The value of most software for most companies has a limit, pushed down both by supply of software engineering labor and that manual labor is actually quite cheap and your software must inherently cost less than what it would to just pay a human to do it.
There are always outliers obviously - whether it's the adtech giants being in the right place at the right time, or niche industries (avionics software costs a lot more to develop, and yet even then the actual SWE salaries there don't reflect it), but by and large the value of software is nowhere near what a decade of market distortion led us to believe.
Remember most of us (including me) are plumbers, not rocket scientists. It's just that a decade of market distortion allowed us to play rocket scientist (with only a tiny minority of those rockets actually ever needing to fly). Now we're back to being what we really are.
> The value of most software for most companies has a limit, pushed down both by supply of software engineering labor and that manual labor is actually quite cheap and your software must inherently cost less than what it would to just pay a human to do it.
Software, in the limit, is FREE. It gets written once for a finite cost and then can be copied ad infinitum. Even stratospheric labor costs are acceptable because of this.
> (avionics software costs a lot more to develop, and yet even then the actual SWE salaries there don't reflect it)
But the cost to develop a piece of software (or anything, really) has nothing whatsoever to do with its value!
Avionics software is expensive not because it's inherently more complicated (high-quality OSS hobbyist implementations like Ardupilot exist for free), but because it has to undergo system integration tests because of FAA regulations.
And, despite lower pay for avionics engineers, the value to the end users of quality avionics is literally the difference between life and death.
> but by and large the value of software is nowhere near what a decade of market distortion led us to believe.
Where do you get this idea? Tech, as a whole, is a wildly profitable sector of the economy. It is a primary driver of US GDP growth and productivity [1].
> Remember most of us (including me) are plumbers, not rocket scientists. It's just that a decade of market distortion allowed us to play rocket scientist (with only a tiny minority of those rockets actually ever needing to fly). Now we're back to being what we really are.
I get that ZIRP is ideologically unpopular, I really do. You can only see so many "Uber but for dogs" startups get multimillion dollar funding rounds before you get cynical.
That being said, interest rates have been falling for centuries [2]. Low interest rates are a fundamental reflection of a society that's gotten better at allocating investments efficiently.
It turns out that software, because of the whole "zero marginal cost" thing, is a spectacularly efficient investment.
There are always outliers obviously - whether it's the adtech giants being in the right place at the right time, or niche industries (avionics software costs a lot more to develop, and yet even then the actual SWE salaries there don't reflect it), but by and large the value of software is nowhere near what a decade of market distortion led us to believe.
Remember most of us (including me) are plumbers, not rocket scientists. It's just that a decade of market distortion allowed us to play rocket scientist (with only a tiny minority of those rockets actually ever needing to fly). Now we're back to being what we really are.