Computational Selves, MinDVis and AI takeoff ā Live and Learn #8
Welcome to this edition of Live and Learn. This time with wisdom from Eric Schmidt, a story of what AI takeoff might look like, and an awesome paper on a theory of computational selves. Enjoy.
āØ Quote āØ
I have no special talents. I am only passionately curious.
ā Eric Schmidt - (source: How Google Works)
Links
The Computational Boundary of a "Self" by Michael Levin. A self is "just" a system that can pursue a set of goals. To pursue goals, systems need cognitive abilities. Within a system, there can be multiple selves interacting with one another in a hierarchy. Selves are made up of smaller selves, everything alive is computing some form of goals. DNA and proteins make up cells, cells make up tissues, tissues organs, organs organisms, and organisms societies. This paper is long and tough, but worth the effort, because it introduces many interesting ideas, like the "cognitive lightcone" or how homeostasis as a simple goal, gives rise to complexity and eventually brains. And how brains are in their cognitive abilities and mechanisms related to the "simpler" selves, formed by single-celled organisms. Highly mind-bending stuff.
Blitzscaling Lecture by Eric Schmidt. Eric Schmidt was CEO of Google for quite some time and used to work for different boards and companies before that. He's quite the legend in the startup/business world. Listening to him tell stories of what they used to do in order to build successful businesses is extremely interesting. A lot of it boils down to: Hire smart people, and give them the tools to build what they want to build.
MinDVis by Zijiao Chen et al.. TLDR: People work hard to create better machine-learning techniques to decode brain activity from non-invasive procedures. In a way, brain reading might be a very real thing in the not-so-distant future. I mean wtf... Brain reading without electrodes implanted directly into brains is not only possible, but we already know how to do it to some extent. Isn't that crazy?
AI takeoff-story by LessWrong - Edouard Harris. With ChatGPT, Midjourney, and other tools on the rise, I've been thinking a lot about the role of AI. Stories like this one, are a nice tool. Because they show different scenarios of how AI takeoffs could possibly happen, showcasing the plethora of possibilities. There is also a collection of AI takeoff stories available for free online.
š Midjourney š
š¶ Song š¶
Favorite Things by John Coltrane
I hope you found this newsletter useful!
Have ideas for improving it? Please let me know.
Cheers,
ā Rico