4 ms·

Andrew Ng: What Artificial Intelligence Can and Can’t Do Right Now

35 points by seycombi 10y ago

ilaksh 10y ago
I guess because business people tend to be confused by hype or anti-hype and sometimes unable to care about five years from now this is a realistic message. But he really emphasized the supervision and input -> output parts which is a bit of an oversimplification. There is definitely less supervision than before and demos from groups like Deep Mind or OpenCog that are powerful interactive agents.
- mojomark 10y ago
  "...the amount of work we can automate with AI is vastly bigger than before. As leaders, it is incumbent on all of us to make sure we are building a world in which every individual has an opportunity to thrive." Andrew Ng is cerainly an AI liminary, but the above excerpt begs for a refreshed discussion summarizing leading ideas on what this new "AI-infused" world looks like. In a 2014 Bloomberg article, Ng stated "There will always be work for people who can synthesize information, think critically, and be flexible in how they act in different situations."(1) This logic worked for the Industrial Revolution (IR), and I'm sure it will certainly hold true for our lifetimes and perhaps a few generations onward. However, while steam engines and special purpose machines of the IR drastically increased productivity, displacing many physical labor jobs, humanity still carried the reserve of our intellect to fall back on in a post IR world. There was still a largely untapped reservoir of intellect and creativity. I suspect the transition to an AI-infused world will occur gradually over a few generations and my great grandchildren will have a much different sense of nastalgia and what it means to live a meaningful life in a world where people gladly receive a base wage and spend the day consuming the latest computer generated song, novel, or scientific white papers. A world where computer creations arrive at a pace and quality faster and higher than humans could never compete with, so why bother? In my current mindset, however, that just doesn't seem like a fun world to live in. Every single thing I do being far inferior to what a computer can do. If I were transplanted in that world, I think there would be many products where I would draw the line and opt for a good portion of non-AI generated products. Like a modern day Amish person I suppose. 1. https://www.bloomberg.com/news/articles/2014-03-12/your-job-taught-to-machines-puts-half-u-s-work-at-risk https://www.bloomberg.com/news/articles/2014-03-12/your-job-...
jstnstwrt 10y ago
ai community needs to put out more clear writing like this to help educate business world on what is possible and what isn't.
- dilemma 10y ago
  The AI community is the source of false information. They will never tell you that AI is fundamentally impossible because a machine cannot think; it can only do as its told by its programmer. That is the bottom line.
  - gumby 10y ago
    How is that different from a human or a dog?
    - hood_syntax 10y ago
      Creativity imo. We can choose to create novel things, with intention, that fulfill a use.
      - gumby 10y ago
        Dogs are creative as well, though not as much as humans. The comment I was replying to simply asserted that people have something a machine inherently cannot. I wonder what that thing is, and if the writer thinks it's a machine-vs-nature issue or a humans-vs-all-other issue. FWIW I don't believe the assertion in the slightest. I do believe it is quite likely to have an AGI that does not recognizably exhibit conciousness but that, to me, is irrelevant.
        hood_syntax 10y ago
        That's fair, I understand your stance on that, and definitely agree with you on the AGI.
    - dilemma 10y ago
      Humans aren't programmed by another human.
agentofoblivion 10y ago
Refreshing. The hype around AI/machine learning in places like here and TechCrunch is exhausting. It doesn't help when people like Stephen Hawking and Elon Musk speculate about nightmare scenarios that seem to have both feet firmly in the distant future. We're so, so far away from anything resembling so-called "general intelligence". It's totally nuts to compare clever engineering scenarios like Google's Go project to general intelligence. It just isn't. My theory is that paid per click advertising/clickbait is making it increasingly difficult to say anything that gets attention on the internet, so people increasingly use hyperbole. "Internet journalists" will trump things up to make it sound more exciting, feeding the hype machine. Those that don't succumb to the practice are just not heard. There are, of course, exceptions. But a well-reasoned, calmly stated argument from a nobody is never going to go viral in the way that an outlandish story from that same nobody could. i.e. Seth Godin doesn't have reach just because he's interesting, it's also because he's been working at it for over a decade, but any idiot can make up some fake story that goes viral and impacts the public consciousness.

35 points by seycombi 10y ago

I guess because business people tend to be confused by hype or anti-hype and sometimes unable to care about five years from now this is a realistic message. But he really emphasized the supervision and input -> output parts which is a bit of an oversimplification. There is definitely less supervision than before and demos from groups like Deep Mind or OpenCog that are powerful interactive agents.

mojomark 10y ago
"...the amount of work we can automate with AI is vastly bigger than before. As leaders, it is incumbent on all of us to make sure we are building a world in which every individual has an opportunity to thrive." Andrew Ng is cerainly an AI liminary, but the above excerpt begs for a refreshed discussion summarizing leading ideas on what this new "AI-infused" world looks like. In a 2014 Bloomberg article, Ng stated "There will always be work for people who can synthesize information, think critically, and be flexible in how they act in different situations."(1) This logic worked for the Industrial Revolution (IR), and I'm sure it will certainly hold true for our lifetimes and perhaps a few generations onward. However, while steam engines and special purpose machines of the IR drastically increased productivity, displacing many physical labor jobs, humanity still carried the reserve of our intellect to fall back on in a post IR world. There was still a largely untapped reservoir of intellect and creativity. I suspect the transition to an AI-infused world will occur gradually over a few generations and my great grandchildren will have a much different sense of nastalgia and what it means to live a meaningful life in a world where people gladly receive a base wage and spend the day consuming the latest computer generated song, novel, or scientific white papers. A world where computer creations arrive at a pace and quality faster and higher than humans could never compete with, so why bother? In my current mindset, however, that just doesn't seem like a fun world to live in. Every single thing I do being far inferior to what a computer can do. If I were transplanted in that world, I think there would be many products where I would draw the line and opt for a good portion of non-AI generated products. Like a modern day Amish person I suppose. 1. https://www.bloomberg.com/news/articles/2014-03-12/your-job-taught-to-machines-puts-half-u-s-work-at-risk https://www.bloomberg.com/news/articles/2014-03-12/your-job-...

jstnstwrt 10y ago

ai community needs to put out more clear writing like this to help educate business world on what is possible and what isn't.

dilemma 10y ago
The AI community is the source of false information. They will never tell you that AI is fundamentally impossible because a machine cannot think; it can only do as its told by its programmer. That is the bottom line.
- gumby 10y ago
  How is that different from a human or a dog?
  - hood_syntax 10y ago
    Creativity imo. We can choose to create novel things, with intention, that fulfill a use.
    - gumby 10y ago
      Dogs are creative as well, though not as much as humans. The comment I was replying to simply asserted that people have something a machine inherently cannot. I wonder what that thing is, and if the writer thinks it's a machine-vs-nature issue or a humans-vs-all-other issue. FWIW I don't believe the assertion in the slightest. I do believe it is quite likely to have an AGI that does not recognizably exhibit conciousness but that, to me, is irrelevant.
      - hood_syntax 10y ago
        That's fair, I understand your stance on that, and definitely agree with you on the AGI.
  - dilemma 10y ago
    Humans aren't programmed by another human.

agentofoblivion 10y ago

Refreshing. The hype around AI/machine learning in places like here and TechCrunch is exhausting. It doesn't help when people like Stephen Hawking and Elon Musk speculate about nightmare scenarios that seem to have both feet firmly in the distant future. We're so, so far away from anything resembling so-called "general intelligence". It's totally nuts to compare clever engineering scenarios like Google's Go project to general intelligence. It just isn't. My theory is that paid per click advertising/clickbait is making it increasingly difficult to say anything that gets attention on the internet, so people increasingly use hyperbole. "Internet journalists" will trump things up to make it sound more exciting, feeding the hype machine. Those that don't succumb to the practice are just not heard. There are, of course, exceptions. But a well-reasoned, calmly stated argument from a nobody is never going to go viral in the way that an outlandish story from that same nobody could. i.e. Seth Godin doesn't have reach just because he's interesting, it's also because he's been working at it for over a decade, but any idiot can make up some fake story that goes viral and impacts the public consciousness.