August Subscriber Post
Join my retreat for snake enlightenment (it's just a petting zoo)
The above is what “Snake Cult of Consciousness” yields when run through an AI slop meme generator. Not entirely bad. But alas, I have been beaten to the punch:
Job search
Lest I be reduced to that, I am currently looking to switch roles (as the current start-up looks dicey). So, if you know of any openings in data science or ML, especially related to NLP in health care or psychological data, please let me know. I am also happy to make any connections if you are adjacent to that space.
Previously on Vectors of Mind
How to prompt chatGPT
Here are some of the most overpowered prompting techniques:
1) Use clear and specific instructions
2) Chain-of-thought reasoning
3) Collect the required information before doing a task
4) Break tasks up into steps
5) Technical terms are your friend...
The Bullroarer: Much More Than You Wanted To Know
“No ethnomusicologist, I think, would stand for plurigenesis as regards the bull-roarers, which even in decorative detail are often alike and are used for the same purpose wherever and whenever found.” ~Jaap Kunst, 1960
Links
Traditions of Conflict with a great piece on male mystery cults. After showing many use flutes (and bullroarers) as a secret ritual object, he asks:
“Still, a question remains: why the flutes, and their phallic symbolism? Other than vaguely appealing to psychological universals, a general tendency among males to be fascinated with their own genitalia, and a common human predilection towards symbolism, I don’t have a good answer. Anthropologists who studied men’s cults have often interpreted the symbolism in neo-Freudian and psychoanalytic terms, and while I do not find that approach particularly satisfying, I can’t say that I have a better answer.”
Interesting conversation with an LLM about its own mind.
Paper: Surprising gender biases in GPT. TLDR male expendability (a hypothesis in anthropology) is baked into language models. Not really surprising if you’ve interacted with them or understand how they are made. Pairs well with this analysis of GPT’s training data:
Djokovic went into the Tennis final as an underdog and fought like hell to take the title. This analysis before the match predicted just that would happen. He argues that an Olympic Gold is the only thing Djokovic hasn’t won, and thus, he would somehow make it happen, even if Acaraz is better on paper. I’m reminded of an interview with Woody Allen, who said he didn’t watch movies much anymore because they were so predictable and the true stories were being told in sports.
Not sure if you've come across this site (https://dreams-of-an-electric-mind.webflow.io/) before, but if you enjoy reading trippy self-reflections by LLMs... it might fit your fancy ;)