3 ms·
I read the article so I'm gonna respond here ;) As a counter-anecdote I'll offer my experience working for a company that built large precision machinery. Man
by YZF 5y ago
I read the article so I'm gonna respond here ;)
As a counter-anecdote I'll offer my experience working for a company that built large precision machinery. Many of the parts of said machinery were built by subcontractors. The quality was extremely high. We did do certain things in house if they require special expertise and we did have the capability to do rapid prototyping but it wouldn't make any business sense for us to purchase all the equipment to make all those parts that we subcontracted.
I think this is common in many industries. (Automotive?)
Even with software this is far less than clear cut. There's plenty of times where you should not build it yourself and the quality stuff isn't always terrible. Seems like there's some cherry-picking going on in the examples re: bad software. Sure, there's lots of bad software, but there's also awesome software that works really well. Building it yourself isn't a guarantee that it'll work well. 90% of the time the people that think they can do better if they build it themselves can't, they just don't know it yet. And sure Kyle is great but there are plenty of solid databases built before he was around (before he was born?) by companies like IBM and Microsoft.
I've seen companies build in-house tooling that's much worse than what's available off the shelf and have an endless drain of resources due to that.
Build vs. buy isn't an easy call. I'll agree with that.
I felt like there was a self contradiction in the essay. You can't get great software but you can get great engineers that will build the great software in house? I don't think your chances of finding the team that can build that are any greater. You're just as likely to hire a bunch of people you think can do this and then find out they can't. 5 years later. Which sort of jives with the commentary of experts doing shoddy work...
EDIT: I'd also say that the examples of the stuff that was built in-house that was superior might be a case of survivorship bias.
EDIT2: Thinking more about machines... motors, sensors, cables, connectors, bolts, power supplies, pumps, linear rails, screws, etc. etc. all of which you buy, never make, the quality is generally very good, and it's almost unimaginable that you can be completely vertically integrated.
- aidenn0 5y agoI've never worked in the automotive industry, but I've worked adjacent to it occasionally. The amount of effort put into ensuring that suppliers do what they need to do and do it well is huge. Companies always have two (or more) suppliers and if you keep only to the letter of the contract, you get cut. They can afford to put so much effort into managing suppliers because volumes are large and margins are low. There are also cultural issues at play; he makes the point that things that are inevitable in one culture are unthinkable in another and different industries have different cultures
- ako 5y agoThe difference between machinery and software in this example is that one is outsourcing manufacturing and the other design. When you outsource manufacturing you're asking the other party to create exact replicas according to very specific specs your own engineers created. With software the outsourcing usually includes outsourcing a large part of the design, which is inevitable, as there is no significant manufacturing you can outsource. Writing the software is creating the detailed specs. In software there is only product design, you don't really have to put a lot of effort into creating exact copies of the original design.
- erosenbe0 5y agoLinear rails -- THK. Pumps -- Iwaki. Optical glass & microscopes -- Kenko, Nikon, Canon. Sensors/Switches/Relays -- Omron. Imaging, Passives, Semiconductors -- Toshiba, Epson, Panasonic. Chips -- Renesas. All made in Japan. Few or no Silicon Valley equivalents. Makes you wonder.