🧵 View Thread
🧵 Thread (8 tweets)
Why does it feel like every computer sim of evolution assume that all agents are competing for fitness in isolated parallel universes? It's like all those shit cut-throat Social Darwinism ideas but applied to AI agent model generation. Even see this wrong when implementing NEAT.
Complex meta-structures aren't ever going to form if you don't simulate *POPULATIONS* of agents interacting in the same space. This kind of simulation is so rare that I've only ever seen 3 examples of it in practice (aside from game simulations).https://t.co/g4swQ1JlSN
(simulated) life comes at you fasthttps://t.co/nef4EkY6bT
"We consider MMORPGs [...] best proxy for the real world among human games: they are complete macrocosms featuring thousands of agents per persistent world, diverse skilling systems, global economies, and ad-hoc high stakes single and team based conflict."https://t.co/I0r0w09atj
I wonder how many models of evolution in computer science are riffing off of myths and misunderstandings.Ring species wasn't really even validated until 2012! Allopatric speciation (darwin's finches) might even be a result of collaboration between gut microbes and host!