Tesla buys robot maker. Hang on, isn't that your sci-fi bogeyman, Elon? Troubled Tesla Inc. has quietly acquired Perbix, which designs robot production lines. Perbix was already a Tesla contractor. "With the acquisition of Perbix, Tesla further advances its efforts to turn the factory itself into a product – to build the machine that makes the machine," the electric carmaker burbled on its website …
Tesla is buying a robotics manufacturing company.
Musk was warning about AI
Much as smoke-and-mirrors purveyors like to portray current robotics and expert systems, ANNs etc as "AI", in reality there is a world of difference.
The Register sensationalising the routine? Well, I never!
A relative who was spirited from the UK out to the West Coast as a senior production automation guru has told me that he's found Tesla really difficult to work for, and told his own employers that he doesn't want to work on the Tesla contract. And that from a complete workaholic, and one of the most well regarded people in his field.
I therefore suspect the reason Tesla are buying contractors is to stop them walking away at the earliest possible opportunity, and to ensure that if chasing other business, that's always lower priority than trying to please the shambolic and overly ambitious management of Tesla.
Message to Perbix technical and operations staff: Now is a VERY good time to look for another job.
I'm not surprised. The modus operandi of any Musk corporation seems to be: Burn em up, spit em out, chew on the next guy. SpaceX and Tesla both have a massive churn rate, with MANY reporting the expected workload and hours to be absolutely insane.
Right from the start Tesla have always shouted they'd built their own damn factory. With blackjack! And hookers! Their timelines and production line set-up were,from what I've been told by someone who works for a party that bid on a contract to do some of the work, "completely and utterly insane". Going completely against everything the rest of the car industry has been optimising for the last few decades. Ignoring many of the lessons learned that experienced contracters were giving them.
"Right from the start Tesla have always shouted they'd built their own damn factory.
Except.. They didn't? So once upon a Freemont suburb, there was a NUMMI. Or a JV between GM & Toyota to build cars and trade production technology. The Freemont factory produced up to 6,000 vehicles a week. Then GM went kinda bust, Tesla stepped in with Toyota. And now the NUMMI MK2 is producing far fewer vehicles than those old dinosaurs GM & Toyota ever managed. How could this be?
OK, so Tesla didn't buy a lot of the factory equiment in the NUMMI sale, but 7 years on, all the signs are that Tesla's made some big mistakes wrt production.. And is now facing increased competition from those dinosaurs who do know a think about manufacturing cars.
What they bought with the nummi plant was a building on a large plot of industrial real estate and a bunch of people with knowledge. The production line for one kind of car is pretty much useless for any other type of car not on the same platform. They had to pretty much strip the entire factory an build it up again from scratch. The people they bought with the factory could have helped them with that, but from what I've been told they were pretty much summarily ignored.
Yup.. I kinda get the feeling there's been too much micromanagement, wishful thinking and ignoring advice from veterans. And maybe some of this ""the economics of having a plant in California so far away from the supplier lines" in the Midwest "just doesn't make business sense" for Toyota to continue running the NUMMI plant.,
From the wiki article on NUMMI. I don't know if that kind of logistics tail holds true for Tesla's production, but they are lagging against their peers/equally efficient operators. In theory, the Teslas are meant to be simpler than ICE, ie fewer parts but that doesn't seem to be translating into production numbers. Seems simple on a whiteboard, not so simple on the lines. I'm rather curious about reliability/service costs, especially when I see things like this-
http://teslabears.club/t/new-thread-keefs-complaints-with-photographs/107
Quite a few Model S vehicles with similar looking suspension failures.
His fault; Musk himself blurs the distinction between robots and AI. From a speech of his this past summer:
“I keep sounding the alarm bell, but until people see robots going down the street killing people, they don’t know how to react, because it seems so ethereal.”
They plainly said as much earlier that their problems are related to certain sections of their assembly lines and that the failings of the contractor handling that were missed. In light of that I can see nothing surprising in wanting to take matters into their own hands, sorry. And factory "robot" intelligence has pretty much bugger-all to do with AI; if anything, most "smarts" involved concerns itself with optimizing the efficiency of the processes involved, not with performing them physically.
In light of that I can see nothing surprising in wanting to take matters into their own hands,
Well, true to form. They're learning the hard way all those "oh-so-easy" lessons of car making, and now they're thinking "what could be so hard about automating production?".
"Can we do it? Yes we can!" might work for Bob the Builder (even the crap, inferior version flogged to the Merkins), but back in the real world, every major car maker trusts and uses external specialists in industrial automation. This is a daft move.
This post has been deleted by its author
The author seems to be mistakenly conflating "AI" and "robot". While there are some applications where the two overlap, the industrial robots used on production lines often involve little or no AI -- and many big AI projects don't involve robotics at all.
One hopes this was an honest mistake and not one made intentionally to produce some synthetic drama for the story...
Cognitive dissonance much?
First we diss Elon for warning about AI (and yes, I'd like some of what he's smoking) and then buying robotics capability, then diss him for needing robots, assuming that the lack of them is the problem and robots are the solution:
" A more sceptical [sic] view is that Tesla is finding it so difficult to manufacture its long-awaited Model 3 sedan that it's hand-finishing them as they drop off the production line."
So you're going to diss the guy for realizing he needs to use the solution you approve of, really?
I'm sure he's (and other mgt) aren't perfect in all eyes - I've done the CEO/Owner/Vision-guy gig, and while things were successful, i'm pretty sure that those I made do what they promised weren't all happy with being held to it. So?
The thing about unreasoning hate, or even sour grapes, is it not only makes you wrong, after awhile you just look dumb. Yeah, no one in a huge startup gets it all done exactly as they said and on time, but this group of outfits does tend to get it done, unlike a ton of others the Reg hasn't bothered to heap abuse on.
Having worked in Automotive Manufacturing, it sounds to me like Perbix were contracted to design and implement sections of the production line and it's equipment, and as is typical, it didn't come up to snuff and was experiencing many unanticipated issues.
Then coupled with the fact that suppliers can get pretty finicky when their stuff doesn't work as expected (who wants to troubleshoot when you can sell and install new stuff), and Elon feeling the pressure to get production humming along while the supplier is dragging ass in regards to fixing it, it makes perfect sense to me that Elon would just buy them out to force them to focus on fixing the problems in a timely fashion.
It also has the side-benefit of growing the vertical integration, in that they now have the knowledge and expertise in-house to be able to design and build these types of production equipment.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
