I don't think the betamax analogy applies. Betamax was a consumer product where technical capability is not the foremost consideration. Licensing IP for SNNs is a different kettle of fish where technical capabilities are paramount. SNN IP customers are not going to buy a product because it does not clash with the aspidistra.On darker days I think excatly that... We have the best product, that I am convinced about. But then we also had Betamax vs VHS. Betamax was definitly the better product. And yet...
The questions @SERA2g posted a page back show that the questioners were familiar with NNs but, unless they were Dorothy Dixers) did not know the details of Akida. For instance, one questioner asked about quantization and sparsity. On quantization, Akida started with (and still has) 1-bit weights and activations and this has now been expanded to 8-bit. Akida takes sparsity to a whole new level with N-of-M coding. But because this question was rolled in with several other questions, Nandan did not address it. There was one slide which showed that a comparison system achieved 92% accuracy (I think it was to do with de-noising, but don't quote me), and Akida with TeNN achieved 97% with significantly lower resources.
I was hoping for more such comparisons in the talk, particularly highlighting the capabilities of ARM M85 combined with Akida 2.