Silicon Valley Decides It’s Just Too Hard to Build a Car | Hacker News Reader

Silicon Valley Decides It’s Just Too Hard to Build a Car

bloomberg.com

29 pointsyabatopia17 comments

The Submarine giveaway here was the zero mentions of Tesla. They had quotes from a Ford family member but none from someone at one of the "tech" companies. I really struggle to feel like the news gives me a good understanding of the reality of the situation.

Yes, how is Tesla not Silicon Valley? Bloomberg used to be a great news source, but it has deteriorated sharply.

I suspect that the car industry's troubles with hiring software engineers as mentioned in the article has more to do with leadership in the auto companies. If they have leadership that understand software and software engineering, then they can set up an environment that appreciates the unique environment that software engineers like. They certainly have the financial resources to hire away competent software engineers from other firms.

I suspect that their troubles have more to do with the need for software developers experienced in stuff that's not JavaScript or some other "cool" technologies. They probably also require people to read/write and understand software spec documents, something most software devs in SV can't do.

I imagine it definitely has something to do with it. When I'm looking for a web dev I get 3+ CV's / day. When I'm looking for someone to help with my end of things (imaging device design and computer vision)... Well, I'm still looking and it's been two years.

People with these skills do exist. Perhaps it has something to do with your location. For instance, your name is EpicEng, which might mean that you work for Epic, the large medical record company in Wisconsin. Epic is not located in an interesting place or a place conducive for career growth with other companies, in contrast with say, SV, NYC, Boston.

Alternatively, if it is important enough to the firm, they should be offering significantly more money than they are.

Oh, sure, they're out there. I work in SD, so we do have access to talent (though we also struggle competing against SF.) My problem has been finding someone who is a software engineer first, e.g., not solely an academic type. We need them too as they're the people with depth in the computer vision arena, but I'm looking for someone who understands digital microscopy, image processing and analysis, but who will also write robust, production level software.

It's a bit of a niche and I was fortunate enough to be exposed to it early in my career.

Maybe you should take an apprentice.

I agree, it's just been hard convincing management to let me hire a junior person. That role always gets pushed in favor of our immediate needs.

Given how much difficultly so many software companies have in doing this, especially sustaining it, it seems unlikely companies from other domains are capable of succeeding, especially in this domain where sustained effort is required.

And car company cultures ... I recommend two things to get an idea about them, US ones at least (for Japanese, the revelations of Toyoda's horrible software practices after their code was subject to 3rd party examination is sufficient for me):

Read a bit of DeLorean's On a Clear Day You Can See General Motors (hardcover: https://www.amazon.com/Clear-Day-You-General-Motors/dp/09603... paperback: https://www.amazon.com/Clear-Day-You-General-Motors/dp/03805...), you'll be impressed by the insanity. Granted, things had to change after GM hit hard times, but when you start out with a base like that, well, nothing in recent decades should come as a great surprise.

And the bits in King of the Killing Zone (https://www.amazon.com/King-Killing-Zone-Kelly-Orr/dp/042512...), the best book I know of about the development of the superlative M1 Abrams tank, about the engineers dealing with their Chrysler car salesman boss.

His main concern, at least as described in the book, was that it looked good. Which I suspect was highlighted because it caused the most friction with the engineers and made a good story. Note that in and of itself, that's a good goal, men should feel proud about their weapons (and I can't imagine the culture of French rifle designers), but....

As it turned out, a lot of his input of this type had good results. One bit prompted the start of a design iteration where they shaved a foot off the height of the tank, which is critical for the European theater (by comparison, prior to the development of the "Silver Bullet", maximum main gun depression which translated into greater heights was key in deserts).

In the other example I remember, he insisted the side armor skirts be extended all the way to the back. Now, they were doing simulations of everything to help decide the very many trade-offs of parameters like size, weight, and protection/offensive capability, and this wasn't recommended by those.

But I guess it was a close enough call that they went ahead and did it, and not only did it improve the looks of the tank, it turned out to be important in limiting the kicking up of debris by the treads, which was critical since the M1 was the first (and only?) turbine powered tank, which requires a lot of air, and the air filtering system can deal with only so much junk.

Probably becuase they're in Detroit.

Most major car companies now have R&D offices in SV. In fact its called "automotive row": http://www.autonews.com/section/map502&iframe=true&width=960...

Maybe theyre Pay Sucks?

TLDR; The article is not only void of any new information but also falsly claims that Google has given up on manufacturing its own self-driving car without citing any sources[1]. Then it goes on to get quotes from "experts" asserting that building cars is too hard outside Detroit. That's basically about it.

Its all bull crap, if you ask me, and potentially paid article by some lobby. If a tiny company called Tesla can build a car out of nowhere, so can others. Although, the risk taking appetite of big tech companies has certainly went down recently and it does remains questionable if they can actually pull off putting car on market.

[1] There has not been any news (like Apple) on Google giving up on car project. In fact as per recent interview, the project is alive and well: http://www.theverge.com/2016/10/25/13307364/google-self-driv...

No car company builds cars from scratch (a few niche players might come close but they are economically insignificant despite the astronomical prices of the vehicles). I have a friend in the industry and his description of car makers is that they are kit assemblers (it's not said disparagingly).

Of course BMW, Mercedes, GM, etc. do a lot of design work but a lot of the parts are either standardised or within the customization range of their suppliers.

This is of course how Tesla managed to get into car making so quickly, by buying as much material and expertise and manufacturing capability from existing suppliers as possible.

If Apple and Google have decided to leave it to the existing big players that doesn't sound like any kind of defeat to me, just recognition that there are organizations that already have the expertise and supply chains up and running available to provide what the market wants.

And in fact even you can build your own car. Just search for kit cars. You can make it even self-driving (almost) with hardware like this: http://www.nvidia.com/object/drive-px.html

Tesla, in contrast, is good at building cars, but sucks at safety-critical self-driving systems.

Google has an effort with Fiat-Chrysler to build 100 self-driving minivans. The article seems to have missed that.

Reply on news.ycombinator.com