I'll head in later and post badly!
It's the technique of running a primary search against some other system, then feeding an LLM the top ~25 or so documents and asking it for the specific answer.
A friend who worked with her is sympathetic to her but does not endorse her: this is a tendency she has, she veers back and forth on it a lot, she has frequent moments of insight where she disavows her previous actions but then just kind of continues doing them. It's Kanye-type behavior.
Stuck in my brain: I used to work at a dating site for Indian people, and one of the things we tried was "LLM-generated pickup lines."
I don't remember most of them and we never made the feature public, but one of them sticks out in my mind for being the most incomprehensible.
Guy-to-girl:
What's your remedy for a Bollywood love affair?
Please share!!!!
If it helps, I know who you are and will still happily tell you incorrect information about yourself and your profession if asked to!
This man's blog is intense, but I am not sure he comes off well! I clicked around and it seems like "wants to be at all the fascist parties, courts acts of violence to complain about on his blog" is, at least in 2024, a really accurate summary of his behavior.
Or, in his words:
I tell them that I’m actually pretty hated and feared by most of these people, and I can only stay around because I criticize particular influential figures in this counterculture so well that they want to fuck me, and so they keep me around to flatter them, to reflect their true hideousness back at them by elevating it to the status of myth, and then they lash out at me like the maenads devouring Orpheus.
Holy fuck! That man does not sound like an engineer. Why is he the CTO of anything?
My opinion is that Jesse Lyu is lying about making any significant changes. (Because otherwise the demo wouldn't have worked)
I don't want bad things for him personally, but I want bad things to happen to people who lie in public.
The code is open source with licensing requirements, so I'm therefore hoping someone Jesse has already made a statement to can write him with these requests:
- For GPL2 licensed components such as Linux: Give me your changes in source form.
- For Apache-licensed components such as Android: What files did you change?
I can imagine him responding in three ways:
- "Sure, here is another lie" -- and then he's locked into an answer which will probably make him look clueless as hell
- "We don't think we have to do that" -- and now the Open Source Reply Guy Brigade instantly hates him.
- -- and now, given that a conversation has actually occurred, he looks evasive.
I read a few of the guy's other blog posts and they follow a general theme:
- He's pretty resourceful! Surprisingly often, when he's feeling comfortable, he resorts to sensible troubleshooting steps.
- Despite that, when confronted with code, it seems like he often just kind of guesses at what things mean without verifying it.
- When he's decided he doesn't understand a thing, he WILL NOT DIG INTO THE THING.
He seems totally hireable as a junior, but he absolutely needs the adult supervision.
The LLM Revolution seems really really bad for this guy specifically -- it promises that he can keep working in this ineffective way without changing anything.
Is there an amount of testosterone that will turn me into a lizardman
NotebookLM was really useful to my friend who has a humiliation kink which he satisfies by erotically roleplaying on Discord: he simply copypasted the chatlogs into the AI input box and received a personalized podcast of two AI voices kinkshaming him.
His primary complaint was that it wasn't longer.