Brain Computer Interfaces are what will give us AGI, not sure if you've kept up with recent developments. But Mary Lou Jepsen of Openwater (Google it) has developed a BCI that can scan the entire brain with infrared light, it can read neurons at the submicron level. It can literally scan your thoughts and as your axons are synapses are firing off. Her device is already ready and the dev kits are set to release this year. She has already beat Elon Musk to the finish line.
We can copy/paste the human brain's architecture and combine it with deep learning(which has recently defeated top pros at Starcraft II and DOTA II even without BCIs), not sure if you've been keeping up with recent events, but AI has been going through a renaissance. Don't be discouraged because the Tinnitus world is moving slow, because a lot of other fields are moving very fast. AI is having a breakthrough every couple of days. And Brain Computer Interfaces are about to combine with deep learning. One field of science helps another. BCIs help will help an already fast moving AI field reach it's end goal faster than expected. These general learning algorithms can then be applied to solving all different kinds of problems humans currently struggle with.
You solve intelligence, you solve everything else. If we get a general learning algorithm in the next few years, consider everything cured. Because it will be far more intelligent than the entire human race combined in a very short amount of time.
We will have generalized learning algorithms by the middle of next decade. After that, human research doesn't matter.
Quite frankly, I agree with Elon Musk, we need BCIs here quick just in case AGI gets here without them. Humanity has no choice but to merge with it.