"4. Generalist technical knowledge." is good, but the ending of "* You need someone who can optimally work on all parts of the stack." is delusional. You want most of your people to be able to do credible work on all parts of your stack, but doing really good work requires some* degree of experience and the specialization that comes with it.
In general while the author's biases are by and large good I can seem someone following all these points ending up with a lot of technical debt, something that companies don't always survive.
The people who are simultaneously good at writing device drivers in ARM assembly, JSP pages using Spring, and natural language processing are indeed rare.
However, most startups don't require that breadth; rather, they need a good grasp of web technologies, a good grasp of databases, a good grasp of networking and a good grasp of the problem domain. THAT'S NOT AS RARE. You just have to know where to look, and how to be attractive enough -- usually these people are happily employed and make good money. To attract them, you need to offer better terms (money; interest; title; whatever)
Historically that has not been "most startups", although maybe in today's world that's severely financing constrained but has a rich web/Internet and FOSS ecosystem that's true.
Which would be a shame, since that means "We aren't inventing the future anymore". I.e. no new microprocessor macroarchitectures, no new weird fun thinks like FPGAs, just endless variations on today's lowest level and most expensive to establish technical foundations, improved by sustaining innovation but very little disruptive.
That said, I'm a bit skeptical that there are all that many people who are really good at the forward facing stuff and seriously scalable backends ... but I could be wrong, and I've seen it reported that many if not most companies seem to be focusing on the first and then if they're hugely successful they spend a year or more paying down their backend technical debt. Some companies fail at that.
Which is the way to do it in my experience. The most frustrating startup failure I experienced was dead when I joined it, since a combination of ultra-stealth mode plus building a massively scalable custom database system ended up spending their first angel's available money (he'd promised more, but then ran into trouble with another company).
So they got some new angels who turned out to be devils, had done no Customer Development (a finer example of "there are no facts inside the building, only opinions" I cannot imagine), but it hardly mattered since the devils preferred to have all of nothing instead of most of something.
5 comments
[ 3.4 ms ] story [ 24.6 ms ] thread- Do-What-It-Takes-edness
- Beer hiring test
In general while the author's biases are by and large good I can seem someone following all these points ending up with a lot of technical debt, something that companies don't always survive.
The people who are simultaneously good at writing device drivers in ARM assembly, JSP pages using Spring, and natural language processing are indeed rare.
However, most startups don't require that breadth; rather, they need a good grasp of web technologies, a good grasp of databases, a good grasp of networking and a good grasp of the problem domain. THAT'S NOT AS RARE. You just have to know where to look, and how to be attractive enough -- usually these people are happily employed and make good money. To attract them, you need to offer better terms (money; interest; title; whatever)
Which would be a shame, since that means "We aren't inventing the future anymore". I.e. no new microprocessor macroarchitectures, no new weird fun thinks like FPGAs, just endless variations on today's lowest level and most expensive to establish technical foundations, improved by sustaining innovation but very little disruptive.
That said, I'm a bit skeptical that there are all that many people who are really good at the forward facing stuff and seriously scalable backends ... but I could be wrong, and I've seen it reported that many if not most companies seem to be focusing on the first and then if they're hugely successful they spend a year or more paying down their backend technical debt. Some companies fail at that.
Which is the way to do it in my experience. The most frustrating startup failure I experienced was dead when I joined it, since a combination of ultra-stealth mode plus building a massively scalable custom database system ended up spending their first angel's available money (he'd promised more, but then ran into trouble with another company).
So they got some new angels who turned out to be devils, had done no Customer Development (a finer example of "there are no facts inside the building, only opinions" I cannot imagine), but it hardly mattered since the devils preferred to have all of nothing instead of most of something.