What We Liked This Week
LLM Readings
Explaining LLMs to my neighbors by EP9嘉宾Felix Cheng（Head of Applied Science at Zendesk)
https://medium.com/@felixjcheng/explaining-llms-to-my-neighbors-d804b6364fc8
RAFT: Adapting Language Model to Domain Specific RAG
LLM这个领域真有趣，前一天Stella和Amy还在讨论有没有针对RAG的general purpose finetuning，第二天就看到了这篇发表才一周的paper……
https://arxiv.org/abs/2403.10131
Distilling Step-by-Step: 大模型教小模型😄
https://arxiv.org/abs/2305.02301
LLM寶藏網站: Comprehensive LLMs comparison
artificialanalysis.ai
LLM花邊新聞
“Attention is all you need” authors 首次齊聚一堂 - 截圖來自GTC 2024:
- None of them stayed at Google，大廠真的是過去幾年發展AI的好地方嗎？原作者們可能身體很誠實，陸續離開Google自己創業去了;
- 所有人都认同虽然这几年GenAI/LLM红红火火，但其实model本身的发展已经停滞了好多年了。(7 years is like a century in the field of AI…)
UX Design
The Real Unfiltered Chat with Don Norman
EP8嘉宾Felix Lee（ADPList Cofounder）采访UX god Don Norman。老爷子好可爱！
本週看到好玩的產品
Vapi https://vapi.ai/
Build, test and deploy voicebots in minutes. (demo) - 實測有點慢，不過工具做了最popular的幾個voice＆llm的串接，祝你找到最佳排列組合。
Random Stuff
在Instagram上卖lightbox的小哥 https://www.instagram.com/lclightbox/
卧薪尝胆三年Hacker的故事（很吓人……）
https://arstechnica.com/security/2024/04/what-we-know-about-the-xz-utils-backdoor-that-almost-infected-the-world/