Hacker News new | ask | show | jobs
by doubloon 1005 days ago
I would like to apologize on behalf of other ycombo users for your top comment being downvoted, and also for people calling workers "oxen".

I feel I seldom hear in tech circles (on the internet at least) the empirical evidence around the Toyota Production System. Maybe I didn't either until I read "The machine that changed the world" by Womack, Jones, and Roos.

It is a historical fact that Toyota produced higher quality with lower cost all through the 1980s vs their American competitors. In the same era at General Motors, CEO Roger Smith said he would build a lights-out factory 100% run by robots. It never worked and they had to abandoned the concept.

When I listen to those Toyota people talk about their system, the concept of mutual respect is the basic foundation. The engineers at Toyota have to spend time working on the line. Exactly what you are describing, partnership vs Taylorization.

Sometimes I feel like this explains Tesla quality issues. When I look at how some people at Tesla think about line workers, they literally call them "Button Pushers" or even use racial slurs. Elon when he described the "machine that makes the machine" was echoing what Roger Smith said in the 80s about a "lights out factory". And here we are ten years later and Tesla does not have a lights out factory and their quality is still very low.

Where I work in the US, it is a service business but some of the concepts still apply. Some of our tools and processes were built from the top down with little user input. It is a struggle to do Poke Yoke or Kaizen, and part of that I think may be the grapevine knowledge to "keep your head down and your mouth shut".

1 comments

> I would like to apologize on behalf of other ycombo users for your top comment being downvoted, and also for people calling workers "oxen".

I appreciate it but no need - the people doing that (typically) are as much victims of this propaganda. They aren't malicious, they have been learned the same culture we're talking about.

Funny enough, the first place I ever interned had attempted to do a lights out factory in the 1980s. Short term, it was a disaster - in part because they literally had not installed sufficient lighting to trouble shoot problems, push automated transfer vehicles back onto control strips, and pry parts apart. The machining centers of the era could do the quality but required experts to operate, monitor, and adjust. When things went wrong they weren't small anymore, they were destructive and shut down production for 12+ hours. The shift had been made in part to leverage concessions from the workforce, and (IMHO) as a result it disregarded all the tacit skills those employees used to keep the place running. Long term, They ended up with higher quality, faster throughput, but more employees. But, it cost them a lot and they needed more employees to move stuff around the factory because they had built this huge space to do lights out and the irredeemable piece of lights out was material movement and handling. What I learned was that robots and machines can't tell you what went wrong - no matter how much you log.

I've got a ton of other examples. Every factory I've worked in I've learned how much of a myth 'low skilled labor' is. On the floor, in every factory, new hires were useless - which is a pretty definitive marker of 'skilled labor'. Same thing in the military - senior NCOs are often critical mentors to new LTs. As an engineer I've typically learned as much from the technicians and operators as from the other engineers. At each, I've also worked with really smart engineers who were terrible at their job because they had been socialized that their intelligence was all they needed and there was no need to learn from the 'floor guys'. Roger Smith (and Elon) thought engineers could make the transition to lights out with their knowledge alone because they had bought into the same narrative about blue collar work be low skill and low intelligence. The result was concerns about air friction on robots but not tolerance stack up of stampings or repairable castings.

The problem is, until you really engage with it, the narrative descriptions of certain work we get from the media and economists become a lens for making sense of reality. I'm not immune from that, I just had it beaten out of me. Oxen aren't stupid either, they're still really useful in situations where tractors struggle.

dude i could listen to you talk for hours.. do you have a book?

i really feel like alot of this same stuff could apply to jobs inside service / data processing industry. It reminds me of "Gladys the L6", the person six levels down in a bureaucracy who is the only one who actually knows how things are actually working with the spaghetti of multiple legacy systems and arcane rules. Michael Lewis wrote about this case at Athena Health