Kevin Slote

One thing missing from JD Vance discourse is the line from his book:

Do you want to be a good terminator, a bad terminator, or a neutral terminator.

Like wtf is a “neutral terminator.” This dude has clearly never seen The Terminator.

I do have one evopsych argument that I concocted: women love bald men. Our aerodynamic heads give us an advantage in speed to save them from predators.

My friend’s student just asked chatGPT what number to set for a random seed.
It answered: 42
Talk about layered

The issue is that the entire premise of the Turing test is "can we deceive people into thinking this is a human?" These systems are accidentally optimized for deception.

Wow, my paper is getting a lot of attention.

Whoa. This is cool

YouTube video by Kevin Slote Online performance estimation for machine‑learning models with zero labelled data.

Check out my youtube video detailing mine and my collaborators method for estimating model performance on unlabeled data streams.

www.youtube.com/watch?v=-cgL...

I really hate cybersecurity trainings. Some boomer clicked on a Nigerian prince scam, once, and now it’s my problem.

Hi. This is cool

The AI slop situation is quickly spiraling into an academic crisis imho. The problem of undergraduates writing essays was mild in comparison to the amount of conceptual gobbledygook I’ve read in the past week.

I just want to say that I really admire your messaging and commitment these beliefs.

This is my favorite subject. I saw someone at SIAM apply this to neural networks recently. Cool stuff

That’s neat

I’ll give the paper a thorough read.

All these people calling the new Superman too “woke” is making me really want to see it.

Nice. I haven’t had a ton of success with the symbolic regression technique on my data

Causal Network relating news media, social media, and gun purchases

Just published in PNAS Nexus!

Using high-resolution, time series from 2012–2020 and applying the PCMCI+ causal algorithm, we find:

The Twitter posts by groups like the NRA is another direct causal driver of firearm acquisition.

academic.oup.com/pnasnexus/ar...

I’m going to dedicate my dissertation to my haters

I think the main gripe people have is when they try to build rag systems in industry gigs. In works great if your search engine works and the data is well organized. Also, RAG performance is near impossible to measure with the existing scientific literature on the topic.

Easy money

If the search engine doesn’t work, the whole process breaks imo.

Oh! I like this!

Before ChatGPT the world was full of scientific awe and wonder. Not I’m just skeptical that everything I read is conceptual gobbledygook

I’m so sick of the machine learning and AI. Like, I study brains, the most complex system known to humans. Why would anyone assume ML/AI is applicable to something that hinges on i.i.d assumptions? Model metrics are just the new p-values.

Belykh-Lab (Biological and Engineering Networks Lab) Org profile for Biological and Engineering Networks Lab on Hugging Face, the AI community building the future.

huggingface.co/Belykh-Lab
I’m trying to start a trend in the applied math community of releasing data sets on @hf.co

“Why wasn’t the warsaw uprising flying Blutfahne?” - the dumbest person alive.

When I have more time, I plan on reading Wilhelm Reich’s The Mass Psychology of Fascism. But you really don’t need a psychology degree to see that sexual repression is the root cause of the fascist neurosis, just look outside.

Fascism is all libidinal economy all the way down

Congratulations

Beaker and gonzo as the main guy and baby in the David Lynch movie Eraserhead