Remix.run Logo
▲ asdfman123 a day ago

I'm actually quite good at building robust systems, cleaning up messes, documenting things well.

As a result I get:

1) The enmity of my lead dev who prefers to do things more fast and loose.

2) Not promoted, due to the first reason and due to the fact that I don't have "impactful" metrics.

I'm not saying I'm perfect, there are definitely things I could have done better. But the fact still remains that the industry grossly undervalues maintenance and optimization that doesn't lead to easily measurable numbers that impact the bottom line.

That in turn is because it's hard to tell if SWEs are actually paying down critical debt or just wasting time. But that's the core problem, not that we're not good at it.

The good news is I can get AI agents to do this kind of work so I can now chase after metrics.

▲sushisource 20 hours ago | parent | next [-]

I think this thinking misses the forest for the trees a lot of the time. Assuming you're as good as you say you are at this, you need to ask how much building robust maintainable systems _actually helps the business_. The number of products where being super duper robust actually matters is probably less than you think. It can be painful as an engineer to accept, but it's often true.

I actually happen to work on a product where reliability _is_ our #1 most critical thing, and even then in certain features and certain situations it is perfectly reasonable to prefer time-to-market over quality as long as we are clear about that to our users.

So, I'd encourage you to think from a product perspective.

▲ctkhn 19 hours ago | parent [-]

We have a similar issue but when you look at it long term it still doesn't make sense from a product perspective. We have incredibly brittle and unreproducible env setups from local (just got working this year) to dev, qa, pre-prod etc. and some tickets to automate away a lot of that work would be paid off within a quarter. Instead everyone up to and including staff engineers does nothing about it. Even our documentation of infra setups in lower environments (that prod support uses to set up higher envs) is very sloppy. But no one is evaluated on it at all. I still do it thoroughly myself just so I don't have to re-learn things when I see them again a year later but otherwise there would be no point.

▲jgord a day ago | parent | prev | next [-]

I think this is a real alignment issue - nobody has excess time/money to make the software robust/secure.

This is possibly due to the high level of economic inequality, meaning less resources in the middle of the economy.

In a decade I think we will have few people who can competently write or even read computer code without AI assistance.. and I think we might say the same about math and science and medicine.

AI assisted learning could lead to a renaissance of human capability .. but thats not where we are headed, we are outsourcing all the hard thinking skills. In future, only a few crazy retro nerds will learn assembler or multivariable calculus, just as only a few crazy nerds now make mechanical watches as art objects.

▲jgord a day ago | parent | next [-]

In my own social circles I hear of capable young people who want to forego university and just take up a trade, because it avoids debt.

Excessive inequality means we wont be training young scientists, and young couples will defer having children until its biologically too late.

High inequality + LLM AIs + climate change is not a recipe for human flourishing.

▲mschuster91 a day ago | parent [-]

University means knowledge work, and knowledge work is already getting fucked over by AI. The clankers can now design rudimentary PCBs. The only university degrees that will not be compromised by AI are the ones that deal with humans in the meatspace - think medicine, psychology or education, although I think we might be getting a revisit of MOOCs augmented by AI to drastically reduce the need for teachers.

Being good at one of the trades however? That's a pretty safe bet. It will be quite the time until physical robot "bodies" are capable of the dexterity, size and endurance of a human. And by the time robots get that capability, we're either living in a Star Trek utopia or in Matrix style neo-feudalism.

▲asdfman123 a day ago | parent | prev | next [-]

> This is possibly due to the high level of economic inequality,

Friend, I work for Google. The high level of inequality works completely in our favor, yet we are incentived to churn out slop that hits metrics.

▲ponector 21 hours ago | parent | prev | next [-]

>> nobody has excess time/money to make the software robust/secure

That is not true. Microsoft is extremely profitable. Apple as well. Amazon. Banks are literally printing money. All of them have resources to produce robust secure software.

They just decided not to go that way.

▲pixl97 21 hours ago | parent | prev [-]

>nobody has excess time/money to make the software robust/secure.

Regulations are written in blood. No industry regulates itself at first. It's always no brakes, no safeties until a horrific event occurs, then the crackdown comes.

▲palata a day ago | parent | prev | next [-]

I feel similarly: I am slow but solid. The industry values those who move fast, break things, and make a mess.

Which makes sense: those in a position to promote others generally don't understand the technical part. What they see is how fast they get a new toy to play with. And if it breaks too soon, they will still be able to blame the developers.

▲cess11 a day ago | parent | prev | next [-]

While you do that, look for opportunities in niches where "fast and loose" is unacceptable, because while they might not market themselves as loudly as most software development organisations they do actually exist.

Typically these places are associated with compliance, audits and regulation with teeth. It might be adjacent to lawyers, in public institutions or weird consultancies. At least that's where I've found such work. Insurance and banking might be decent bets, and they sometimes also come with oddball tech stacks, perhaps mainframes, perhaps Smalltalk.

▲RetroTechie an hour ago | parent [-]

And a whole field like that is about to explode: robotics. Add self-driving vehicles too.

▲mock-possum 11 hours ago | parent | prev [-]

sounds like you’re just at the wrong company / on the wrong team - and unfortunately the job market is scary enough that it’s not so easy to just jump ship these days. Very frustrating.