Jinspire
Writing about curiosity, insights, and reflection.
Artifacts →2025年年终总结
1 今年真的是无比精彩的一年。精彩意味着,跌宕起伏。意味着无数的挫折,遗憾,不甘,痛苦,和挣扎。”ample doses of pain.” 然后便是成长。先总结,然后慢慢道来。
Summary of the Year 2025
1 This year has truly been an incredibly wonderful year. Wonderful means ups and downs. It means countless setbacks, regrets, unwillingness, pain, and strugg...
顶级AI领袖 - 能力雷达
这是Gemini 3如何感知顶级AI领袖在不同维度上的能力的视觉化。我特意使用了1-10的完整量表,以避免聚集在中位数周围。
Top AI Leaders - Ability Radar
Here is a visualization of how Gemini 3 perceives the abilities of top AI leaders across different dimensions. I intentionally made use of the full scale of ...
AI仍然无法读取Google Scholar的数字
祝贺 Yoshua Bengio 达到 100万引用! 我一直想知道谁拥有 最多的引用,以及 前10名。所以我用AI帮我寻找答案。 令人惊讶的是,AI搜索甚至深度研究都没有找到正确的结果,并且产生了很多幻觉 —— 尽管这真的 很简单的 验证。 分数: GPT 5 + Web...
AI still can't read Google Scholar's numbers
Congratulations to Yoshua Bengio for hitting 1 million citations! I’m always wondering who has the most citations, also the top 10. So I used AI to help...
LLM训练对我自身学习过程有哪些实际启示?
我最近更多地了解了LLM是如何训练的,我开始注意到与 我自己的学习过程 的相似之处。
What are some practical implications of LLM training for my own learning process?
I recently learned more about how LLMs are trained, and I began to notice parallels with my own learning process.
从零开始花100小时构建嵌入搜索引擎,我学到了什么
我花了100个小时从零开始构建了一个向量(嵌入)搜索引擎,结果我学到了更多。我想分享我学到的东西,以便在你需要经历类似旅程时为你节省一些时间。我将我的理解总结为 简史,然后是 基本和基础概念,然后分享我学到的 最佳工程建议和实践。
What I learned from using 100 hours to build an embedding search engine from ground up
I spent 100 hours to build an vector (embedding) search engine from ground up, and I ended up learning much more. I want to share what I learned, to save you...
半年用AI写了一百万行代码,我学到了什么?
今天标志着我的代码行数达到了100万。
What I learned from using AI to write ONE MILLION lines of code in half a year?
Today marks my lines of code count to 1 million.
创业日记 - 持续更新
入职?周了(Sep. 28 2025)
Startup Diary - Continuously Updated
? weeks into the job (Sep. 28 2025)
为什么说世界的本质是DP?
为什么说这个世界的本质就是动态规划(Dynamic Programming,DP)?
Why the Essence of the World is DP?
Why is the essence of this world Dynamic Programming (DP)?
机器学习资源索引
主要目的是列出我收集的所有关于机器学习 (ML)、人工智能 (AI) 和大型语言模型 (LLM) 的资源。
ML Resources Index
The main purpose is to list all the resources I’ve collected about Machine Learning (ML), Artificial Intelligence (AI), and Large Language Models (LLM).
图像生成101:入门指南
简介
Image Generation 101: an Introduction
The Introduction
Gemini 2.5 绝对令人惊叹
去年,我花了无数个夜晚与GPT-4o交谈,深入探讨技术、传记、历史等。就像旧时光一样,查阅维基百科,点击链接,然后再点击更多链接。上周,我又开始用Gemini 2.5做这件事,一直学到凌晨5点,完全没意识到时间过得有多快。它只是给出了更深思熟虑的结果。我从未想过取消我的ChatGPT订阅,但这确实是我第一次认真考...
Gemini 2.5 is strictly amazing
Last year, I spent nights talking with GPT-4o, diving deep into tech, biographies, history, etc. Just like old times, checking wikipedia and click into links...
并行计算101:LLM并行性入门
问题
Parallelism101: Introduction to LLM Parallelism
The Problem
用Deepseek读《物演通论》
最纯粹的哲学,最终极的追问。
Reading 'The Evolution of Things' with DeepSeek
The purest philosophy, the ultimate inquiry.
可视化LLM生成过程
法国的首都是___
Visualize Generation in LLM
The capital of France is ___
人工智能, 创业, 与反思
最近的生活
AI, Startup, and Reflections
Recent Life