Close

Will AI Take our jobs

Postby AVClub » November 28th, 2016, 12:53 pm


AVClub
step two
step two
 
Posts: 64
Joined: November 22nd, 2014, 9:32 pm
I knew this was on the horizon but I just saw a video of the AutoDesk Dreamcatcher program slated to release next year. What do you guys think? Will AI take our jobs? The program seems a bit more engineering focused, but talks about "Designers" heavily. Whether or not that is CAD Designers, ME Designers, etc, who knows. I Imagine this is how Graphic Designers felt when Illustrator and Photoshop were released.

https://autodeskresearch.com/projects/dreamcatcher

Re: Will AI Take our jobs

Postby Dan Lewis » November 28th, 2016, 3:52 pm

User avatar

Dan Lewis
step four
step four
 
Posts: 315
Joined: December 6th, 2005, 5:47 pm
AVClub wrote:I knew this was on the horizon but I just saw a video of the AutoDesk Dreamcatcher program slated to release next year. What do you guys think? Will AI take our jobs? The program seems a bit more engineering focused, but talks about "Designers" heavily. Whether or not that is CAD Designers, ME Designers, etc, who knows. I Imagine this is how Graphic Designers felt when Illustrator and Photoshop were released.

https://autodeskresearch.com/projects/dreamcatcher


I'm not worried about being replaced or becoming obsolete. As AI develops it will be another tool. Designers (all types) and Photographers were thrilled by the arrival and development of vector drawing and image editing programs, CAD and 3D CAD -- more tools, more options. I started my career long before computer anything -- it's way better now.

Re: Will AI Take our jobs

Postby KenoLeon » November 28th, 2016, 10:37 pm

User avatar

KenoLeon
full self-realization
full self-realization
 
Posts: 615
Joined: October 23rd, 2012, 12:22 pm
Location: Mexico City
One of my pet peeves of AI is that the term is just plain wrong, even in literature ( Artificial Intelligence a modern Approach - usually considered a good introduction to A.I.), they somehow lament the term and propose something like computational rationality instead. ( try marketing that though).

One issue is that from the other side of the argument ( Human Intelligence ), in the fields of cognitive science, logic, psychology and neuroscience, there is not a hard,simple or unified definition of intelligence, yet most of us have a definition we more or less agree with in our minds and usually transplant the term to A.I.

So in this case, the threat would come from an A.I. that not only designs the product in the nuts and bolts kind of way, but also comes up by itself with all the other things that a designed product is : A solution to a human need embedded in the current culture, with fine aesthetics, environmental considerations and on and on.

What most of the current and past ( A.I. started around the 1940's) A.I.s do and do increasingly well, is efficiently find solutions by borrowing know-how from the programming,biology and scientific world, but, and this is the crux, it is generally to a very specific problem that is usually known beforehand and it is strictly defined: a weight bearing strut system, is the picture a cat, can I drive or vacuum in this area, how to cut some shapes efficiently etc,etc.

So it seems we are living in a period where A.I.s will raise the bar by providing solutions, insights and variations to design while being guided by designers in a partnership.

The loss of jobs is harder to weight, technological unemployment has been a constant in human society, for the last centuries it has mainly been driven by machines that replace human manual labour, more recently it has been data manipulation,and even more recently services, but long term global unemployment has remained relatively stable (~6-8% ) due mainly to our ability as society to self regulate.

For instance, the other day I was designing the guts of a small electronic thing, it took me about a week and basically it involved providing structural support to a few electronic boards & components & making sure it would fit on the outer enclosure while minimizing material costs, let's say a studio does this at a greater scale with a team of 5 designers, a new software tool appears in the market and provides this same functionality, what happens next?

Well you could fire the 5 designers, or you could realize that now you can crank designs 5 times faster with your existing team, what's more, other studios will be doing the same, so maybe hire an additional designer ? It's a simplistic example, but I believe it shows how the markets could adapt to this type of disruption.
Eugenio (Keno) Leon
k3no.com Linkedin Instagram

"Go where you are celebrated, not merely tolerated"

Re: Will AI Take our jobs

Postby Mr-914 » November 29th, 2016, 8:17 am

User avatar

Mr-914
Moderator
Moderator
 
Posts: 5688
Joined: January 12th, 2004, 7:44 am
Location: Montréal, QC
1. Looking at how horrible those bike frames look on the Autodesk site, I'm not worried.

2. At the end, we are designing for human beings, and have in big advantage in being human ourselves.

Take a look at the big data ad world. It was supposed to be the ultimate in delivering ads to the people who wanted them most. In reality, online ad rates are declining and less and less people seem to be clicking on them. When you consider the amount of data and time spent on ads for such a poor result, I think design, which is much more complicated, is safe.

3. Like Dan says, I think it will be a tool. It would be really cool if I could just click a button in Solidworks and it would propose a fastener arrangement. That's something that I hate working out anyways...
Ray Jepson

"The key to success in this business is to find a boss who doesn't care." - Mike Rowe

Re: Will AI Take our jobs

Postby Cyberdemon » November 29th, 2016, 8:53 am

User avatar

Cyberdemon
full self-realization
full self-realization
 
Posts: 3227
Joined: February 7th, 2006, 11:51 pm
Location: New York
What you shared is really considered topology optimization and it's been around for years.

http://www.3ds.com/products-services/si ... imization/

Graphic designers were not "scared" when photoshop came out. While some people may "Belittle" a tool in it's early stages, most designers realized their would be a point where there workflow had to adapt new tools. The same way there was a tipping point where film photographers, who claimed that digital could never reproduce the quality or resolution of film said "nope, I was wrong, this is better now".

Advanced FEA/optimization tools like this will never design a pure end product. As Industrial Designers your job isn't just to design a mechanically optimized structure, your job is to take into account aesthetics, design for manufacturability, design for disassembly, the use of multiple material types, etc.

Topology optimization is currently great for visualizing concepts if you're designing suspension components, airframe assemblies - or other single material, function driven components.

As long as you don't say "bah I don't need them computers!" in 5 years then you'll be just fine.

Re: Will AI Take our jobs

Postby AVClub » November 30th, 2016, 1:14 pm


AVClub
step two
step two
 
Posts: 64
Joined: November 22nd, 2014, 9:32 pm
Good discussion guys! Now that I have had some more time to think this over, I think the technology like most of you have said could be a cool tool, but is really nothing more. Part of this is that the technology from my understanding uses existing archetypes and data points, where as much of what we do as designers is reliant on progressing from the past. And there is truth to it being considered topology optimization, something I know Ross Lovegrove (among others) has been using extensively for some time now.

Re: Will AI Take our jobs

Postby MK19 » December 1st, 2016, 12:28 pm


MK19
step two
step two
 
Posts: 67
Joined: August 19th, 2016, 10:13 am
Generative design is not necessarily Topology Optimisation. Topology optimisation optimises a given topology, i.e. you have to give it a 3D form to start with. Generative design builds variable and numerous different 3D forms which can meet the requirements of the load cases given.
That said, and as previously mentioned, it is not AI in any sense. It is just a mathematical model which has been defined by humans.
There are massive potentials for AI in CAD programs, for example in giving guidance for Design for Manufacture aspects, or for pulling in ergonomic data in order to optimise the size and form of components. Etc.
Potentially, in my opinion, AI could replace a lot of product designers. Only those on the real forefront of understanding psychological and emotional human needs and desires are irreplaceable for longer. Products which are functional, aesthetically targetted to the intended user group and affordable (with mechanical design, market research, material selection, manufacturing cost optimisation, etc all being possible in the near future with AI) will be easily designed by true AI (not mathematical models).

Re: Will AI Take our jobs

Postby Mr-914 » December 5th, 2016, 8:34 am

User avatar

Mr-914
Moderator
Moderator
 
Posts: 5688
Joined: January 12th, 2004, 7:44 am
Location: Montréal, QC
This thread reminded me of a paper I read (and posted on C77!) in 2006/2007. It was an OCED report on manufacturing. It showed that the real number of manufacturing jobs worldwide was falling, including China. I'm sure this trend has continued since.

Despite losing manufacturing jobs to robots, automation and better management, global poverty has continued to rapidly fall and the global economy seems at least stable.

More importantly for design, I think there has never been a bigger variety of products. This will give us a huge market for our skills for the next 20 years at least. Even after that, computers are good at synthesizing existing data. I wonder if they can make the kind of abrupt shifts that humans can (like grunge music or psychedelic graphic design)?
Ray Jepson

"The key to success in this business is to find a boss who doesn't care." - Mike Rowe

Re: Will AI Take our jobs

Postby design-engine » May 5th, 2017, 4:29 pm

User avatar

design-engine
full self-realization
full self-realization
 
Posts: 702
Joined: December 14th, 2004, 4:27 pm
Location: United States of America
AI assists Engineers now exp with FEA work. I could get specific if anyone cared. My thought on this is that design will be the last to go with artificial intelligence.
Bart Brejcha
DESIGN-ENGINE|EDUCATION

http://design-engine.com/category/courses/ SolidWorks surfacing classes


Return to software and technology