I mean, these are topics that have been discussed countless times over the years and in some cases decades.
It's all well and good to say we need generalizable machines, and something other than backprop, and something closer to traditional programs, but we all know this. The issue is that no one knows what this would even mean, never mind how one would go about implementing it. In the few cases we do know how, the results are horrible compared to the methods we already use.
We use the methods we do today because they work, not because we think they are the best, or because we don't understand the limitations of our models.
True, there's been discussions, but from what I've seen it's mostly flag planting or vague pop-eng fodder that project directors dish out to tech journalists. Having Keras make a statement on this carries far more weight, because fchollet is not selling a product, or pushing an agenda, or creating a walled garden of some sort.
The only thing that's a bit off about Keras is that it's mostly the efforts of one guy. Sure, there's many other contributors, but they don't seem to be acknowledged. I've never seen anyone else speak for the project. I'd really like to see a neutral party emerge for deep learning practice and tooling, before the whole industry gets sucked into a single dominant ecosystem like AWS.
It's all well and good to say we need generalizable machines, and something other than backprop, and something closer to traditional programs, but we all know this. The issue is that no one knows what this would even mean, never mind how one would go about implementing it. In the few cases we do know how, the results are horrible compared to the methods we already use.
We use the methods we do today because they work, not because we think they are the best, or because we don't understand the limitations of our models.