Vlady Veselinov

Here's a blog post about it vladinator.net/blog/launchi...

In short - launched a live music platform that helps emerging artists get paid, keep local venue doors open, and make it easy for fans to find the live music they’re missing.

there's an Xzibit joke in there somewhere

A beautiful loop: An active inference theory of consciousness Can active inference model consciousness? We offer three conditions implying that it can. The first condition is the simulation of a world model, whic…

@steveklabnik.com models of models 👀 www.sciencedirect.com/science/arti...

YouTube video by Yupty Live Music Organizing live shows is hard

If anyone is curious about what I've been up to for the past 2 years www.youtube.com/watch?v=UYos...

Anyway, don't know where I was going with this, but the context engineering thing seems cool

And the neat part is these actions don't have to affect the world, but can be purely perceptual, e.g. turn your head or squint to see something new that resolves your uncertainty and improves your context.

Once you select a plan, its "paths" can be resolved by something akin to reflexes which turns them into action.

Then when you do the inference (produce the paths), you now have a "plan in your head", you can do that a few times to get a few plans and then do some kind of search operation between them to pick the one you like the most.

If you're into the neuroscience angle, there's also a really cool concept about planning, where a plan can be looked as a series of paths in a latent space.

So you "choose your axes" given a certain criteria, e.g. energy preservation, reducing uncertainty.

Seems useful though

Can't help but read *insert LLM term* + "engineering" as "bolting on stuff to LLM", but this might be the cynic in me talking.

There's a thing in neuroscience for this - Bayesian model reduction, could be interpreted as removing things from "context" that aren't useful. Also, epistemic foraging - actively seeking out new data to reduce uncertainty.

jail

Can somebody teach me how to write em dashes, I think I'm ready to start using them.

We out here being farmed

I guess the confidence thing comes from a sense of belonging to a community of skeptics in the face of hype

I think Yann LeCun's opinion has to do with learning efficiently from small amounts of data. LLMs don't do well in that aspect currently.

Also, a lot of people are looking at the end result, without considering the data implications.

It's a matter of definitions and what you care about specifically.

👀

Something non-equilibrium steady states something

Max giving a talk w the slide in OP

lol this may be the most cogsci cogsci slide I've ever seen, from @maxkw.bsky.social

"before I got married I had six theories about raising children, now I have six kids and no theories"......but here's another theory #cogsci2025

Still doesn't change the fact that it feels surreal to watch people do it though

Saw this recently and I'm not sure how I feel about it, but I'd lie to you if I said I didn't see some people traversing it as the years pass

My childhood friends were all "skinhead, do business and underpay everyone" so everything after that feels mild to me😅

What do you suggest more specifically?

@mrdiscopop.bsky.social I'm bootstrapping a live music service to help emerging artists get paid. Here's a press kit in case you're interested yupty.live/press

idk they look pretty chill to me

Would love to watch if it's recorded