Luhmann's phrase "second memory" is a rehashing of an idea—a continued exploration. Using the ZK method is one way of formalizing the continued review of ideas. Without a formal process, it is too easy to fall into old bad habits and not work towards "the serendipity of combinatorial creativity. "
Reply to Will Simpson at https://forum.zettelkasten.de/discussion/comment/17939/#Comment_17939
There should be more conversation about zettelkasten as both a "ratchet" as well as a "flywheeel". Sometimes I feel like it's hard to speak of these things for either lack of appropriate words/naming and/or having a shared vocabulary for them.
Even Luhmann's "second memory" has a mushiness to it, but I certainly see your sense of it as a thing which moves forward. I have the same sort of sense with the Aboriginal cultural idea of a "songline" which acts as both a noun as well as having an internal sense of being a verb to me. The word "google" has physically and specifically undergone the transition from noun to verb in a way which "second memory" and "songline" haven't, though perhaps they should? The difference is that the word google is much more concrete and simple while second memory and songline have a lot more cultural material and meaning sitting with them if you know them and their fuller attendant practices.
Expansion is led by focus. By taking time to edit, carve up, and refactor our notes, we put focus on ideas. This starts the Great Wheel of Positive Feedback. All hail to the Great Wheel of Positive Feedback.
How can we better thing of card indexes as positive feedback mechanisms? Will describes it as the "Great Wheel of Positive Feedback" which reminds me a bit of flywheels for storing energy for later use.