Categories
BW Member Blog

Generative Part Design

With Bill fearing that AI mechanical design will put engineers out of work, I thought I’d post some recent articles on it:

https://www.mmsonline.com/articles/what-this-seat-bracket-says-about-the-future-of-automotive-manufacturing

https://www.mmsonline.com/articles/generative-design-generates-new-interest-in-a-range-of-manufacturing-options

https://www.machinedesign.com/3d-printing-cad/article/21121975/data-ai-and-the-generative-design-revolution

I’ll point out that for now there is NO danger that AI part design will end human engineering.  I doubt it will even reduce the amount of human engineering.  In the same way that automation reduced the cost of items like shoes and automobiles so much that the huge increase in purchasing resulted in a substantial increase in employment in those fields, I would expect that reducing the cost to custom engineer something would result in a huge increase in the amount of engineering going on.  Yes, the computers are learning how to optimize part design.  Soon they will be optimizing assemblies.  But who determines what products need to be developed, and what the tradeoffs need to be in their design?  It will be a LONG time before an AI can guess what non-existing product you wish would be made better than you can.  (Although I am reminded of Ford’s claim that “If I had asked the customer what they wanted, they would have said ‘a faster horse'”).  What I think will happen instead is that people will be working in the garden or playing with their kids or plinking at the range or doing some other task and think “I wish I had something to make this specific task easier, more fun, quicker, more efficient, etc… and, whereas before the effort to turn that wish into a real product would be prohibitive unless it was done purely as an unprofitable personal quest or else sold to a large market of other people, that person can spend a much smaller amount of time working with an AI CAD system and tinkering with a couple of iterations out of their FabMachine and then have the item they want without having to be a jack-of-all-trades engineer/woodworker/machinist/welder.  Once you change that break-even point for when it makes economic sense to spend the effort to build or improve a product for a niche or one-off demand then we will have a lot more such custom or low-volume improvements being done.

What would be the cumulative effect of a continuous stream of these sort of customized improvements saving even just a few seconds a day for us all?  https://www.youtube.com/results?search_query=lean+is+simple

Leave a Reply