Keep on to blur preview images; turn off to show them clearly
RT @HkUsInsight: 刚玩推特没多久，无意间刷到fin神的这篇文章，看完之后深表认同且大受震撼，这让我想起了赫拉利的【人类简史】这本书，赫拉利在【人类简史】中讲过三个核心观点： 1. 认知革命（约7万年前）--人类学会用语言、符号、故事传递“抽象信息”，这是第一次“…
Prompt Engineer, dedicated to learning and disseminating knowledge about AI, software engineering, and engineering management.
Today, I built myself a clipping tool using @warpdotdev. It was super fast - only took a few minutes to build! ⚡️ This tool: 1. Transcribes a given video file 2. Parses the transcription and finds the best moments using AI 3. Clips parts of the video using ffmpeg Within a minute, it parsed a 4-minute video into 5x 10-20 second clips that I can use for content. Brilliant! 🤩
Btw, this was the prompt I gave @warpdotdev (the one in the purple rectangle) 👇 After that, it did its job; all I had to do was add an @OpenAI API key, and boom. It just worked!
RT @latentspacepod: We asked @corbtt about the rise and fall of LoRAs: “Predicated on the fact that you're doing fine-tuning at all, LoRAs…
achieve ambition with intentionality, intensity, & integrity - @dxtipshq - @sveltesociety - @aidotengineer - @latentspacepod - @cognition + @smol_ai
核心点在于「我写的代码我更容易维护」并不等同于「代码更容易维护」 “我写的更顺手”是熟悉度， “谁来都能接手”才是可维护性。 如果 AI 和同事的代码同样让你犹豫，问题不在“AI”，而在规范与协作。
🖥️ Indie Maker 🛠️ 星球「海哥和他的小伙伴们」 📌 油管「海拉鲁编程客」 🌸 沦为程序员的段子手/猫咪
RT @kalomaze: @oxycblt work with real data from real sources instead of unrealistically "clean" baselines. especially don't treat existing…
RL and efficient distributed pretraining • eXperiments lab • memes and training lores
Yea this thing is not coming into my house.
Founder 📈 @parqetapp Host of 🎙 @minimalempires Prev. @stripe