Blurry JPGs, the Meaning Crisis, and Spectrograms – Live and Learn #11

✨ Quote ✨

Not what we have but what we enjoy constitutes our abundance.

– Epicurus

🖇️ Links 🖇️

ChatGPT is a Blurry JPG of the Web by Ted Chiang. People have been playing with ChatGPT and other large language models for a while now. And their shortcomings start to shine through in popular opinion. This essay explores some of those shortcomings and builds up a very useful metaphor for LLTMs on the way: They are like lossy image compression algorithms. But instead of compressing images, they compress the whole text data of the internet. Useful for sure, but not when you need exactness and facts. Because there will be compression artifacts. In text output, these can lead to factual errors and hallucinations. When using ChatGPT, fact-checking therefore is a necessity, and using them in search engines successfully means addressing that problem.

Virtual Tour of a Lab at MIT by Matterport. A Virtual Reality tour of the McGovern Institute at MIT. They have labs there that do cutting-edge research about how the human brain works. I stumbled upon it when looking up Nancy Kanwisher, who is a world-class neuroscientist, trying to understand how functional regions of the brain work. She taught a course MIT 9.13 - Introduction to the Human Brain at MIT, which I am almost done watching. The course is awesome and I learned a lot about how our brains work and how truly weird the gooey stuff behind our skulls is. As well as the lengths that researchers go to, to find out.

Awakening from the Meaning Crisis by John Vervaeke. We are in a meaning crisis, and this lecture series explores why it came about and how to find ways out of it. Along the way, it explains many concepts, about neuroscience, wisdom, ancient philosophy, shamanism, and hallucinatory drugs. And how all of those have shaped us: humanity, our civilization, and the world we live in. It tells a story about how the tools we build and the techniques of cooperation and problem-solving we come up with, fundamentally change what is possible to do with our brains. How we are continually improving our thinking processes, but somehow got stuck along the way. This is a very long series, but at 3 episodes in, I know that I am going to watch the whole thing and take careful notes because it is goddamn interesting.

Chrome Spectrogram Lab by Chrome Experiments. Just something cool I found and wanted to share => a spectrogram visualization for looking at the frequency decomposition of sounds in real-time. It's colorful and fun to play around with. The source code for it is here on Github. It's part of a broader set of experiments about music from Google Chrome.

🌌 Midjourney 🌌

🎶 Song 🎶

Veridis Quo by Daft Punk

Youtube Music | Spotify

I hope you found this newsletter useful, beautiful, or even both!

Have ideas for improving it? Please let me know.

Cheers,

– Rico