There's been a lot of discussion lately around concerns that GenAI is displacing human creativity. Is there an alternate narrative where this technology could be used to broadcast the things we create?
This is something @vakkermans.bsky.social and I have been working on to explore that question.
Bad things often seem to happen when we start with premises like "[the economy/the climate/some complex system] is simple, it's just like [running a household/heating an oven/something relatively simple]". Maybe part of the problem is simplifying complex systems leads to heavy-handed top-down fixes.
Is it fair to say that the prevalent trend in collaborative humanoid robotics at the moment is towards integrating action planning with some combination of generalisation techniques (e.g. imitation learning) and training on large-scale data? If this is right, are there any interesting alternatives?
My second blog for today is about an approach to robotics that takes into account the way humans become language users, and why now might be the right time to start thinking about how this could work:
www.newsituationists.com/2024/11/whyh...
Okay here goes for real. I've got two new blogs to share. The first one involves a thought experiment about how humans use language versus how language models process language, from an evolutionary perspective:
www.newsituationists.com/2024/11/mind...
Well now what's going on in here...
