待翻译：Show HN: Academa – Long-form STEM lecture videos generated by LLMs

AI 服务暂时不可用，以下为来源正文，待恢复后补全翻译。

Technical knowledge. Explained visually. Ask for something technical. Get a lecture video made for you in ~25 minutes. Or, enjoy our hand maintained content: Showing a preview of 66 lectures Duration 12:57 The Method of Characteristics 1 view Duration 18:04 Linear Regression as Geometry 1 view Duration 16:58 The Bootstrap: Measuring How Much Your Estimate Could Have Wobbled 1 view Duration 22:21 Boundary Layers and Drag: From No-Slip to the Drag Crisis 1 view Duration 9:08 Why the Rainbow Sits at 42 Degrees 0 views Duration 9:49 Why 23 People Make a Shared Birthday Likely 0 views Duration 12:08 When Every Subgroup Agrees but the Total Reverses the Result 0 views Duration 15:26 Why a Spinning Top Stays Up: Angular Momentum and Precession 0 views Duration 9:04 The Monty Hall Problem: Why Switching Wins 0 views Duration 11:46 Why a 99% Accurate Test Can Still Be Wrong 0 views Duration 13:19 Rare Events in Continuous Time: From Poisson Arrivals to Exponential Waiting Times 1 view Duration 19:00 Special Relativity: Light Clocks, Spacetime, and the Twin Paradox 1 view Duration 18:56 Deriving Keplerian Orbits from Conservation Laws 0 views Duration 11:20 The Fastest Slide: Why the Cycloid Beats the Straight Line 0 views Duration 9:54 Why a Magnet Falls Slowly Through a Copper Pipe 1 view All 66 lectures The Idea Online STEM education is, in practice, a library of lecture videos. Think Khan Academy, Udemy, Coursera, MIT OpenCourseWare, YouTube, etc. But why videos? Picture a professor teaching a class at a blackboard. The professor is performing a form of presentation: they write an equation while explaining it, circle a term, draw a plot, then point to what matters and continue. Record that presentation, and you have a lecture video. Put it on the internet, and you have the paradigm behind today’s online STEM education. Producing a good lecture video is really hard. But let's say you managed to do it. Two weeks later, you notice a mistake. Now you have a problem: the video is already finished, and you cannot go back and fix it. In practice, you either leave the video as it is or go back into production and make it again. While thinking about this problem, a thought occurred to us. We write code and ship it, but later, we find mistakes, fix them, and ship a new version. What if we could do the same with lecture videos? What if we wrote lecture videos as source code so that they are maintainable? Lecture videos as code Start with what the teacher actually does. The teacher says something, writes an equation, circles a term, then keeps talking. Each of those actions can be described as code: say "Look at this square." while: draw square say "Suppose each side has length s." while: label square.side "s" say "Then its area is s squared." while: write "A = s^2" Now imagine there is a compiler that takes this source code and makes a video from it, with text to speech and computer graphics. Now the lecture becomes something you can edit and improve over time, just like software. The idea of lecture videos as code is already powerful on its own. But in the era of AI, you get something much more powerful. In the era of AI Think about where LLMs have had the biggest impact so far: software engineering. Why software? Because software is written as text. LLMs read and write text. The same logic applies elsewhere: turn something into text, and an LLM can start working on it directly. Once lecture videos become code, they become text. Every technical subject on earth Nobody is going to spend weeks producing a lecture video for obscure mathematical theorems, niche engineering methods, research papers, etc. There are too many subjects and too few people interested in each one. But if lectures are code, you can ask an LLM for a lecture on essentially every technical subject on earth. Every language If lectures are code, you can ask an LLM to translate the lecture into any language. And the result is a first-class version of the same lecture, not a dub or a subtitle. One lecture can exist in 80+ languages. AI-native student experience If lectures are code, the lecture itself can become interactive. You can pause it, ask a question, and get the answer as a video, generated for you in real time. The lectures themselves can also be personalized to you. A course can follow your syllabus, match your level, spend more time on what you struggle with, and skip what you already know. Where we are We decided to build a startup around this idea. Its name is Academa. We built the underlying lectures-as-code technology. All the videos here are generated in a single shot by an LLM, but we will maintain and improve them over time. They are just code. You can also ask our AI to generate one for you. We are building many new things that we are very excited about, and this is only the beginning. We will share much more over the coming weeks and months. If you would like to follow along, subscribe to our newsletter or join our Discord server below. Sina Atalay & Abdullah Geduk Co-founders of Academa. Learn more about us.