The Increasing Demand for Software Engineers
The factors inhibiting software engineering growth at companies is related to the economy, not AI. Costs are increasing faster than profits, and that makes it hard to hire very expensive software engineers. Even still I am still seeing massive offers being made by big companies. Even OpenAI offers are bananas. If they were truly automating away software engineering jobs in 6 months, why would they keep spending all this money on engineers and not more GPUs?
History does not necessarily predict the future, but every exponential improvement in software engineering has only made the field more expansive and lucrative.
Giving software engineers the ability to write higher level code because of compilers instead of assembly was a game changer, and it not only enabled more people to become engineers, but it made them all way more efficient.
We got Rollar Coaster Tycoon written in pure assembly, but now folks who are not quite as legendary can make games too. And that’s not a bad thing! More games meant more competition, which meant more engineers.
It is my prediction that the need for Software Engineers will increase as the economy recovers. Now, I am not an economist so I have no idea when that will happen, I am only relying on the fact that it will eventually. Interest rates continue to slowly come back down, and while were are in the middle of the dumbest trade war of my lifetime, I am bullish on a bright future.
It is also worth mentioning that the day we can replace software engineers with AI, means that the AI can improve its own code. That is the path to the AI Singularity, which means none of us will need to work. The machines will be capable of anything, and their rate of growth will be limited only by the energy we can consume. I’m not AI (or so I will have you believe), but my goal would immediately be Dyson Sphere (a megastructure that absorbs all the suns energy).
That doesn’t mean we won’t start to replace some engineers with less efficient LLMs, but it’s a big leap to say we’ll replace many or most engineers. If LLMs are cost effective, even if they aren’t as good we will see shifts, but folks wanting to move fast and build great things will continue to need skilled engineers until an AI Singularity.
So today is a great day to learn software engineering, don’t let any hypster tell you otherwise without real proof.
I love AI tools, and they can enable incredible things, but as someone exposed to very large code bases, I have not seen any signal that indicates software engineers are going anywhere. The role simply evolves as it so far always has.
The market is hard because the economy changed. The AI hype and people screaming about the "end of the engineering world" certainly hasn't helped, but we have a doomsday culture, it is ingrained in us to look at how bad things are.
The amount of world ending events I've lived through is stupid. Smart people get tricked all the time by people yelling that the sky is falling. During Y2K, planes were supposed to literally fall out of the sky.
Remember the Mayan Calendar 2012 world ending? I knew some very smart people saying some very dumb things simply because the Mayans didn't make an infinite calendar.
Life Updates
My focus remains ever present on the course! For anyone new, I am working on a course to help new and aspiring engineering managers who are uncertain, overwhelmed, and stressed become confident leaders so that they can successfully guide their teams and drive results.
To that end, I have been filming and uploading some content to YouTube to practice making videos and become more comfortable putting myself out there. I shot all of this on my phone and my Sony a6000, and just today I unboxed my new Sony FX30! I am absolutely in love with this camera, and I am learning all about it so I can start filming.
My next step is to film Module 5 on building trust. I’ve made all the slides and content, now I need to film and edit and figure out the process. I don’t have a strong reason for filming this module first, other than I didn’t want to do the intro first so I just picked one to trial out.
Once I’ve done that and I’m happy with it, I will go film everything else. I have a lot of course content and extras to make, but my primary focus is on the core course content. I may start selling the course before all the extras are done, and then add the ones I have planned over time. This will mean I can help people sooner. I will release all that extra content for free for everyone who purchased the course. My goal is to keep making it more valuable over time!
A huge thank you to everyone who has offered to help with the course! I super appreciate you, and I may take some of you up on that offer, but I need to make it first :)
> We got Rollar Coaster Tycoon written in pure assembly, but now folks who are not quite as legendary can make games too. And that’s not a bad thing!
While I agree that it's a good thing that more people can make games than ever before ― I've played many very good indie games from people that are much better designers/artists than they are programmers ―, there is something to be said for having very deep technical knowledge within a field. I'd argue that one reason that many companies need more software engineers than ever before is because the skill level of the average software engineer has been on the decline. Having to write in assembly to make a game, for example, filters out anyone who doesn't fully love programming as a craft, whereas the 6-figure salaries at large tech companies attract a very different crowd.
This isn't to be entirely pessimistic though; there are more resources than ever before to learn advanced things like assembly language, GPGPU programming, or whatever else you can think of. Those skills will always be in-demand, and LLMs have a very long way to go before replacing those jobs. In my experience, LLMs are only good at doing things they've seen before, whereas even the most standard of tech jobs require at least some basic ability to reason about relatively complex systems and come up with semi-novel, context-dependent solutions.