General AI is a "wave of the future -- always has been, always will be" technology. Only when a piece gets broken out from general AI do we get practical benefits. The benefits from applying all that matrix algebra aren't small, though!
Would agree that in this article AI is a stand-in for what used to be called Math, Statistics, and Computer Science?
Our scientific output is an artifice, and it is "intelligent" in its ability to guide our action. To me it isn't then strange to think of the sciences as the "general AI".
I have no doubt through that you'll be able to come up with a reason why each of these examples isn't what you really meant by "human level in creative tasks." Which is rather the point.
Just to be clear, I'm not trying to argue with you in the sense that you convey in that second part of your comment, I just enjoy supposing about this particular subject.
We probably can claim 0.1 or 0.01 on such a scale.
Interesting and left too undiscussed IMO are the economics of productizing what currently goes under the AI umbrella. It is so expensive, and such a big task to automate with even the shard of AI we pursue today, that the numbers in many (though not all) cases don't seem to pencil out unless the benefits can accrue to a few dominant oligopolic providers, from nearly all human-staffed positions where each individual position is largely dominated by a single low-cognitive skill repeatedly applied. But interestingly, lots of physical tasks are left alone, which seems to point to a weakness in that economic model: despite joking talk of the coming robot apocalypse upon jobs, human jobs that require diverse physical and perceptual engagement with the environment are nowhere near common automation.
I see this in a lot of domestic household chores that have yet to experience disruption at the hands of AI. That's at least a 50M-count global market population willing to spend say, $2500-10000 USD per year on a robot that can in the beginning perform the housekeeping chores a hotel housekeeping crew routinely performs for every room. Be trainable on a per-item type basis to put up dishes and cutlery the same way into a dishwasher and back to storage, fold clothes and put them up, and mow the lawn: all highly-repeatable tasks that are only addressed by single-purpose, very expensive "AI solutions" today, tell me that if you are investing into any kind of general-purpose AI, invest long-range, like Uber-scale "point a money-hose at it" long-range.
[1] http://www.rebekkahniles.com/2012/03/word-box-sapience-vs-se...
No physical laws preclude a machine that can operate independently and is more capable than humans at any human activity. If we ever make a machine like that, you won't be able to say that "oh, that's just computer science". Well, of course you could, but you would also have met the original definition of general AI.