@volpeon cybernetics is hard actually :neocat_woozy:

non-spike nets have known stability issues that everyone is just ignoring for some reason. spike nets are stable under reinforcement learning but google doesn't care.

i think we probably do have or close to the algorithms needed but it needs some experimentation and tuning and some of that tuning is mind bendingly difficult since we can't even answer questions like what alive-ness looks like on a normal day.