Author Topic: AI and Robotics: new opportunity areas?  (Read 195 times)

The Gorn

  • I absolutely DESPISE improvised sulfur-charcoal-salt peter cannons made out of hollow tree branches filled with diamonds as projectiles.
  • Trusted Member
  • Wise Sage
  • ******
  • Posts: 21452
  • Gorn Classic, user of Gornix
    • View Profile
Re: AI and Robotics: new opportunity areas?
« Reply #15 on: September 04, 2017, 02:40:56 PM »
Gorn, serfdom is good if you're making money and treated ok at a job.  Not everyone wants anything more than that.  I always had good jobs until I got older.  After that, I got to pick from the trash.

I agree with that thought.

Except that after you were at your career peak, and this applies to me, too: the "peak" of a typical IT career is so much less today than it was 20-30 years ago in terms of respect and real earning potential.

In other words it's pot luck to identify a technical specialty that will pay off, and THEN the payoff isn't even that much today.

And that picking through the trash phase can come much faster today than it did a decade ago.
Gornix is protected by the GPL. *

* Gorn Public License. Duplication by inferior sentient species prohibited.

unix

  • Trusted Member
  • Wise Sage
  • ******
  • Posts: 3251
    • View Profile
Re: AI and Robotics: new opportunity areas?
« Reply #16 on: September 04, 2017, 03:48:28 PM »
I can see all technical devices around us becoming smart.

Your fridge. There is no reason you cannot control it via a remote app. An app that will monitor temp but most importantly, the contents. and come up with a shopping list the next time you shop.

Your security. A remote app that controls cameras around your premises and notifies you.

Your car. apps built in to monitor everything. An app to unlock the car door, the house door perhaps and much more. so you have a log file who came, who left. An app that automatically turns off AC if there is nobody in the house.

I can see a lot of devices going smart in the next generation.  From the common household bulb to the fridge to whatever you can think of. I think the old ways of doing things are obsolete. Who wants to carry around a set of keys for your car and your house when all you need is a password? Or a 2-factor authentication like a finger print + a password.

Same goes for banking.  Punch in your account number, then fingerprint and password.  Or something even simpler than that, but more secure than what we have now.

Or an RFID chip embedded in your phone that acts like a wallet and a key for everything (but still needing a password if it's stolen).

Now, what jobs will all the generate, you tell me.  I don't know if that's what you mean by AI.

I think servicing a dishwasher in the year 2030 won't be like doing so in 1980.

Brawndo. It's got what plants crave.

ilconsiglliere

  • CCF Winner's Circle - Supporter
  • Wise Sage
  • *
  • Posts: 2634
    • View Profile
Re: AI and Robotics: new opportunity areas?
« Reply #17 on: September 05, 2017, 10:06:54 AM »
We'll have to clean up talk like referring to offshore labor as tards in the new site...

 :D :D :D :D :D

Kidding!

Just like any boom time economy, the Indians thought the gravy train would run on forever. That's as simple an analysis as it gets. Capitalist economies all experience booms and busts.

I'm personally happy the sector is experiencing pain.  It's well deserved for ruining and then nepotistically running IT in the US.

I could have easily just put Replaceable IT Tards :) . LOL.

I have a background in AI as I did it in college and than in defense. I know way to much about it. Device intelligence like what unix posted about fridge, security, your car and other household appliances is not really true AI. Early AI was glorified rule patterns that were like this:

If ABC = XYZ than
   Do  DEF
Else
    If GHI than
        Do DEF

This is mostly what these devices are doing. What has held AI back in the past was computing power and cost. The computing power was not there and even it was the cost was astronomical. On top of this the languages used for AI tended to be quite bloated and required tons of computing power. You couldnt really have one or the other. Where AI starts to shine is software that learns patterns as it goes and starts building out its own rules. Over time it just keeps learning different patterns and putting in more and more rules. The problem in the past was the rule structure keeps getting bigger and bigger and bigger. Eventually it becomes far larger than the hardware can handle.

This is not really the case now. Hardware is dirt cheap so all kinds of things are possible that werent before. The end result is you are starting to see device intelligence that wasnt possible before. Now you can just keep scaling the computing power.

All this stuff now of being able to control the lights, heating, fridge and whatever else in your house remotely IMO is quite dangerous. I have seen locks you can unlock via your iPhone. Do you honestly want your house remotely controlled so that a hacker can lock you out of your house and take control of everything in it? No thanks. The same thing applies to cars - do you really want a car that the hackers can come visiting when they want. Again, no thanks.

Now true AI intelligence where the machine makes its own decisions about things it can do is a ways off. I have read quite about the chess computers that have beat the chess masters. Those machines have beat the masters via brute computing power, its not that they are actually intelligent. They just have huge lists of rules. Stuff like the movie Colussus is a long way off.

https://www.youtube.com/watch?v=iRq7Muf6CKg

As far as AI being the next big thing for IT tards, I would not take that bet. They said solar and 3D printing was going to be the next big thing for IT as well. All the solar is now being made in China now, 3D printing - eh I dont know.

If you want to stay in IT I am of the opinion you need to find a niche that the H1B hoards have not invaded or cant invade.  I have read that functional programming is a growing thing:

https://en.wikipedia.org/wiki/Functional_programming

But its very difficult to learn which should keep the script kiddies and idiots out of it. Or you can focus on something they cant do, I took the latter approach and focused on human skills that cant be easily learned from a book and there is no cert for it.

The Gorn

  • I absolutely DESPISE improvised sulfur-charcoal-salt peter cannons made out of hollow tree branches filled with diamonds as projectiles.
  • Trusted Member
  • Wise Sage
  • ******
  • Posts: 21452
  • Gorn Classic, user of Gornix
    • View Profile
Re: AI and Robotics: new opportunity areas?
« Reply #18 on: September 05, 2017, 10:40:51 AM »
I have a background in AI as I did it in college and than in defense. I know way to much about it. Device intelligence like what unix posted about fridge, security, your car and other household appliances is not really true AI. Early AI was glorified rule patterns that were like this:

If ABC = XYZ than
   Do  DEF
Else
    If GHI than
        Do DEF

This is an expert system. Rules - based automation. Essentially you have access to higher level testable conditions (IE, you're not testing program variables as such) and you trigger real world actions based on the tests. I never worked with expert systems but I recognize the basic concept.

If Trump wins 2016 election than
   Do  Pepe_memes
Else
    If Hillary wins 2016 than
        Do Nuke Haiti


The other major branch of 1980-90s AI was neural networks, which had the attribute of fuzzy logic. Today you have many models of counter top rice cookers that use fuzzy logic (I guess it avoids burning the rice.)

A good short definition here at the top o' the page: https://www.google.com/search?q=neural+network+fuzzy+logic&ie=utf-8&oe=utf-8

I surmise that neural networks can "learn" and grow in functional capabilities. Rules based systems are more like compiled software.

In terms of digital computing I don't know of any other AI type tech.

By modern standards of newness, both of these technologies are older than snot.
« Last Edit: September 05, 2017, 11:00:16 AM by The Gorn »
Gornix is protected by the GPL. *

* Gorn Public License. Duplication by inferior sentient species prohibited.

unix

  • Trusted Member
  • Wise Sage
  • ******
  • Posts: 3251
    • View Profile
Re: AI and Robotics: new opportunity areas?
« Reply #19 on: September 05, 2017, 07:03:13 PM »
Interesting points you make.
Brawndo. It's got what plants crave.

benali72

  • CCF Winner's Circle - Supporter
  • Wise Sage
  • *
  • Posts: 2434
    • View Profile
Re: AI and Robotics: new opportunity areas?
« Reply #20 on: September 11, 2017, 03:33:04 PM »
I worked with expert system software years ago on a project. It was great for that particular application, which would have been very hard to program otherwise. But it only addresses a certain set of problems -- the need for expertise where the domain is known and the rules can be precisely expressed.

I think JoFrance has an interesting idea that AI may be at long last bursting out of its previous limitations. I guess we'll see over the next 5 or 10 years.

JoFrance

  • CCF Winner's Circle - Supporter
  • Wise Sage
  • *
  • Posts: 2055
    • View Profile
Re: AI and Robotics: new opportunity areas?
« Reply #21 on: September 12, 2017, 03:10:27 PM »
Japan is doing a lot of interesting work with humanoids that will become care givers to their older population.  I sure wouldn't mind having a robot around the house when I'm too old to do a lot of things.

https://www.cbsnews.com/videos/cbsn-on-assignment-explores-japans-future-of-humanoid-robots/

There is a definite need for this technology in Japan and also, as the Boomer generation gets older, we will have the same need here.