Activity for mensmachina.com
Loading activity...
One of the mad science things I’ve been throwing tokens at is to see if I can create cognitive modules in models through the right scaffolding and evolutionary pressure. This is a very neat paper that suggests maybe it’s less mad, more science.
The human brain is strikingly modular: distinct networks for language, formal reasoning, social reasoning, physical reasoning. Is this fundamental to intelligent systems, or an accident of evolution? In our new preprint, we find the same modular organization emerges in LLMs.
This isn’t exactly my area of expertise, but I believe that there are a bunch of 1-way doors you go through when designing and building a data centre. If you don’t go through them right, you can never truly establish trust again. I wonder if all those doors in all those DCs were traversed right.
It's not "local" as in on your desktop, but that is at "community" scale. My first job out of high school, I worked for a local ISP. They had a shopfront in town, and they ran the server side out of one of their homes. I remember shelves upon shelves of US Robotics 56K modems blinking away. 1/
LPC. Now that’s a name I’ve not heard in a long time. A long time.
Coding on a MUD in LPC. You're an object. Others are objects. The rooms you're standing in are objects. You're carrying objects. You can code functionality into objects you develop, as can they. You can patch (call) arbitrary functions on objects around you. The dynamics were amazing: code fights.