I did actually get a hit after adding "I can fix your furniture" to my tinder bio. I then managed to fumble it and not get a date because I don't know how to flirt. Still, progress!
This is a fun slice of life. I'm glad y'all had a good time!
I've been reading through a lot of your posts and I'm trying to think of how to apply this knowledge to an RL agent, specifically for a contest (MineRL) where you're not allowed to use any hardcoded knowledge besides the architecture, training algorithm, and broadly applicable heuristics like curiosity. Unfortunately, I keep running into the hardcoded via evolution parts of your model. It doesn't seem like the steering subsystem can be replaced with just the raw reward signal, and it also doesn't seem like it can be easily learned via regular RL. Do you have any ideas on how to replace those kind of evolved systems in that kind of environment?
Another issue with stacking is that your measurement and prediction error stacks, so at a certain depth you have basically no certainty about any conclusions at that depth.