Ask HN: In hindsight, what would you do differently for a startup tech stack?

[lumberjackstian]

I'm mostly curious about the context of a startup (because I'm doing that now). And I'm wondering what kind of things people wish they did in hindsight given their situation.

I for e.g. already wish I had spent a "bit" less time on ideating and a bit more time on getting in to flutter (as now I've gone in to an iOS mode and I know that's not scalable at the start since we won't have man power to maintain an android and ios codebase)

[souprock]

It's all behind us now, but...

Don't run all network services on a Gentoo box.

Don't write yet another C++ strings library.

Don't write yet another C library. (that is, a libc implementation)

Don't put developer workstations on a network that connects to the internet.

[ScottFree]

> Don't put developer workstations on a network that connects to the internet.

Is there a story behind that one? That one is very common and not traditionally considered harmful.

[souprock]

There is the general principle. Connecting things to the internet reduces security. You'll get hacked. You should protect your financial information, employee private data, customer private data, and work product.

It's especially important when the work product includes hazardous information. We had enough to cause large-scale disaster. Maybe I'd best leave it at that.

Even the smaller stuff is a problem. Imagine full data dumps from places like the Federal Reserve, Google (particularly Gmail and Project Zero), or Verisign.

Things can build over the years. At first it doesn't seem like a big deal to be on the internet. It's convenient. You don't have to buy two computers for everybody. Eventually, if you notice it or not, you may have a large amount of dangerous data.