A lot of degrees of freedom for the algorithm to develop strategies within a framework motivated by human decision-making ... not quite "Human-like AI" imo but still great at playing Go

Yeah, I haven't seen actual intelligence so far. And currently everyone and their mother calls their rigid voice input/bot interfaces AI.

Child's play. Computers have been better than us at finite tree games for ages. As they get more powerful they can go down the trees faster. No big deal.

Real intelligence will be inferencing across the whole web. Not just the googleweb. The shared web, your personal web, the public web and content. And tailoring an optimal web experience for you. Actually for your brain chemistry.

This is not tree based thinking. It's graph based. Where every statement on the web is a peer and yields a first order logic of productions and new inferences. These inferences become highly connected by spontaneously linking to other inferences and forming new unexpected structures. True intelligence is an emergent property from highly connected systems. It cannot be made in a lab.

Yes indeed. The web of logic will blow you mind. Here's a taster :

http://weboflogic.com/

I largely agree, but I wouldn't say "child's play" about this one. Go numbers are so large that calculating solutions isn't enough to win a game. That's why chess is a completely solved game, but Go has been and continues to be a game where human intuition (i.e. human intelligence) has an edge over computers.

Sure, it's a very big tree. And also the style of play is quite zen-like and counter intuitive. With chess you have a rough points system that normally lets you know where you are, with go it's alot harder because everything can change so fast. But that simply means you need to discover. or auto discover new metrics for scoring. Seems they've done that. Dont get me wrong, it's quite a feat. But it remains a narrow rule based topic which is where computers tend to excel. When the rules are thrown out the window and order needs to emerge from chaos using feedback loops (ie similar to Darwinian evolution) then you'll get whole new branches of problem solving and even problem spaces. ie the solution to games without a rule book, will make what came before seem primitive.

Amen. :)