So after a long break I get to return now to my C++ neural network library which implements our neuromodulator framework to be used for learning in embodied agents. And I am now wiser by the following two facts:
-
Any length of time greater than one day of break will put a thick layer of rust on your C++ skills.
-
C++ sucks. I’d so much prefer if I could use Python. But until our embedded devices are powerful enough for real-time Python, I’ll keep banging my head against the table.
Why is it that I always end up spending two or three days implementing something spiffy only to find out then that there’s some preprocessor loop or C++-braindeadness preventing me from doing what we get working in Python in a matter of hours?
