New image classification video, page-121

I think your point was well taken: the same functionality of the Akida neuromorphic hardware can be simulated on Von Neumann architecture.

Quantum computers can be simulated using Von Neumann architecture as well. However, they are extremely slow and their performance decreases exponentially with each qubit simulated.

It is not uncommon to use existing hardware to simulate new processors for different architectures. This allows developers to create tools as well as become acclimated with, and write code that runs on the new architecture before it is baked into silicon.

As also noted, a simulation is not representative of the actual benefits of said architecture employed in actual hardware. The cost, size, power, and speed are not realized until the hardware has been manufactured, validated, and then tested.

I applaud your attempt at downplaying the actual benefits of new computing architecture that industry wasn't sure would be a viable solution until Brainchip proved it could be done and commercialized it.