I guess we disagree on what innovation is. I define innovation as technology that either no one has, or capability to do things for a lower material cost. For

I guess we disagree on what innovation is. I define innovation as technology that either no one has, or capability to do things for a lower material cost. For example, lets say I have a server rack at home. There is nothing in AWS that exists that I can't replicate on my server rack at home. The only difference is that time required to implement all of it on my server rack vs in AWS, which is optimization. If AWS comes out with some closed source ML model that is good at doing something, thats innovation - i can't replicated that at home without knowing specifics. Likewise, if they have super cheap EC2s with RISC or dedicated ML hardware that are cheaper per hour to run than anything I can build at home with power costs, thats also innovation. The interviews aren't fully useless, I agree that leetcode provides valuable insight. However IMO it should be the bare minimum.