Sublime

Can somebody make this in the studio ghibli style? https://t.co/39kUvex0LB

#Juchats 🐂🍺这交互设计！妥妥学习神器🚀 https://t.co/xYkO8qS0cL

Wrote a piece on the unfortunate drama (and borderline scientific fraud) that has been unfolding for the last year in the academic community around the environmental impacts of deep-sea mining. https://t.co/95ZHOus3cp

风光二十年，还是窝囊一辈子？ 2025年，我要把自己关在房间里，一年不出门，专心研究链上和项目。 工作室已经租好，三室两厅三卫，房租一年4万，设备也陆续到位。 再请一个阿姨，专门负责做饭和打扫卫生，年薪6万。 另外两个房间留给两个粉丝——有没有想来的？吃喝拉撒住，全都包了。 作息安排 08:00... See more

> We find that LLMs’ intelligence [...] almost linearly correlates with their ability to compress external text corpora > compression efficiency [...] serves as a reliable evaluation measure that is linearly associated with the model capabilities IQ for LLMs just dropped https://t.co/B9UOeLctRS

探索 AI Agents 方案和代码实现 概述：著名开源作者 @nutlope 开源的面向开发者的实用指南网站，受 Anthropic 「Building effective agents」启发，提供了六种常见的 AI Agent 工作模式及其代码实现示例，帮助开发者快速在自己的应用中构建高效的 AI 功能 1. 提示链(Prompt Chaining) 将多个 LLM 调用串联起来，前一个调用的输出会作为下一个的输入，实现结构化的推理和任务分解 2. 路由模式(Routing)... See more

I have been posting (and reposting) tutorials for years now, if you want to see my process and thinking more in depth with talks and live demos, considering signing up for my summer classes at @UnderpaintAcad https://t.co/0T6hrpWI6z