The vast majority of startups aren't doing anything remotely complex enough to be able to save money on electricity by paying for developers to write more efficient software. Or if they can - the instances are few and far between.

The vast majority of startups are also heavily cash-flow negative - so anything you do likely won't pay for itself.

It's almost always a bet on a rosy future.

Not true. The vast majority of startups hemorrhage money with inefficiencies. I routinely cut tens of thousands of dollars a month off a startup’s AWS bill. It takes me 30-45min. Why didn’t someone else do it? Too busy, didn’t know about it, didn’t know how.

I keep expecting to find a place that doesn’t need some sort of efficiency cost-saving. I’ve yet to find it.

The fact is there’s a crap-ton of beneficial work to be done anywhere you look. It’s not hard to justify a good salary in the software world.

For example: what if your company could double the speed of your CI/CD system and halve the price? If you move your runners to spot instances in an auto-scaling group you can do that. What’s the return on increasing the productivity of your eng team? Maybe eng salaries times percentage productivity improvement? That number is probably… large.

The vast majority of engineers earning these salaries work for "startups" like Google minting huge amounts of money on software built by these engineers. They are underpaid if anything. Do you really think management or shareholders (that get paid more for doing less) are the underpaid ones in this equation? Where else would the money go?

The amount of revenue per employees at some of these companies is in the 7 figures.

The "vast majority of startups" is not a useful unit of measure. Look at where all the people and the money actually are (FAANG).

Define FAANG.

Facebook, Apple, Amazon, Netflix, and Google employ <1M engineers in the US. There are 4.4M engineers in the US.

That does not change the point that most of the engineers earning insane salaries are working for FAANG

There are 4.4M engineers in the US, but almost all of them make far less than the one working for FAANG.

Fair point.

I didn't know we were talking exclusively about engineers in the top 5% of pay or so.

I thought this was about engineers in general.

I was paid $165k a year by a neobank startup to.....build their bank. It's now responsible for over $200mm a year in revenue. I was the sole engineer on the project.

Good engineers are worth their weight in gold.

Yeah when I read takes like the guy you're responding to, I have to wonder: where should the money go if not to the people that built the product? Management? Shareholders? You can say that but those people are doing less work for more money already, so I don't buy it.

Engineers like many people that build useful things, provide orders of magnitude more value than what they get for their labor.

Well...op left the company, so why should he still get a cut of their current revenues ? That does not seem very fair to the current engineers who a currently able to manage the $200 MM. As far as we know, the company really started working well when they were finally able to replace his crappy code with a better implementation.

The money probably goes to management, who stuck during the entire story - which OP didn't. It also goes to current employees who, by their number alone require more money. And of course to stakeholder who paid for OP $165k when the bank was making $0.

Actually no. They replaced me with 9 engineers to do the same amount of work and had to delay the project 6 months because they had me working 100 hours a week and renegged on a raise and a vacation.

This one.

I did unimaginable things for one company, only to understand the value years later.

You know who else asked themselves that same question? Karl Marx.

The Workers control the deployment to Production!

By that logic, any salary < 70% of AWS cost is justifiable.

In reality there is a job market, and there are office politics. These determine salaries more than "marginal revenue of labour". This means that as long as someone who is employable by office-political standards will do the job cheaper, the salary can be contested.

I get your point. But to add to it, there is a labor shortage of good software engineers and that contributes to the high salary. We are still digitizing our economy, we are still disrupting non tech savvy businesses with tech first startups, and we are barely beginning to apply ML all the areas it could make gains. Please see more_corn's comment in this thread about how they haven't found a job yet where they couldn't immediately save the company 10k a month. There is no shortage of work to be done and a real shortage of people who can do it. So the reality is yes, any salary < 70% of AWS cost is justifiable in this example.