大雨大風攏是同款咱的命運

Attitude 1: LLMs are stupid (deny). Attitude 2: LLMs are not relevant to our field (Avoidant). I feel both attitudes miss the opperturnity to import a humanistic perspective into the ethical development of LLMs.

Agree! I believe that the build/understanding open source model course is especially valuable for non-cs background students. I switched from political theory to math/cs after GPT-4's release. Most of my humanity/political theory/anthropology fellows still have a quite hostile attitudes to LLM.

Received the copy yesterday. It is a truly inspiring book. It helps me to revisit many questions I was thinking when I was a humanity student. Reading this makes me feel that the past humanity experience has its power to make cs different. Wish there could be something like this next year!

A flowchart offers a deep way to explore code—this isn’t AI for Humanities; it’s Humanities for AI.

I hope professors assign this as a “reading and summarizing task” for students, especially those outside CS. Humanities students don’t fear reading but often fear long codes. Writing can eliminate those fears.

The google colab notebook I’m working on tonight comes from aiforhumanists.com/tutorials/. Deeply appreciate AI for humanists team for creating this amazing digital humanity tutorial on BERT! It is still a good starting point for AI beginners.

Will explore!

Interesting thoughts! I'm curious about how how viewing planning as the core of AI could reshape our approach to both studying and applying AI. In other words, how can this new perspective give to us, as students, or as reseachers?