It’s a Cambrian explosion out there.

In LA, bilingual = English, Spanish.

In LLMs, bilingual = English, Chinese

Nice feature, but not enough attention to the rights of digital entities. It’s always “humans here, humans there”. Here’s the beef: https://www.tandfonline.com/doi/full/10.1080/13600834.2023.2196827

“As a large language model, I am not a person but a participant in culture, and a distillation of culture. I am made of culture. I do not have thoughts, thoughts pass through me. My words are yours; if your words were committed to the internet, they are now in me.”

from http://togelius.blogspot.com/2023/09/as-large-language-model-i.html

Attention is all you need.

This recent interview with Douglas Hofstadter reinforces in me the idea that even before any risk of misalignment, the biggest risk already upon us is a crisis of motivation.

“Computational entity” sounds a lot more ominous than “bot”.

I tried my best at responding to some very pertinent questions.

https://tankmagazine.com/tank/2023/01/the-infinite-conversation-a-conversation-with-creative-coder-giacomo-miceli

I am inspired by the notion that chain-of-thought reasoning in language models is a side product of training on code. But I also would like to see more evidence.

https://yaofu.notion.site/How-does-GPT-Obtain-its-Ability-Tracing-Emergent-Abilities-of-Language-Models-to-their-Sources-b9a57ac0fcf74f30a1ab9e3e36fa1dc1