What's new
What's new

Interesting article on AI / Robots and future employment

Job schmob. Behold the Utah Data Center.

aerial-utah-data-center.jpg


The Singularity isn't about jobs.

.
 
I can't seem to read the article without logging in.

The problem I see is that this "AI" seems to be a pant load harder to get working than everyone keeps telling us. We've been waiting for robots to "take over" for almost 50 years. It just doesn't seem to be happening. We're still a very long way from a robot replacing a human.
 
I can't seem to read the article without logging in.

The problem I see is that this "AI" seems to be a pant load harder to get working than everyone keeps telling us. We've been waiting for robots to "take over" for almost 50 years. It just doesn't seem to be happening. We're still a very long way from a robot replacing a human.

There is a dialogue box for signing up for e-mailed news articles but you can just close it with a button on the upper left corner of the box and read the article.
 
I can't seem to read the article without logging in.

The problem I see is that this "AI" seems to be a pant load harder to get working than everyone keeps telling us. We've been waiting for robots to "take over" for almost 50 years. It just doesn't seem to be happening. We're still a very long way from a robot replacing a human.
Robots have already replaced 10's of thousands of workers in this country alone. The company I worked for 1st started using them 28 years ago and within a dozen years hand many of them. Example: between each large press on a 4-5 press line making automotive stampings there would be 2 workers removing the stamping off the lifters. They would then transfer the part to the next press for more operations. You would have 8-10 workers involved in that line prior to the robots and now they 2 workers removing the panel from the final press to rack. In between the presses they switched to robots and when a different panel is to be run the tooling head for the new panels are changed. Where workers use to spot weld a robot will now do the work. There easy to train, don't take breaks, don't require holiday pay and work for peanuts.. In the 90's we were paying around $75k for a ABB robot and it paid for itself in 1.5 years.
 
Robots will take your job - The Boston Globe

Particularly of interest are comments related to the loss of jobs due to technology and proposed solutions to provide for the unemployable.

Lets see 2033 i will be 73 years old...maybe i will have enough to retire. maybe invest in some AI companies today? Interesting perspective...Let the kids know, maybe there is money in arts in the future. Or do you think emotions will be taken over by machines as well?

dee
;-D
 
Robots will take over the easy stuff and have in may industries. It's entirely different to build a robot that can do what a human can do like feed a child, climb stairs, cook dinner, or clean the garage. We're a long way from those things being done economically and reliably by robots.
 
It seems to me that AI & Robots eliminate some jobs but create others. The fly in this ointment is that the jobs created require a higher degree of human intelligence, skill, & versatility, which leaves the people who were only capable of doing the "lesser" jobs unemployable. A certain former President of the United States stated "we will retrain the people whose manufacturing jobs are off-shored for service jobs (burger flippers)". As I see it, when everybody is working at flipping burgers, nobody will be able to afford to eat out.
 
From the article: "smartest way to distribute the wealth is by implementing a universal basic income."
There is a name for that isn't there? Not seeing a big trend towards it.

Interesting the "new" interest and buzz about AI. Mention that you were building a company around neural nets or expert systems in the 80's and investors would throw money at you faster than you could spend it.
It was the "darling" of wall street. Now perhaps enough time has passed that those who lost all their money are not the ones investing today.
Pick your winners carefully and learn from history. This is not the first time the media projected exponential growth. At one time people were ga-ga over Eliza.

While I'm not impressed by the Go game player it is a different world and time. A net training set that took 48 hours to run now will run in 4 minutes or less allowing for a lot more "tweaks".
Bob
 

Attachments

  • UAWposter.JPG
    UAWposter.JPG
    47.2 KB · Views: 224
i heard that NC will put all the turret lathe operators out of work and riots will ensue. we must fight this.
 
The auto industry at the top level has been automating both hard and then soft basically forever.
Big. big push in robots and such in the 80's. Insane amounts of money spent and big expectations.
The floor people in highly automated areas ended up having much easier jobs and the biggest paycheck so that seemed very good.
I built a system that automated one lifelong friend's wife out of a job and bumped her back to the line. That seemed very bad.

Then along came Japan and Toyota in particular. They opened their plants and production model for us to come visit.
Turns out, despite the thought to the contrary Japan was miles behind us in robots. They were not investing the money we were here.
Within three years a hundred or so small robot/computer automation companies would file bankruptcy.

Interesting now that the pendulum is swinging back.
What do we do with all our "worker bees"?
IMO we as employers have a duty to teach them the nifty new shit way past what is needed to "get the job done".
This of course means investing real and serious cash into your workforce. For sure not good short term and maybe impossible if cash strapped, maybe good long term.

You bring in a kid and he/she knows nothing about feeds/speeds or the difference between carbide and HSS
You teach them this. Why can't they now learn how to make a robot balance a broomstick, navigate the shop floor moving parts or play ping-pong?
Mostly I see much of underestimating our "worker bees" capability and not wanting to invest in them in the interest of short term profits
This is just as true in small shops as it is in big multi-nationals. Perhaps worse in the small/medium sized ones.
(Please do not get wound up with the term "worker bee". I use this with very much respect, more a poke at those who for some brainless reason feel above this level)
Bob
 
I think in some cases it depends on the definition of a robot.

http://www.nytimes.com/2016/02/28/magazine/the-robots-are-coming-for-wall-street.html?emc=eta1&_r=0

Technology is advancing and being deployed in ever more sophisticated ways.

We are developing a grinder that is driven straight out of solid works with no additional cam system . . . It uses an acoustical sensor for touching off on the wheel and proceeds to grind a complex 3D profile holding tenths. It learns the wear rate of the wheel by continually checking the wheel and can dynamically adjust the workpiece offsets to ensure accuracy across a run of 1000's of parts.

Tended by a robot and setup over a network by the engineer who designed the part being ground.

There will eventually be a half dozen such grinders that will replace a total of 18 less sophisticated grinders and the operators that run them will move onto other tasks in the company.
 
I think in some cases it depends on the definition of a robot.

http://www.nytimes.com/2016/02/28/magazine/the-robots-are-coming-for-wall-street.html?emc=eta1&_r=0

Technology is advancing and being deployed in ever more sophisticated ways.

We are developing a grinder that is driven straight out of solid works with no additional cam system . . . It uses an acoustical sensor for touching off on the wheel and proceeds to grind a complex 3D profile holding tenths. It learns the wear rate of the wheel by continually checking the wheel and can dynamically adjust the workpiece offsets to ensure accuracy across a run of 1000's of parts.

Tended by a robot and setup over a network by the engineer who designed the part being ground.

There will eventually be a half dozen such grinders that will replace a total of 18 less sophisticated grinders and the operators that run them will move onto other tasks in the company.

It really is about the AI and less about the robot. Thanks for posting the article. It gives an interesting perspective into possible future employment.
 
.....
We are developing a grinder that is driven straight out of solid works with no additional cam system . . . It uses an acoustical sensor for touching off on the wheel and proceeds to grind a complex 3D profile holding tenths. It learns the wear rate of the wheel by continually checking the wheel and can dynamically adjust the workpiece offsets to ensure accuracy across a run of 1000's of parts.
......

I just have to ask. :willy_nilly:
Check the wheel or check the parts coming off? No machine defection as the wheel loads and dulls before the next dress?
Realtime checking of the wheel (or cutting tool dia.) gets me in the ballpark but has not been able to control the process itself tight enough.
For me no matter how big and stiff the iron or how small the feedback loop increments, everything bends as the tool pressure changes.

Just cycling the part through twice without unclamping will often blow my entire tolerance limit so the machine "learn" a back-off on a rerun cycle and this back-off is different on varying parts.
It's not really a "smart" machine but it does sort of "think" about what the output gauge tells it and tries to find a good average number to use.
It also knows and tries to adjust this number differently at 2 pieces on the part counter into the dress vs. 500 on the part count.

As another aside but perhaps more to what do we do with our workforce.
How many people will these machines you are building put out on the street? There is only a fixed amount of dollars.
Will your check come from increased market share or less people getting paid? If you increase market share isn't somebody loosing a job in another plant?
Everything you automate or make run faster means bad news for somebody somewhere.
How do you feel good about this and what you do eliminating jobs? What happens, what do you think when you haved unemployed someone close to you in the family?
Not so bad when faceless people.... close relatives and good friends, been down that rabbit hole once or twice.

Are you the good guy or the bad guy? Does this produce sleepless nights?
Bob
 
I just have to ask. :willy_nilly:
Are you the good guy or the bad guy? Does this produce sleepless nights?
Bob

You know - I do think about that question often and I am not sure I know the answer, but it doesn't keep me up at nights nearly as much as wondering how to keep an ever growing payroll of people busy pushing manufacturing technology forward.

We use a schmitt industries acoustical sensor to measure the wheel directly. The part fixture is mounted to a 3-axis trunnion (X-Axis linear motor with 1/20um feedback, A and B axes with 10 arc sec repeatability, 30 arc sec accuracy) and the spindle is mounted to the Z-Axis linear motor also with 1/20um feedback.

Grinding wheel is a 350mm diameter x 100mm long cylinder (abrasive with polymer binding) with speeds from 30m/s to 70m/s depending on abrasive and binding. Spindle wound up being our own design using 120mm ID angular contact bearings preloaded against each other and labyrinth seal arrangement.

Acoustical sensor is on the fixture and measures OD in X and face distance in Z looking for a specific acoustical signature based on touching off a sacrificial carbide which wears on the wheel less than 0.0001" per 10 touches.

Our prototype is working and we are on to the next prototype stage where we build a unit that will run in production for six months at the customer facility to learn what we can learn and then we tweak where needed and build the production units. A lot of stuff to develop all at the same time from the controls, to the mechanisms, coolant with variable speed pump to attempt to match surface speed of wheel, different abrasives, different materials being ground, having as few operator tweaks as possible, and getting all this technology to be as simple as possible for someone to set up and put into production - very software intensive!
 








 
Back
Top