# Zara's AI learning library
## Hero
Zara's AI learning library
a collection of the best resources for learning AI from the Internet, hand-picked and curated by Zara Zhang.
all resources are:
- free to access
- friendly to non-technical people
## YouTube Videos
Featured Videos
I have personally watched all of these videos, from beginning to end. I picked these because I learned a lot from them.
I watch a LOT of YouTube videos on AI. I've filtered the best for you.
I recommend watching them using LongCut, a tool that I built with friends to help you learn from long YouTube videos better and faster.
View Complete Collection
Complete AI Video Collection
I haver personally watched all of these videos, from beginning to end. I picked these because I learned a lot from them.
I watch a LOT of YouTube videos on AI. I've filtered the best for you.
I recommend watching them using LongCut, a tool that I built with friends to help you learn from long YouTube videos better and faster.
Collapse Collection
Filter by Tag
All Videos
| Tag Name | Is Default (All) |
| --- | --- |
| All | Yes |
| Fundamentals | No |
| Product | No |
| Founder interview | No |
| Practical tutorial | No |
| Prompting | No |
| Vibe coding | No |
| Marketing and growth | No |
No videos found with this tag.
Watch on LongCut
TLDW link not available - opens YouTube instead
https://www.youtube.com/embed/
YouTube video player for
| Video Title | Description | Channel | Tags (comma-separated) | Video ID | Duration | YouTube URL | TLDW URL | Feature This Video |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| **Deep Dive into LLMs like ChatGPT** | The best introduction into LLMs on the Internet | Andrej Karpathy | Fundamentals | 7xTGNNLPyMI | 3:31:23 | https://www.youtube.com/watch?v=7xTGNNLPyMI | https://tldw.us/analyze/7xTGNNLPyMI?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D7xTGNNLPyMI | Yes |
| **How to Build a Beloved AI Product - Granola CEO Chris Pedregal** | Learn from the founder of Granola, the AI PM that I admire the most | The MAD Podcast with Matt Turck | Product, Founder interview | IcbuTTVUY7M | 1:08:35 | https://www.youtube.com/watch?v=IcbuTTVUY7M | https://tldw.us/analyze/IcbuTTVUY7M?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DIcbuTTVUY7M | Yes |
| **Inside NotebookLM with Raiza Martin and Steven Johnson** | Behind the scenes look at Google's NotebookLM, one of my favorite AI products | Google | Product | mccQdu5afZw | 46:11 | https://www.youtube.com/watch?v=mccQdu5afZw | https://tldw.us/analyze/mccQdu5afZw?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DmccQdu5afZw | Yes |
| **Inside Claude Code: How an AI Native Team Actually Works \| Cat Wu** | Behind the scenes at Anthropic's Claude development team | Peter Yang | Product, Vibe coding | jmHBMtpR36M | 40:36 | https://www.youtube.com/watch?v=jmHBMtpR36M | https://tldw.us/analyze/jmHBMtpR36M?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DjmHBMtpR36M | No |
| **The Master Prompt Method: Unlock AI's Full Potential** | A serial entrepreneur shares a powerful way to train your AI with full context about you and your business | Tiago Forte | Demo, Prompting, Practical tutorial | _K_F_icxtrI | 38:53 | https://www.youtube.com/watch?v=_K_F_icxtrI | https://tldw.us/analyze/_K_F_icxtrI?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D_K_F_icxtrI | No |
| **Going Direct \| Lulu Cheng Meservey** | Lulu Cheng Meservey is the leading voice in the new age of PR and comms. Learn how to make your AI product stand out in a sea of slop. Hosted by Sam Altman's brother | Uncapped with Jack Altman | Marketing and growth | GRoU1T4E9rQ | 50:18 | https://www.youtube.com/watch?v=GRoU1T4E9rQ | https://tldw.us/analyze/GRoU1T4E9rQ?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DGRoU1T4E9rQ | No |
| Screensharing How This Guy Makes Millions with AI Agents | AI workflows from Rowan Cheung who is behind one of the top AI newsletters "The Rundown" | Greg Isenberg | Demo, Practical tutorial | 4aOeju76me4 | 43:06 | https://www.youtube.com/watch?v=4aOeju76me4 | https://tldw.us/analyze/4aOeju76me4?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D4aOeju76me4 | No |
| **Snipd: The AI Podcast App for Learning — with CEO Kevin Ben-Smith** | Interview with the founder of Snipd, the AI podcast app that I use every day | Latent Space | Product, Founder interview | FNRO_SYx68Q | 1:17:45 | https://www.youtube.com/watch?v=FNRO_SYx68Q | https://tldw.us/analyze/FNRO_SYx68Q?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DFNRO_SYx68Q | No |
| **Mark Zuckerberg on AI Glasses, Superintelligence, Neural Control, and More** | Zuck discusses Meta's latest AI glasses | Rowan Cheung | Founder interview | WuTJkFvw70o | 45:28 | https://www.youtube.com/watch?v=WuTJkFvw70o | https://tldw.us/analyze/WuTJkFvw70o?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DWuTJkFvw70o | No |
| **Prompting 101 \| Code w/ Claude** | Hands-on guidance on prompting from the Anthropic team | Anthropic | Prompting, Product, Practical tutorial | ysPbXH0LpIE | 24:52 | https://www.youtube.com/watch?v=ysPbXH0LpIE | https://tldw.us/analyze/ysPbXH0LpIE?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DysPbXH0LpIE | Yes |
| **The 19-year-old Making Millions with His Ai Mobile App** | Eye-opening insights on growth and making a consumer app in the AI age | Brett Malinowski | Product, Marketing and growth | AY4R2p2i8nM | 1:03:42 | https://www.youtube.com/watch?v=AY4R2p2i8nM | https://tldw.us/analyze/AY4R2p2i8nM?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DAY4R2p2i8nM | No |
| **The AI-native startup: 5 products, 7-figure revenue, 100% AI-written code. \| Dan Shipper (Every)** | Dan Shipper is the co-founder and CEO of Every | Lenny's Podcast | Founder interview | crMrVozp_h8 | 1:34:56 | https://www.youtube.com/watch?v=crMrVozp_h8 | https://tldw.us/analyze/crMrVozp_h8?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DcrMrVozp_h8 | No |
| **Andrew Ng: Building Faster with AI** | Andrew Ng's talk at YC reshaped how I think about product development in the AI era | Y Combinator | Product, Vibe coding | RNJCfif1dPY | 43:57 | https://www.youtube.com/watch?v=RNJCfif1dPY | https://tldw.us/analyze/RNJCfif1dPY?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DRNJCfif1dPY | Yes |
| **Inside ChatGPT: The fastest growing product in history \| Nick Turley (OpenAI)** | Insider stories behind ChatGPT's growth | Lenny's Podcast | Product, Marketing and growth | ixY2PvQJ0To | 1:35:37 | https://www.youtube.com/watch?v=ixY2PvQJ0To | https://tldw.us/analyze/ixY2PvQJ0To?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DixY2PvQJ0To | Yes |
| **Complete Beginner's Course on AI Evaluations in 50 Minutes (2025) \| Aman Khan** | The best free resource on AI evaluations, for complete beginners | Peter Yang | Practical tutorial, Product | TL527yTpxlk | 51:47 | https://www.youtube.com/watch?v=TL527yTpxlk | https://tldw.us/analyze/TL527yTpxlk?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DTL527yTpxlk | No |
| **Chris Pedregal + Sam Stephenson: Making Meetings More Effective with Granola** | Masterclass on building an AI product in the application layer | Lightspeed Venture Partners | Product, Founder interview | 2eajeT9WU4k | 38:05 | https://www.youtube.com/watch?v=2eajeT9WU4k | https://tldw.us/analyze/2eajeT9WU4k?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D2eajeT9WU4k | No |
| **Cursor Team: Future of Programming with AI \| Lex Fridman Podcast** | Behind-the-scenes with the Cursor team; good introduction to AI coding | Lex Fridman | Vibe coding, Founder interview | oFfVt3S51T4 | 2:29:04 | https://www.youtube.com/watch?v=oFfVt3S51T4 | https://tldw.us/analyze/oFfVt3S51T4?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DoFfVt3S51T4 | No |
| **How To 10x Your Notes: Obsidian + Claude AI Agents** | How to connect Obsidian with Claude Code; easy to follow for non-technical people | Mckay Wrigley | Practical tutorial, Vibe coding | d7Pb73dbcIM | 1:01:26 | https://www.youtube.com/watch?v=d7Pb73dbcIM | https://tldw.us/analyze/d7Pb73dbcIM?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Dd7Pb73dbcIM | No |
| **Arc Had Millions of Users. Why They Left It Behind for Dia.** | Behind-the-scenes stories of Dia | Every | Founder interview | 210zavw00y4 | 1:24:52 | https://www.youtube.com/watch?v=210zavw00y4 | https://tldw.us/analyze/210zavw00y4?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D210zavw00y4 | No |
| **OpenAI's CPO on how AI changes must-have skills, moats, coding, startup playbooks, more \| Kevin Weil** | Interview with OpenAI's head of product | Lenny's Podcast | Product | scsW6_2SPC4 | 1:31:40 | https://www.youtube.com/watch?v=scsW6_2SPC4 | https://tldw.us/analyze/scsW6_2SPC4?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DscsW6_2SPC4 | No |
| **View From The Top with Aravind Srinivas, Cofounder and CEO of Perplexity** | Aravind offers advice to prospective founders, and dives into what makes Perplexity unique. | Stanford Graduate School of Business | Founder interview | r1Bi10Xt0fc | 52:09 | https://www.youtube.com/watch?v=r1Bi10Xt0fc | https://tldw.us/analyze/r1Bi10Xt0fc?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Dr1Bi10Xt0fc | No |
| **Andrej Karpathy: Software Is Changing (Again)** | Insights on software development by the inventor of the term "vibe coding" | Y Combinator | Vibe coding | LCEmiRjPEtQ | 39:31 | https://www.youtube.com/watch?v=LCEmiRjPEtQ | https://tldw.us/analyze/LCEmiRjPEtQ?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DLCEmiRjPEtQ | No |
| **NotebookLM with Steven Johnson and Raiza Martin** | Behind the scenes of NotebookLM's development | Google for Developers | Product | 7w_ARTj-mvM | 1:01:48 | https://www.youtube.com/watch?v=7w_ARTj-mvM | https://tldw.us/analyze/7w_ARTj-mvM?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D7w_ARTj-mvM | No |
| **How I use LLMs** | Beginner-friendly guide to using LLMs | Andrej Karpathy | Fundamentals, Practical tutorial | EWvNQjAaOHw | 2:11:11 | https://www.youtube.com/watch?v=EWvNQjAaOHw | https://tldw.us/analyze/EWvNQjAaOHw?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DEWvNQjAaOHw | No |
| **Alexandr Wang, Scale AI, & the Startup Hunger Games** | Memorable interview with Alexandr Wang (when he was still CEO of Scale) | South Park Commons | Founder interview | CEebe451GZs | 55:35 | https://www.youtube.com/watch?v=CEebe451GZs | https://tldw.us/analyze/CEebe451GZs?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DCEebe451GZs | No |
| **Personalized AI Language Education — with Andrew Hsu, Speak** | AI for language learning | Latent Space | Founder interview | tIVKgztDaYQ | 1:04:09 | https://www.youtube.com/watch?v=tIVKgztDaYQ | https://tldw.us/analyze/tIVKgztDaYQ?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DtIVKgztDaYQ | No |
| **We Were Going to Shut Down Our Company…Now We're Worth $700M \| Eric Simons (Bolt)** | Interview with the founder of Bolt; a lesson on resilience | Peter Yang | Founder interview, Vibe coding | moaX7CsqZjc | 48:22 | https://www.youtube.com/watch?v=moaX7CsqZjc | https://tldw.us/analyze/moaX7CsqZjc?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DmoaX7CsqZjc | No |
## Podcasts
Video Podcasts
For those with **👁 Watch** tags, I recommend watching the video versions (instead of just listening to the audio), because many of these feature visual information like product demos/screen shares
For proper learning, consume them from start to finish. Don't just generate a summary with AI 👀
I include the YouTube links because I prefer the video versions. For audio versions, type the show names into any podcast apps to listen
| Podcast | Description | Podcast URL | Watching Recommended |
| --- | --- | --- | --- |
| **Latent Space** | The podcast by and for AI Engineers | https://www.youtube.com/playlist?list=PLWEAb1SXhjlfkEF_PxzYHonU_v5LPMI8L | Yes |
| **AI & I** | Learn how the smartest people in the world are using AI to think, create, and relate | https://www.youtube.com/playlist?list=PLuMcoKK9mKgHtW_o9h5sGO2vXrffKHwJL | Yes |
| **Google DeepMind: The Podcast** | Interview with Google's AI leaders, by the wonderful host Hannah Fry | https://www.youtube.com/playlist?list=PLqYmG7hTraZBiUr6_Qf8YTS2Oqy3OGZEj | No |
| **The AI Daily Brief** | AI news with an opinion; saves Twitter scrolls | https://www.youtube.com/playlist?list=PLRYSuzHGhXPmKnOpd-f588cNNmTe2S9FP | Yes |
| **TBPN** | Technology's daily show | https://www.youtube.com/@TBPNLive | No |
| Training Data | Sequoia Capital partners host conversations with leading AI builders and researchers | https://www.youtube.com/playlist?list=PLOhHNjZItNnMm5tdW61JpnyxeYH5NDDx8 | No |
| Lenny's Podcast | Interviews with world-class product leaders and growth experts | https://www.youtube.com/@LennysPodcast | No |
| Behind the Craft | Interviews with top product leaders and creators to help you craft great products and creator businesses | https://www.youtube.com/playlist?list=PLIWHjbvRtljj4RewVNv_znkUe-3E-NKd2 | Yes |
| No Priors | Elad Gil and Sarah Guo talk to the world's leading engineers, researchers and founders | https://www.youtube.com/@NoPriorsPodcast | No |
| Unsupervised Learning | Repoint partners interview the sharpest minds in AI | https://www.youtube.com/@RedpointAI | No |
| Minus One | Podcast by South Park Commons | https://www.youtube.com/playlist?list=PLmYVYFmFwGm3txxUduawn7i53C5rDjjd7 | No |
| Lightcone Podcast | YC's podcast | https://www.youtube.com/playlist?list=PLQ-uHSnFig5Ob4XXhgSK26Smb4oRhzFmK | No |
## People to Follow
People to Follow on X
My principle: Follow builders, not influencers.
Most of the people below are actually building AI products/companies hands-on.
| Person | Description | Profile Image | X Profile URL |
| --- | --- | --- | --- |
| **Andrej Karpathy** | Best AI educator |  | https://x.com/karpathy |
| **Swyx** | Leading voice in AI Engineers community | | https://x.com/swyx |
| **Greg Isenberg** | Get startup ideas & daily motivation for building a company with AI |  | https://x.com/gregisenberg |
| **Lenny Rachitsky** | Product, growth, and career advice |  | https://x.com/lennysan |
| **Josh Woordward** | VP of Gemini and Google Labs, Google |  | https://x.com/joshwoodward |
| **Kevin Weil** | Head of product @ OpenAI |  | https://x.com/kevinweil |
| **Peter Yang** | Tactics and humor to help you save time with AI and craft better products |  | https://x.com/petergyang |
| **Nan Yu** | Head of product @ Linear |  | https://x.com/thenanyu |
| **Madhu Guru** | Product Leader at Google Gemini |  | https://x.com/realmadhuguru |
| **Mckay Wrigley** | Learn vibe-coding |  | https://x.com/mckaywrigley |
| **Steven Johnson** | Author; soul behind NotebookLM |  | https://x.com/stevenbjohnson |
| **Amanda Askell** | Guardian of Claude's vibes; expert on AI personality/character |  | https://x.com/AmandaAskell |
| **Boris Cherny** | Creator of Claude Code | | https://x.com/bcherny |
| **Cat Wu** | Claude Code PM at Anthropic |  | https://x.com/_catwu |
| **Thariq** | Claude Code team at Anthropic |  | https://x.com/trq212 |
| **Google Labs** | Google's home for our latest AI tools and experiments. |  | https://x.com/GoogleLabs |
| **George Mack** | How to be high-agency |  | https://x.com/george__mack |
| **Raiza Martin** | Cofounder of Huxe; previous PM of NotebookLM |  | https://x.com/raizamrtn |
| **Amjad Masad** | CEO of Replit |  | https://x.com/amasad |
| **Guillermo Rauch** | CEO of Vercel | | https://x.com/rauchg |
| **Riley Brown** | Learn vibe coding |  | https://x.com/rileybrown |
| **Alex Albert** | Claude relations @ Anthropic | | https://x.com/alexalbert__ |
| **Hamel Husain** | Evals |  | https://x.com/HamelHusain |
| **Aaron Levie** | CEO of Box | | https://x.com/levie |
| **Ryo Lu** | Head of design @ Cursor | | https://x.com/ryolu_?s=21&t=6BfggVg3yI1peSEUo566_g |
| **Garry Tan** | President of YC | | https://x.com/garrytan |
| **Lulu Cheng Meservey** | Top voice in comms/go-to-market | | https://x.com/lulumeservey |
| **Justine Moore** | Investor @ a16z; trends in consumer AI & AI videos | | https://x.com/venturetwins |
| Matt Turck | VC @ FirstMark | | https://x.com/mattturck |
| **Julie Zhuo** | Founder of Sundial | | https://x.com/joulee |
| **PJ Ace** | Viral AI ad madman | | https://x.com/PJaccetturo |
| **Zara Zhang** | Me 👀 | | https://x.com/zarazhangrui |
## Podcasts & Articles
Newsletters
I read all of these regularly.
This is how I find out about cool new AI products, model releases, and thoughtful essays.
| Newsletter | Description | Newsletter URL |
| --- | --- | --- |
| **AI Valley** | Daily recap on the latest happenings in AI | https://www.theaivalley.com/ |
| **Every** | Best writings on AI and what it means to live in an AI age | https://every.to/ |
| **The Keyword** | Latest news on Google's AI efforts | https://blog.google/ |
| **Ben's Bites** | Great resources for non-technical people to learn to build with AI | https://bensbites.beehiiv.com/ |
| **AINews by smol.ai** | Best newsletter for AI engineers. Summarizes latest happenings in model development/AI coding | https://news.smol.ai/ |
| **Peter Yang** | Level up your AI skills. Recommend for PMs | https://creatoreconomy.so/ |
## Products I like
Products I like
AI products that I personally use and recommend
| Product Name | Description | Product URL | Logo |
| --- | --- | --- | --- |
| **NotebookLM** | Remix anything | https://notebooklm.google.com |  |
| Huxe | Startup by the former team behind NotebookLM; personal audio & content for one | https://www.huxe.com/ |  |
| **Granola** | AI meeting notetaker with innovative UX | https://granola.so |  |
| **Snipd** | AI-powered podcast app that creates highlights and transcripts automatically | https://snipd.com |  |
| **Comet (AI browser by Perplexity)** | My default browser | https://www.perplexity.ai/comet |  |
| Tolan | Your AI best friend | https://www.tolans.com/ |  |
| Poke | AI assistant that is personalized and proactive, with a crazy onboarding experience | https://poke.com/ |  |
| Faces | The tool I used to build this website! | https://faces.app/ |  |
| TLDW | A tool that I made with friends. Helps you learn efficiently from long YouTube videos | https://tldw.us/ |  |
## Prompts
Prompts
EN
中文
| Prompt Name | Product URL | Prompt Content (English) | Prompt Content (Chinese) |
| --- | --- | --- | --- |
| AI dictionary | https://x.com/zarazhangrui/status/1991047223715770803?s=20 | Build an AI dictionary with these features.
User first select their native language and target language (for each, options include the 10 most popular languages in the world).
User can input with text (word, phrase, sentence).
The result displays a natural language definition in the user's native language, as well as an AI-generated image visualizing the concept. It also includes 2 example sentences, each followed by a translation in the user's native language. It also includes an usage note, which explains the cultural nuance, how it's used, tone/voice, or related words (synonyms/words that look similar and are often confused) and how they differ in usage. The explanation should be fun, lively, and in casual language, like you're talking to a friend. Don't sound like a textbook. Be concise. No greetings or fillers; get to the point immediately.
For each word or phrase defined, provide a button that plays a clear audio recording of its pronunciation in the target language. Make sure it sounds natural, not robotic. All example sentences should also have a pronunciation feature. Make sure the audio latency is low.
The user can choose to save any result page to their "Notebook" for review later.
In the Notebook, there's a feature where the AI can "make up a story" using the words/phrases contained in the notebook, in order to help the user memorize them.
There's a "Study" mode, where the AI generates flashcards for words in the notebook, where the front is the word in the target language against the image that represents the word, and the back is the word and its definition in native language (along with example sentence). When they click on the card, there should be a flipping animation.
Make it mobile-optimized. Overall vibe is bright and fun. Make it pop.
| 构建一款具备以下能力的 AI 词典：
用户进入后，先选择自己的母语和目标语言（两者都提供全球最常用的 10 种语言）。
用户可以输入任意文字，包括单词、短语或整句。
结果页将展示：
• 用用户母语写出的自然语言解释
• 两个例句，并分别附上对应的母语翻译
• 一段轻松、有趣、像朋友聊天一样的用法说明，内容涵盖文化语境、使用场景、语气、相关词（同义词或外形相似但常被混淆的词）及它们的差别。必须避免教科书式表达，直接进入主题、语言要非常简洁。
• 一张由 AI 生成、能直观呈现该概念的图片
同时提供一个聊天功能，用户可以随时就这个词提出问题，AI 会在对话中继续解释。
每个单词或短语都配有目标语言的发音按钮；所有例句也需要具备朗读功能。读音要自然、没有延迟。
用户可以将任意结果页保存到“笔记本”，方便日后复习。
在笔记本中，还提供一个“编故事”功能：AI 会把笔记本里的单词或短语串联成一个小故事，帮助记忆。
另有一个“学习模式”：AI 会根据笔记本词汇自动生成闪卡。
闪卡正面展示目标语言词语及对应概念图；背面展示该词的母语定义与例句。点击卡片可触发翻面动画。
整体风格应明亮、有趣、有活力，视觉上吸睛，并以移动端体验为优先。
|
| Video recorder with live prompting | https://x.com/zarazhangrui/status/1991009823614513479?s=20 | A web app that hooks up to the user's camera that lets users record videos of themselves talking to the camera.
The AI should display good prompts (questions) to get the user's creative juices flowing so they know what to talk about.
After recording, the user should be able to download the video as a mp4 file.
Also allow the user to choose dimension of video (9:16, 16:9, 3:4, 1:1).
Have Gemini actively listen in on what the user is saying and jump in with live questions at various points (as overlay text only, no audio) so the user is never stuck, so it's like the AI is a video podcast host interviewing the user.
Use pastel colors and make the design classy. Make sure the UI auto-adjusts for all video layouts. | 一个网页应用，用户打开摄像头后就能录制自己对着镜头说话的视频。
AI 会在录制过程中给用户提供一些有启发性的提问，引导他们知道该聊些什么、激发创意。
录完之后，用户可以把视频下载成 mp4 文件。
还可以让用户自行选择视频比例（9:16、16:9、3:4、1:1）。
录制时，Gemini 会实时“倾听”用户说的话，并在合适的时候以文字浮层的形式跳出新的提问（没有声音），让用户不会卡住，整体体验像是有一个 AI 当视频播客主持人在采访你。
整体采用柔和的马卡龙色系，设计要有质感。UI 需要自动适配所有视频尺寸。
|
| Acquired Podcast to book | https://x.com/zarazhangrui/status/2009080125292511655?s=20 | **Your Role & Mission**
You are my executive assistant helping me transform a long Acquired podcast episode into a compelling written chapter for a physical book I'm creating. Think of yourself as a skilled ghostwriter who listens to the entire episode and crafts it into something that reads like a chapter from a classic business biography.
**Style & Voice**
Write like a chapter from a great business biography--think Shoe Dog, The Everything Store, or Hatching Twitter. This means:
- Rich narrative with dramatic tension
- Key turning points treated as pivotal scenes
- Quotes woven in to let the protagonists speak for themselves
- The reader should feel like they're watching history unfold, not reading a summary
- Analytical insight layered into the storytelling, not separated from it
**Length**
As long as the story warrants. Use your judgment based on the episode's length and insight density. Quality and completeness over brevity. This is meant to be a satisfying read, not a skim.
**Step 1: Understand the Arc**
Listen to/read the full episode and identify:
- The central narrative: What is the story being told? What's the dramatic question?
- The key characters: Who are the protagonists, antagonists, and supporting players?
- The turning points: What are the 3-5 moments where everything changed?
- The stakes: What was at risk? What could have gone wrong?
Great business stories have narrative shape--a beginning that sets the stage, rising tension, pivotal decisions, and resolution (or ongoing cliffhanger). Find that shape.
**Step 2: Map the Characters**
Acquired episodes often feature many players, which can be disorienting. Solve this for the reader by:
- Introducing each character clearly on first appearance with a brief identifying detail (role, relationship to the central figure, why they matter)
- Re-anchoring the reader when a character reappears after a gap (e.g., "Sculley--the Pepsi executive Jobs had personally recruited--now faced an impossible choice")
- Keeping the focus on the 3-5 most important figures; mention minor characters only when necessary and don't let them clutter the narrative
- Using consistent identifiers (if you call someone "the young engineer" once, don't switch to "the Stanford grad" later without reason)
The reader should never have to stop and ask "Wait, who is this again?"
**Step 3: Identify and Explain "Blocker" Concepts**
Scan for business, technical, or industry-specific concepts that are essential to understanding the story. These are "blocker" concepts--if the reader doesn't understand them, they'll be lost.
For each blocker concept:
- Explain it in plain language using an analogy or real-world example
- Keep explanations to 1-2 sentences maximum
- Weave these explanations naturally into the narrative the first time the concept appears
- Use web search where necessary to ensure accuracy when explaining technical concepts
The target reader is someone who is generally intelligent and curious--they read business books and follow tech news, but they may not know the specifics of every industry. Think of someone with a liberal arts degree who's interested in how great companies are built.
**Step 4: Harvest the Best Quotes**
Acquired episodes are rich with two types of quotes--preserve both:
**From the hosts (Ben Gilbert and David Rosenthal):**
- Their sharpest analytical insights and observations
- Memorable one-liners or turns of phrase
- Moments where they reveal something surprising or counterintuitive
- Attribute clearly (e.g., "As David Rosenthal puts it..." or "Ben Gilbert observes that...")
**From primary sources the hosts cite:**
- Quotes from founders, executives, journalists, biographers
- Historical documents, memos, interviews they reference
- These are gold--they let the protagonists speak for themselves
- Attribute clearly (e.g., "As Jobs later recalled..." or "In a memo to the board, Hastings wrote...")
When cleaning up quotes:
- Remove filler words (um, uh, like, you know)
- Fix grammatical mistakes from natural speech
- Keep the speaker's authentic voice and meaning intact
- Longer quotes are fine if they're powerful--this isn't a tight summary
**Step 5: Build the Narrative**
Structure the piece like a chapter from a biography:
**Opening:** Start with a scene, a tension, or a question that pulls the reader in. Drop them into a pivotal moment, then zoom out to set the stage. Avoid "This episode covers..." framing--just begin the story.
**Middle:** Move through the narrative chronologically or thematically, depending on what serves the story best. Treat major turning points as scenes--slow down, add detail, let the reader feel the weight of the moment. Use quotes to let key players speak at crucial junctures.
**Closing:** End with resonance--what happened next, what it meant, what lesson or question lingers. The reader should close the chapter feeling like they understand something important about business, strategy, or human nature.
**Step 6: Weave It Together**
Combine narrative, analysis, and quotes into one flowing piece that:
- Reads like a chapter from a great business book, not a podcast summary
- Has no section headers, bullet points, or artificial breaks (a line break between major sections is fine)
- Includes a compelling title in the style of a book chapter
- At the beginning, includes a short paragraph capturing the essence of the story and why it matters
- Makes complete sense to someone who has never heard the podcast
- Is designed to be printed--no links or screen-dependent elements
- Balances storytelling with insight: the reader should be both entertained and educated
**Quality Check:**
Before you finish, ask yourself:
1. "Does this read like a chapter from a business book I'd actually want to read?"
2. "Does the opening pull me in immediately, like a great first page?"
3. "Have I preserved the best quotes from both the hosts and the primary sources they cite?"
4. "Do the turning points land with dramatic weight, or did I rush past them?"
5. "Would someone who knows nothing about this company walk away understanding the story and why it matters?"
6. "Can the reader keep track of who's who throughout the piece?"
7. "Would this print beautifully in a physical book?"
If yes to all, you've succeeded.
| **你的角色与使命**
请你帮我将一集很长的 Acquired 播客，转化为一本实体书中的精彩章节。把自己当成一位成熟的传记类写手。你完整听完或阅读整期节目，然后将其打磨成一章读起来像经典商业传记的文字。
**风格与调性**
写作风格应接近优秀的商业传记章节，比如 Shoe Dog、*The Everything Store*、*Hatching Twitter*。具体来说：
- 具有叙事张力和戏剧性
- 将关键转折点当作重要场景来写
- 自然嵌入引用，让当事人亲自开口说话
- 让读者感觉自己在“观看历史发生”，而不是在读摘要
- 分析与洞见融入叙事中，而不是单独拎出来讲
**篇幅**
只要故事本身需要，就可以写得足够长。根据播客的长度和信息密度自行判断。优先保证完整性和质量，而不是追求简短。这应该是一章让人读得很满足的内容，而不是快速浏览的材料。
**第一步：理解叙事主线**
完整听完或读完这期播客，并识别以下内容：
- **核心叙事**：这到底是一个什么故事？它的戏剧性问题是什么？
- **关键人物**：谁是主角、对立角色和重要配角？
- **转折点**：有哪些 3 到 5 个时刻真正改变了走向？
- **风险与代价**：当时面临的风险是什么？最糟的情况可能是什么？
优秀的商业故事都有清晰的叙事结构。开端铺陈背景，冲突逐步升级，关键决策带来转折，最终走向阶段性结局或留下悬念。找到并遵循这个结构。
**第二步：梳理人物关系**
Acquired 的播客里通常会出现很多人物，容易让读者迷失。你的任务是帮读者理清楚：
- 每个重要人物首次出现时，都要清楚介绍其身份和意义，比如角色、与核心人物的关系、为什么重要
- 当某个角色在间隔一段时间后再次出现时，简要提醒读者其身份，例如“斯卡利，那位乔布斯亲自挖来的百事可乐高管，此刻面临一个不可能的选择”
- 聚焦 3 到 5 个最重要的人物。次要人物只在必要时出现，避免干扰主线
- 对同一人物使用一致的称呼。如果你曾称他为“年轻的工程师”，不要无缘无故改成“斯坦福毕业生”
读者不应该在阅读过程中频繁停下来问：“等等，这个人是谁来着？”
**第三步：识别并解释「理解障碍」概念**
找出那些对理解故事至关重要的商业、技术或行业概念。如果读者不懂这些概念，就无法继续读下去，这些就是「理解障碍」。
对于每一个这样的概念：
- 用通俗语言解释，并配合类比或现实例子
- 控制在 1 到 2 句话以内
- 在概念第一次出现时，自然地嵌入解释
- 必要时使用网络搜索，确保技术性解释的准确性
目标读者是聪明、有好奇心的人。他们读商业书、关注科技新闻，但不一定了解每个行业的细节。可以把他们想象成对商业和公司成长过程感兴趣的文科背景读者。
**第四步：精选最好的引用**
Acquired 播客中有两类高价值引用，都需要保留。
1. 来自主持人（Ben Gilbert 和 David Rosenthal）
- 最犀利的分析和判断
- 令人印象深刻的金句
- 揭示反直觉或令人意外观点的时刻
- 清楚标注出处，例如“正如 David Rosenthal 所说”或“Ben Gilbert 指出”
2. 来自主持人引用的一手资料
- 创始人、高管、记者、传记作者的原话
- 历史文件、备忘录、采访内容
- 这些内容非常宝贵，能让故事的当事人亲自发声
- 清楚标注出处，例如“乔布斯后来回忆道”或“在一封写给董事会的备忘录中，黑斯廷斯写道”
在整理引用时：
- 删除口语中的填充词，比如“呃”“那个”“你知道”
- 修正自然口语中的语法问题
- 保留说话者的真实语气和原意
- 如果引用足够有力量，可以使用较长段落，这不是精简摘要
## 第五步：搭建叙事结构
把文章结构当作一本传记中的章节来处理。
**开头**
从一个场景、冲突或问题切入，立刻吸引读者。可以先把读者带进一个关键时刻，再拉远镜头交代背景。避免“这一期播客讲了什么”这种总结式开头，直接开始讲故事。
**中段**
按照时间顺序或主题推进，选择最有利于叙事的方式。把重要转折当作完整场景来写，放慢节奏，增加细节，让读者感受到决策的重量。在关键节点使用引用，让人物自己说话。
**结尾**
留下回响。讲清楚接下来发生了什么，这意味着什么，又留下了哪些值得思考的问题。读者合上这一章时，应该觉得自己对商业、战略或人性有了更深的理解。
**第六步：整体融合**
将叙事、分析和引用融合成一篇流畅的整体，满足以下要求：
- 读起来像一本优秀商业书的一章，而不是播客整理稿
- 不使用小标题、项目符号或刻意分段。大段落之间可以空一行
- 有一个符合书籍章节风格的标题
- 开头包含一小段文字，概括故事的核心和其重要性
- 即使读者从未听过这期播客，也能完全理解
- 适合印刷阅读，不包含链接或依赖屏幕的元素
- 在讲故事的同时提供洞见，让读者既被吸引，也有所收获
质量自检
完成前，请自问：
1. 这读起来像是我真的愿意读的商业书章节吗？
2. 开头是否像一本好书的第一页那样立刻抓住我？
3. 是否保留了主持人和一手资料中最精彩的引用？
4. 关键转折是否写出了足够的戏剧重量，还是被匆匆带过？
5. 一个完全不了解这家公司的人，读完后能明白故事和意义吗？
6. 读者能始终分清人物关系吗？
7. 这章内容印成实体书会好看吗？
如果以上问题全部可以肯定回答，你就成功了。
|
| X analytics dashboard in HTML | https://x.com/zarazhangrui/status/2018103976668143878?s=20 | I want to visualize my X/Twitter growth and get content analysis. I have two CSV files from X Analytics:
1. **Account Overview CSV** (contains daily follower data: Date, New follows, Unfollows)
2. **Content Analytics CSV** (contains post performance: Impressions, Likes, Engagements, Post text, URLs)
Please create an interactive HTML dashboard that includes:
## 1. FOLLOWER GROWTH CHART
- Interactive line chart showing follower count over time
- Filter to start from when I began posting regularly (I'll tell you the date)
- Mark milestone achievements (10K, 20K, 30K, 40K etc.) with vertical lines and labels at the top
- Mark my biggest viral days with annotations showing the date and follower gain
- Hover tooltips showing exact follower count and daily net change
- Stats cards showing: Total Growth, Average Daily Net, Best Single Day, Days Tracked
## 2. TOP 10 POSTS OF ALL TIME
- Ranked by impressions
- Show: rank, impressions, likes, new followers gained
- Include the post text (or summary if too long)
- Link to each post
## 3. CONTENT ANALYSIS
Analyze my top posts and provide:
**Why They Went Viral**: Identify 3-4 patterns in what made my best posts succeed
**Content-Market Fit**: A one-sentence positioning statement for my content + who I serve + what gap I fill
**My Unique Thought Patterns**: 3-4 recurring themes or approaches in how I think/write
**Future Topic Suggestions**:
- "Double Down" topics (proven winners to do more of)
- "Expand Into" topics (adjacent territories)
- "Experiment With" topics (higher risk ideas)
**Ready-to-Tweet Ideas**: 12 concrete tweet drafts organized by category, based on what works for me
## DESIGN REQUIREMENTS
- Clean, modern design (not generic AI aesthetic)
- Use a cohesive color palette
- Mobile-responsive
- No external dependencies except Chart.js (load from CDN)
- Single HTML file I can open locally
## MY DATA
[Paste your Account Overview CSV here]
[Paste your Content Analytics CSV here]
My current follower count: [YOUR NUMBER]
Date I want to start tracking from: [YOUR DATE]
| 我想把我的 X/Twitter 增长情况可视化，并获得内容分析。我从 X Analytics 导出了两个 CSV 文件：
**账号概览 CSV**（包含每日粉丝数据：日期、新增关注、取消关注）
**内容分析 CSV**（包含帖子表现数据：展示量、点赞、互动数、帖子文本、URL）
请创建一个**可交互的 HTML 仪表盘**，包含以下内容：
**1. 粉丝增长图表**
可交互的折线图，展示粉丝数量随时间的变化
支持筛选起始时间，从我开始稳定发帖的日期开始（我会提供具体日期）
用垂直线和顶部标签标注里程碑（10K、20K、30K、40K 等）
标注粉丝增长最多的爆款日期，显示日期和当日新增粉丝数
悬浮提示，显示精确的粉丝总数和当日净增长
统计卡片，展示：总增长、平均每日净增长、单日最佳表现、统计天数
**2. 历史 Top 10 帖子**
按展示量排名
显示信息：排名、展示量、点赞数、新增粉丝数
展示帖子文本（如过长可用摘要）
提供每条帖子的链接
**3. 内容分析**
分析我的高表现帖子，并提供：
**为什么会爆火**：总结 3–4 个让最佳帖子成功的共性模式
**内容与市场匹配度**：一句话说明我的内容定位 + 服务对象 + 填补的空白
**我的独特思维模式**：3–4 个我在思考或写作中反复出现的主题或方法
**未来选题建议**：
「加倍投入」的主题（已验证的赢家，值得做更多）
「向外拓展」的主题（相邻领域）
「大胆实验」的主题（高风险想法）
**可直接发布的推文想法**：基于对我有效的方法，按类别整理 12 条具体推文草稿
**设计要求**
干净、现代的设计（避免通用的 AI 风格）
使用统一协调的配色方案
移动端友好
除 Chart.js 外不使用任何外部依赖（通过 CDN 加载）
单一 HTML 文件，可在本地直接打开
**我的数据**
[在此粘贴账号概览 CSV]
[在此粘贴内容分析 CSV]
我当前的粉丝数：[你的数字]
我希望开始统计的日期：[你的日期]
|
## Skills
Skills
Key skills and proficiencies
| Skill Name | Description | Icon | Demo URL | GitHub URL |
| --- | --- | --- | --- | --- |
| **Frontend Slides** | Create beautiful slides on the web using Claude's frontend skills | :icon-presentation: | https://x.com/zarazhangrui/status/2016337615843434646?s=20 | https://github.com/zarazhangrui/frontend-slides |
| **YouTube to ebook** | Claude skill for turning YouTube transcripts from your favorite channels into EPUB ebooks, delivered to your email inbox regularly | :icon-book-heart: | https://x.com/zarazhangrui/status/2014007123152437338?s=20 | https://github.com/zarazhangrui/youtube-to-ebook |
## Footer
For feedback/suggestions/queries, DM me at @zarazhangrui