Expert systems of this kind have existed for decades now, and what Watson does could have been done, on a smaller scale commensurate with CPU/memory power of the time, in the 1990's.
It is a question not of system existence, but of performance relative to your average doctor. And once commercially deployed systems would start outperforming your average doctors these high-paying doctor jobs are bound to change somehow.
The answer to that question was a definite no, some 20 years back. And no, some 10 years back. Now, the answer is yes. And in some 20 years it will be resounding yes. Why? First, because medicine gets less and less 'shaman based' and more and more evidence based. Some 20 years back you could't scan a region of your DNA and figure out that you have a particular versions of couple of proteins that wrecking havoc in your particular case. Now you can. Second, there've been improvements in statistical models and natural language processing. To the level that computers can actually outperform humans in these murky areas. An IBM success with Jeopardy is a good example. Combine these two, take into account that computers are less likely to overlook hard evidence (like DNA data, incopatibilities with medications, etc), and you would also want an expert system to look at your health data, diagnose, recommend treatments and monitor the progress.
Either way. I still think that a "med school is probably Ok" could be a good advice to a kid.
I understand that this is HN and people are optimistic about the potential of technology, but betting against technology changing everything tends to be a winning bet more often than the otherwise.
When I was a kid, I wanted to be an aerospace engineer. I read books in the 80's talking about how we'd have manned Mars missions in 25 years. I actually went to engineering school for aerospace, but by then I had realized humans wouldn't go into space in a real way within my career.
Is it possible that some derivative of Watson is going to replace doctors? Maybe. Or maybe it'll just change what doctors do for a living in the same way auto-pilot changed what pilots do for a living.
My bet is that 20 years from now isn't going to look that dramatically different from 20 years ago, in a fundamental way. I remember 1993 (a year before President Clinton's first term and a year before Amazon was founded) and it wasn't all that different. We made phone calls and sent faxes instead of using e-mails, there was no Twitter, Facebook, or smart phones, but life was pretty similar.
Dude, I look at 1993 and compare it to today and am totally amazed at how much life has changed.
In 1993, your circle of friends was pretty much limited to the people in your town. Maybe you had a pen pal, but that was considered to be a rare novelty and took real work to maintain. Now, I have friends all over the world, that I keep up on their day-to-day lives. When an earthquake hit SF in 1989, my reaction was "Oh, that's too bad, I feel bad for anyone who died. I'm sure we'll be talking about this in school today." When an earthquake hit Japan in 2011, my reaction was "Has anyone heard from Jennifer yet? Is she okay? I know she was in Sendai..." And I had friends all over the globe - India, Colombia, DC, Boston, Morocco, LA - asking the same thing.
In 1993, if you were meeting a friend for an event at an unfamiliar place, you called them up beforehand, specified an exact time and place that you would be, wrote down the directions or brought along a road map, and prayed they would be there when they said they would. If you got lost, you had to pull over until you could look at the map and memorize new directions. If you couldn't find each other in the crowd, you were out of luck. If you hit traffic on the way in, you'd hope they didn't think they were being stood up and apologize profusely. Now, your calendar pops open a notification when it's time to leave, taking into account traffic. You hit a button, and your cell phone navigates to the destination, automatically rerouting you if you take a wrong turn. You can't find your friend, you call or text them. You don't know what they look like, you call them and see whose phone rings.
In 1993, if you wanted to start a company, you worked your butt off for a big employer to learn the ropes of your industry. You spread your network as far as it could go. You sought out mentors. You begged for funding. And then you hired lots of people to make your dreams come true, and if your dreams were wrong, they all got laid off a couple years later. Now, you Google [startup]. You read a bunch of blogs. You download lots of open-source software. You build something, you launch it, you post it on Hacker News and a bunch of blogs, and if your dreams were wrong, you've wasted maybe a year of your life.
If your life isn't all that different from 1993, it's more likely because you're still living in the past than because the world is the same as 1993. "The future is here, it's just not evenly distributed yet." For some people, the world is wildly different from what they expected 20 years ago.
