I agree but I am more pessimistic about it. There are 2 inter-related factors that will stop AGI from occurring at the level we all dream of: Competition and Specialization. An AGI is not something that will occur in a vacuum. It will be developed and created in the world we all are living in. The one with dirt and mortgages and greedy people. More importantly, it will come about in a world where half of the people are dumber than average. The AGI does not need to solve the Riemann-Zeta hypothesis. It just needs to afford to have it's owners/creators bills paid and maybe make some on the side for profit. That is all that it needs to do. Yes, it can do more, a lot more, but it only needs to be a little bit smarter than the average human so that it's owners can scam us by proxy. Look at the Ashley Madison hack. You had real human men paying real world dollars to suss out if they were talking to a bot or not. That's all that AGI really needs to do to sustain itself over a short time interval. How long a time interval do you want it to go for will then add the complexity and get it to 'human' levels of general intelligence as it tries to deal with rarer and less common failure modes (like itself getting hacked in the Ashley Madison case). AGIs will have to fork out of the super-targeted AIs that we now have that make monetary sense. That will take a lot of time and money to do this, and the returns on the investments are uncertain to me.