Dawson Church

@TomMariaRicci @karpathy The shift isn’t that LLMs extract tacit knowledge, but that they force us to structure it. What was implicit becomes explicit through interaction.

LLM Knowledge Bases Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So: Data ingest: I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them. IDE: I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides). Q&A: Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale. Output: Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base. Linting: I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into. Extra tools: I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries. Further explorations: As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows. TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

人在抑郁痛苦的时候，会忍不住产生黑暗想法。 例如：如果我死了，谁都不会记得我。他们只会嫌我死了给他们造成了麻烦，这很不负责任，然后很快就该干嘛干嘛.... 确实，大部分人死后，看起来似乎很快就“被所有人忘了”。 但实际上人生并非如此，可以不必如此。和 gpt-4.5 聊了一下这个想法，感到治愈。 （最近某歌星去世，在时间线看到，想起这个之前想过的问题。又想起推上著名程序员去世的事情）

@zls0910 dajidali

您的磨难已结束，不会回来了！如果你刷到金佛，留下一句：“大吉大利”！折磨你的事彻底消失，好运连连到。

@suiaca1 谢谢

二月份可以按照五日均线持股，所有的牛股都不会轻易跌破5日均线。跌破了就不是牛股了。

Explore the advanced AI model ChatGPT, Claude & Gemini for FREE! sider.ai/invited?c=9fe2…

Do you remember when you joined X? I do! #MyXAnniversary

治愈四千万人的“正念呼吸法”，高效减压，缓解情绪，促进催眠。睡眠不好的试一试！

On #ChineseNewYear's Eve, let's start the countdown together as the new overtakes the old amid the crackle of firecrackers. #ChineseWisdom #SpringFestival #XiQuotes

Free Python ebook for statistics:

Peaches and plums do not speak, but they are so attractive that a path is formed below the trees. #ChineseWisdom #XiQuotes 桃李不言，下自成蹊。 From the keynote speech by President Xi Jinping at the opening ceremony of the first Belt and Road Forum for International Cooperation, on May 14, 2017. #XiJinping chinadaily.com.cn/a/201705/14/WS…

Excited to try out ComfyUI V1 desktop app! Get on the waitlist here: comfy.org/download

深入浅出白话解读Transformer架构！ 什么是 Transformer？ - 定义：Transformer 是一种神经网络，特别擅长理解和生成序列数据。比如，它能理解一句话的上下文，并基于此生成新的句子。 - 特点：它完全依赖自注意力机制，不使用循环神经网络（RNN）或卷积网络（CNN）。 Transformer 就像一个黑箱子 - 输入：你给它一句话，比如西班牙语的“¿De quién es?”。 - 输出：它会翻译成英语“Whose is it?”。 - 黑箱子内部：看起来像是魔法，但其实是复杂的计算过程。 编码器/解码器架构 - 编码器 (Encoder)： - 把输入句子转换成矩阵形式，抓住句子的本质信息。 - 例如，把“¿De quién es?”处理成结构化数据。 - 解码器 (Decoder)： - 接收编码后的数据，逐步生成输出句子。 - 最终把编码的数据翻译成英语句子“Whose is it?”。 Transformer 的内部架构 - 编码器 (Encoder)： - 每个编码器层依次处理输入数据，每一层都包括自注意力机制和前馈神经网络。 - 数据从一个层传递到下一个层，逐渐提取更深层次的信息。 - 解码器 (Decoder)： - 每个解码器层也有自注意力机制，但它们还需要处理来自编码器的数据。 - 解码器逐层生成最终的输出。 假设你在国际会议上，需要实时翻译演讲者的话语。使用Transformer架构的翻译系统，可以快速准确地将演讲内容从一种语言翻译成另一种语言。通过编码器抓取说话内容的核心信息，再由解码器生成目标语言的翻译，使得跨语言交流变得流畅无阻。这正是Transformer在实际应用中的一个典型场景。

麻省理工人工智能实验室 @MIT_CSAIL 强烈推荐的 130 页机器学习备忘录清单！详细了解 #MachineLearning 核心概念 🔥 目前已开源，6.7K Star ⭐️ 非常全面！包括概率、生成式模型、高斯模型、贝叶斯统计、线性回归、逻辑回归、EM 算法、核、蒙特卡洛推理、深度学习等 github.com/soulmachine/ma…

@tuturetom 好

麻省理工这本爆火的《深入理解深度学习》的免费书终于出中文版了🔥 深入讲解了深度学习核心概念，每章节搭配 PPT 和 Python 练习代码！ 内容包括监督学习、神经网络、卷积网络、Transformers、扩散模型、强化学习等。 原书地址：udlbook.github.io/udlbook/ 中文版👉🏻：github.com/careywyr/Under…

KAN: Kolmogorov–Arnold Networks is a really elegant paper. It's refreshing to see that things worked by design rather than by accident in ML. I made a small KAN experiment comparing it to MLP and spline fitting. (notebook in comments!)

当然 Reor 支持 ollama.com/library 中任何本地模型，只需要输入模型名称即可下载使用！

这样批卷子成本是降低了。

SQL Query execution order morioh.com/p/f5733a5d2fa9… #sql #nosql #mysql #database #python #programming #developer #morioh #programmer #coding #softwaredeveloper #computerscience #webdev #webdeveloper #webdevelopment

Prompt: A movie trailer featuring the adventures of the 30 year old space man wearing a red wool knitted motorcycle helmet, blue sky, salt desert, cinematic style, shot on 35mm film, vivid colors.

Introducing Sora, our text-to-video model. Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. openai.com/sora Prompt: “Beautiful, snowy Tokyo city is bustling. The camera moves through the bustling city street, following several people enjoying the beautiful snowy weather and shopping at nearby stalls. Gorgeous sakura petals are flying through the wind along with snowflakes.”