I don’t know that I would have linked to this piece in the Times today, as its points are neither new nor profound, but in the authors’ discussion of “Google Duplex,” there was this line:

Schedule hair salon appointments? The dream of artificial intelligence was supposed to be grander than this

Ouch.

cordlessly,

…of machine sentience. Of course the Wikipedia entry speaks merely of “a test of a machine's ability to exhibit intelligent behavior” (emphasis added), and perhaps I’m splitting hairs here, but I draw a distinction between exhibiting something and manifesting it. I think that during the next few years we are going to see astonishing prodigies of machine intelligences persuasively simulating “intelligent behavior” in contexts increasingly broader, but to the extent that these feats manifest sentience, it will be the sentience of teams of gifted and hardworking programmers and engineers.

This being said, I do not in the least deprecate what these teams have accomplished. I’m inclined to agree with the authors of the referenced article that this salient of research and development will not itself lead to machine sentience, but I believe that it will prove ultimately to be one of the major tributaries feeding into the realization of that goal, should human industrial/technological civilization endure long enough to produce this. And I do believe that the goal is feasible, that there is nothing that meatware can do in this line that can’t be matched inorganically.

It’s odd to be approaching a decade, the 2020s, that a few of us here are unlikely to ride to its end. I’ll never know the Martian equivalent of Neil Armstrong, but I’ve seen detailed pictures of what Neil—or Nell—will see. I’ve looked upon images of the gas and ice giants, to say nothing of Pluto, that planetary astronomers of the 1950s would have given their left nuts to examine. And while I don’t expect I’ll see a better-housebroken version of HAL 9000, I believe that (with the caveat noted above), something very like it will emerge before my youngest collateral relatives have reached middle age, and I’ll predict further that this machine intelligence, likely a distributed emergent phenomenon, will have existed for a few years before anyone becomes aware of it.

cordially,