Ann: Weebit Nano and the Technion collaborate on ReRAM project, page-29

Not downplaying the announcement, I am ecstatic they are doing these things.

BUT, nobody else is going for the heart of big opportunity right now, they are all pussyfooting around.That is because dumb-a MBAs are trained to over analyse markets these days and they end up niche targeting, often losing perspective on whether their product can actually go for the broader market. Ours can potentially.

Ours is the simplest replacement for standard flash available. We should not be subsetting our markets when we have the most general purpose solution. I dont want this being classified purely as an AI chip, because it isnt just that. Certainly proving it solves a wider range of use cases than other comparable product adds weight to the claim but it should not be the goal this early. Widely applicable is much better than narrowly defined.

If the GPU guys remained saying there chips were only for graphics processing they would have missed massive opportunities by not selling to the more general computational uses.

The principle of "delayed binding" is important here. Do not define your market, let the market do that.
I still believe the next generation general NVM replacement after flash is a market there for the taking. Most other rerams have not taken a swing at it, partly I believe due to their complexity and manufacturing cost.

We should because we have to product to do so.