I don't think that style of interview provides as much signal as you think it does. I would probably say that watching someone work through a problem provides

I don't think that style of interview provides as much signal as you think it does. I would probably say that watching someone work through a problem provides more signal than having someone explain a problem they have already solved. That's why I feel that all of the companies I've been at, eventually ended up shifting the interview process to do less "explain a project" and converging more on "talk us though these different types of problems". And then the past experience is what is being demonstrated when they show off how deeply and quickly they can think through the different types and bring their experience to bear. I also just generally do not think that most people can even put together a 2-3 hour presentation of their past work that goes over well. Most of the time people can't really talk more than 30 minutes about past projects. That requires a whole different set of skills, which there are certain contexts where I'd value that more highly. But my initial impression is that if we set up our interviews this way, we'd also end up filtering out a lot of people who'd be great, since not picking the perfect project to demo basically dooms the entire interview to be a flop. With multiple interview types, you increase the chance that there's an interview they really shine in.