re: AI: The future of humanity and artificial intelligence

I am by no means an expert but started reading into it a couple of years ago. A problem with the idea is that it conflates computing power with actual intelligence or sentience. So while Moore's Law has been somewhat reliable (at least until now), advances in software have not necessarily been made in a similarly exponential fashion. Moreover, machine learning itself, while closer to human cognition, does not entail sentience. An example I have mentioned on here before is that Watson may win Jeopardy but it has no idea it won...same for AlphaGo.

Getting to a level such that a machine enjoys human-level intelligence would be an unbelievable step change vs. just being able to pwn humans in Chess or Go (or install tires on a Tahoe for that matter).

Still, an interesting topic. We just have no way of knowing how go predict something like this, and it's odd that your Verge and Kurzweil types coincidentally point to the event horizon falling within either of their respective life spans.