They took the system, excluded the hardware where the all the behavioural features they could directly observe resided (RAM, ROM and its contents) from their an

They took the system, excluded the hardware where the all the behavioural features they could directly observe resided (RAM, ROM and its contents) from their analysis, and then concluded that their neurobiological-inspired technique was useless for understanding complex systems because it didn't tell them anything useful. If they hadn't done that, they'd have been in serious danger of gaining actual understanding and wouldn't have been able to use the exercise to mock the neuroscience practices they dislike.