Conclusion

We know how to implement things with CAs and FLIF neurons.
Multi-level approach needs lots of further exploration.
We need to figure out more about learning those things so that we don't have to program them.
Physical robots are one way forward.
What do you all think?