In fact, I had just read a CVPR '18 paper that did the kind of thing he mentioned--presented an 'interpretable neural network' that assumed what interpretability meant...
I have to say, the impressive scale of this data collection probably means it will still have some uses. And for Chinese health & medicine, it's simply a massive deal, if everything was done correctly.
A Reinforcement Learning machine would have to be re-trained for changes such as that, but not from scratch. Depending on the size of the change, it would take much less time to adapt to such a change.
Several notes:
-The article notes that a lot of the hard parts of human video-game playing have been done 'for' the AI by hardcoding. It doesn't actually have to look at the screen to parse information--it has direct access to game variables that player would have to access through menus. More relevant to your question, a lot of the strategic decisions like item purchasing, ward placement, and probably character placement were just picked by the programmers or set to be ignored. The whole thing is less impressive than the headline sounds. I think it was just learning--where to run on the map and when to attack things?
-RL is based on deep learning, and there are fundamental issues with deep learning's ability to adapt to genuinely new scenarios. None of these system can presently adapt to something genuinely unprecendeted in a time frame you would consider 'safe'. They need at least several opportunities to observe how the world works after the changes and the consequences of their actions. To try to make this concrete--they can't reason about what implications a flood/power outage/landslide has for their traffic management. They can only learn from trial and error <-(the important part) how traffic behaves during a disaster.
Have to agree that this seems like a lot of name-calling and rhetorical suppressive fire which fails to address any actual arguments.
Giving the devil his due, Kurzweil has been proved wrong on almost every single prediction he's made up to the current day--at least from my reading of his 'Singularity' book. Mostly as a consequence of the bottom falling out of Moore's Law.
But like most articles addressing the singularity thesis, the author completely fails to understand it even in the most rudimentary way. The singularity hypothesis depends on several plausible but not guaranteed steps of cause and effect; the author could have discussed anyone, and, as other commenters have mentioned, there are valid reasons to think the singularity highly unlikely. But like so many bloggers, their thought process only went as far as 'sounds weird to me personally--must be a religion!'
More generally, it's both lazy and extremely dangerous to dismiss an idea simply because it seems unusual to you. This is called the Absurdity Bias, and is fairly well demonstrated.
Just for example, there are comments in this thread saying that while some of the radical predictions of technologists are possible, there's 'also nothing to suggest humans are capable of developing it.' Sounds a lot like something Lord Byron would have said about antibiotics or powered flight a few years back.
In fact, I had just read a CVPR '18 paper that did the kind of thing he mentioned--presented an 'interpretable neural network' that assumed what interpretability meant...