当现实世界成为AI可以调用的API
用Apple Watch在嘈杂机舱录音,Whisper Large V3 Turbo竟完整还原了周围所有对话。实验表明:低质量硬件+好算法已超越人类听力,我们正站在现实API可调用的门口。
Computing Life · An engineering notebook
Long-form notes on agentic systems, engineering judgment, astrophotography, hardware, coffee, and the tools that make a life easier to inspect and improve.
用Apple Watch在嘈杂机舱录音,Whisper Large V3 Turbo竟完整还原了周围所有对话。实验表明:低质量硬件+好算法已超越人类听力,我们正站在现实API可调用的门口。
An Apple Watch + Whisper Large V3 Turbo gave me superhuman hearing in a noisy airplane cabin. This experiment reveals we're at the threshold of a Real-World API—where AI can perceive and structure reality continuously.
o3从问答机器变成真正主动干活的Agent:看图推理修咖啡机、端到端计算拍卖落地成本、分析政策影响、主动优化购买决策,使用量因此暴涨。
How o3 transformed from a query-answer machine into a proactive agent that reads diagrams, calculates landed costs, analyzes policy impacts, and even optimizes purchasing decisions.
用AI拆解银行推荐的结构化票据,理解其收益结构、期权原理和隐含权衡,无需金融背景也能看透产品设计逻辑。
Use AI to demystify structured notes and understand their payoff structure, option mechanics, and trade-offs without financial expertise.
真正的问题不是AI会不会取代你,而是你能不能带好AI。三种Agent角色(教练、秘书、搭档)、默契护城河、构建AI-native的世界。
The real question isn't whether AI will replace you, but whether you can lead AI. Three agent types (Coach, Secretary, Partner), the rapport moat, and building an AI-native world.
探索AI如何处理非结构化语料,利用百万级上下文窗口让AI阅读三年微信群聊记录,并通过反思-提取-构建的循环协作方式,从中提炼出关于AI协作的深刻洞见与结构化知识。
Fed 3 years of messy WeChat history into GPT-4.1's 1M token window—it synthesized themes, generated insights, and wrote a cohesive book about AI collaboration.