I also think neuromorphic computing is the future for energy efficient AI. But sadly development is getting a bit sidelined by other popular AI research areas. It doesn't help that it's much easier to train ANNs than spiking neural networks (SNNs). :/

I think neuromorphic might even be a distraction right now.

The hyperparameters for a SNN that performs a specific task are extremely elusive. If you don't even know what kind of neuron model or fan out ratio might work, how the hell can you start burning these as constants into some hardware contraption?