展示HN:从视频中提取并运行代码

1作者: kanemarais22 天前原帖
我花了很多时间通过教程学习编程。视频的一个问题是,它将信号和噪声压缩成一个单一的流。信息密度很高,但可获取的密度却很低。 我创建了YouNote,使其更像一个数据库,除了它是多模态和互动的。你可以在一个界面中点击、提取代码、运行、修改和重新运行。它会保留你所做的一切,这样你就可以随着时间的推移建立一个有用信息的知识库。 试试看 - 无需注册: <a href="https://younote.co/kane/share/VMj-3S1tku0?token=1CIznh5eGzDNVEIM" rel="nofollow">https://younote.co/kane/share/VMj-3S1tku0?token=1CIznh5eGzDN...</a> 在任何单元格上点击“运行”并直接进行实验——它通过Pyodide在浏览器中执行,并支持NumPy、Pandas、SciPy、Matplotlib、SymPy和Statsmodels。较重的工作负载(如PyTorch)将在Colab中打开。 当前限制: 单元格执行 目前仅支持YouTube 共享工作区为只读,但功能齐全(代码可以运行) 框架(EdArXiv): <a href="https://osf.io/preprints/edarxiv/fntyu_v1" rel="nofollow">https://osf.io/preprints/edarxiv/fntyu_v1</a> 希望听到你对这个工具在学习/实验中是否有用的看法——或者你认为它可以做得更好的地方。
查看原文
I’ve spent a lot of time learning how to code from tutorials. The problem with video is that it compresses signal and noise into a single stream. The information density is high but the accessible density is low.<p>I built YouNote to make it more like a database. except it&#x27;s multimodal and interactive. You can click, extract code, run, modify and re-run in one surface. It preserves everything you do so you can build a knowledge base of useful information over time.<p>Try it out - no signup needed: <a href="https:&#x2F;&#x2F;younote.co&#x2F;kane&#x2F;share&#x2F;VMj-3S1tku0?token=1CIznh5eGzDNVEIM" rel="nofollow">https:&#x2F;&#x2F;younote.co&#x2F;kane&#x2F;share&#x2F;VMj-3S1tku0?token=1CIznh5eGzDN...</a><p>Click &#x27;run&#x27; on any cell and experiment with it directly — it executes in-browser via Pyodide and supports NumPy, Pandas, SciPy, Matplotlib, SymPy, and Statsmodels. Heavier workloads (like PyTorch) open in Colab.<p>Current limitations: Single-cell execution YouTube-native for now Shared workspace is view-only but fully-featured (code will run)<p>Framework (EdArXiv): <a href="https:&#x2F;&#x2F;osf.io&#x2F;preprints&#x2F;edarxiv&#x2F;fntyu_v1" rel="nofollow">https:&#x2F;&#x2F;osf.io&#x2F;preprints&#x2F;edarxiv&#x2F;fntyu_v1</a><p>Would love to hear if you find this useful for learning&#x2F;experimenting — or what you think it could do better.