It’s really great hearing some of the ideas from a generation of engineers younger than me. I’ve been out of the game so long I have forgotten some of the constraints.
I meant to comment earlier on the topic of most CS grads coding like they’re ignorant of how he machine works. I’ve seen this a lot but the fact it gets mentioned shows it is self correcting. As an embedded systems developer I’ve had to “break in” sw guys who didn’t know how to code small systems - small program memory, slow clocks, fixed point. After the initial naive complaints they’d realize why the constraints were there (cost, size, power consumption, speed) and they’d start coding. Some got really good at it. The university trains on particular platforms. My school had one semester of computer organization which students feared because you had to code close to the hardware. In retrospect it was a great requirement and may have been a student’s only exposure to low level programming over an entire career. In short, this stuff can be picked up. I don’t see it as some kind of tragedy that a developer doesn’t understand the inner workings of a hardware multiplier when he first sits down.
The prevalence of prepackaged code has always been an influence on developer value. I suspect like any profession it pays to be a little ahead of the crowd. If there are tons of linkable libraries with everything done for you and code generation sw to patch it all together the value of the customized goes up. It will go up as more and more people fall back on the code configurator sw and cannot go that last 1/4 mile in the race. I know a lot of friends in physics who learn just enough Python to get good salaries as data scientists but they’ll never cut out a lucrative niche unless they leverage their overdeveloped math skills or become more sophisticated machine customizers. Not sure I’m being totally clear here but that is my observation lately. You want to be the person who can take things to the next level. I never had enough friends to get hired for just doing the basics…there was always that little bit more that was needed that nobody in house had.
I think that goes for all professions beyond the digital ecosystem.
As for hardware limits, the chip shortage etc. yep, the shortage is a problem and I’m living right in the middle of it (my side hustle uses some 40nm tech). But people have been warning that Moore’s law would hit a wall (Tjunction=temp of sun) and it would stop twenty years ago. The ecosystem’s reaction? Multi core distributed processing and flattening clock rates. Yes there are limits but like a n insurmountable wall getting covered by kudzu the machine seems to find its way around them and continue its spread. At some point we will run out of sand and the ability to melt it to wafers but on the scale of the engineering career half life it’s irrelevant. We can sit around worrying about it or we can make money now. That has some pretty serious sustainability and selfishness problems but seriously I have to survive in a world I cannot easily change alone.
A lot of the semi shortage is climate driven. That’s a sad reality finally coming home to roost. I suspect the industry will find its way out like it always has and continue to grow.
It’s exciting to see the growth. I remember sitting with a boomer (a real one
) in 1999 and he said it’s an exciting time to be working in tech. It still is and more so than back then.