It's good to see you here. You've hit on some timely key points, and if you keep at it I think you'll pick up all kinds of influences and nuance as you go alon

It's good to see you here. You've hit on some timely key points, and if you keep at it I think you'll pick up all kinds of influences and nuance as you go along. I've always liked ideas along these lines, that's why I seriously dabbled in ML using scientific instruments as soon as I possibly could, which was not much given the hardware limitations back in 1980. There was no way AI was anything but a futuristic dream, but it seemed like there wouldn't be very much chance of machine intelligence without the fundamental ability of machine learning. And I think it follows that the more actual learning, the more intelligence can result. OTOH, the more intelligence to begin with, the more capable the learning can be. All I could do at the time was concentrate on making the absolute best use of the hardware for the type of ML that would leverage my own intelligence, and it wasn't much. Read that however you like ;) Anyway, with this in mind that's why I thought it would be a good idea to put in at least twice as much effort compared to other natural scientists using the same instruments, to gain and maintain a point of view that might eventually lead to some intelligent performance that wouldn't be achievable any other way. Along with getting my regular work done. That's 44 years ago now, still haven't come close to having the equivalent of 88 years of dedicated experience, but I'm not done yet ;) And in the meantime while waiting for adequate computer hardware (for potential intelligence duty) to come within reach I just focused on advancing natural science in ways that nobody else was doing (or doing very well). Actually have done a bit of computer science as I go along too, it may not be "natural" but it surely is as naturally confusing or worthy of investigation. Ever since the scientific software business arose it's plain to see there has been tremendous room for improvement :\ For me being in the lab business before that happened, it gives me lots of ideas that you don't have people thinking about in the 21st century any more. Especially when there wasn't much back then either :) How many people can recall the feeling of "Hey, we're doing pretty well now but just imagine what it would be like if we had computers!"? With the proliferation of office computers starting in 1982, which were disappointingly weaker than what I had been using to barely do a little ML, and the abundance of technical minds that flocked to the rapidly-growing field where they were direly needed, this left natural science with a skeleton crew ever since. Which is one reason why you need robots to do more experiments, to make up the difference. They just haven't come along soon enough. I knew I'd be way outnumbered by brilliant computer scientists, and I fully expected them to make incredible progress in machine decision-making as soon as real powerful desktop computers became available, so for me I wanted to be as well-prepared as possible to be able to provide a natural complement if a realistic situation were to arise. Plus I had chemicals up the wazoo and electronics to boot. Or fail to boot which pissed me off that I had to waste time on that. When high memory and fast processors showed up in the '90's we were supposed to be there, but the neural offerings at the time were never going to cut it at my company then. And I was a pioneer in computerized lab results. That's been 30 years ago by now, its amazing the improvements to natural decision-making that can be made given enough real-time building on previous accomplishments. Like real-time decades rather than a brilliant short-term project intended to simulate decades. Great name for FutureHouse since this is all still inspiringly futuristic, even after you've been considering the options for a good while.