Doubling Down

Hardly a week goes by without some tech “visionary” proclaiming the end of software engineering. Even industry veterans, in the face of the growing number of autonomous AI agents like Devin, are calling it the last glorious years of a golden run. There simply may not be a need for as many, or even any human software engineers in the not so distant future. Sounds ominous.

Well, I’m not so convinced.

The outrageous claims like “there will be no human programmers in 5 years”, or that you should no longer learn to code, or that you’ve built an autonomous software engineer and then proceed to list vacancies for actual software engineers sounds exactly like what you expect to hear from people who’s vested interest is convincing the masses that AI is taking over the world.

Those GPUs aren’t going to sell themselves. And how else are investors going to give a bunch of competitive programmers, a skill which involves solving small, well-defined problems (the diametric opposite of a SWEs remit) over $21 million for what can best be described as a GPT-4 wrapper? And wouldn’t company leaders love to put profit before quality, removing those annoying salary expenses from the income statement? We all know how that turns out.

I’m not saying software engineering will not change. In fact it already has. But my bet is firmly in the camp that artificial “intelligence”, if that’s what you want to call the current breed of token-predicting processes, will augment programmers, not replace them. I agree with Patrick McKenzie- I’ll question if the world itself still exists as we know it if 50 years from now, software engineering no longer exists.

If you find yourself believing in the worrying prophecies of our industry’s “thought leaders”, I think it says more about the complexity of work you’re doing rather than about any actual threat to the industry. As Patrick Mckenzie adds, saying AI will take over is equivalent to saying that “no humans have to understand complex systems at a granular level in the future”- like him, I’ll short it to zero.

An important corollary is then to avoid the metaphorical Carthaginian plains salted by AI, we need to head to higher ground.

Which is why I’m doubling down. Moving up the food chain. And diving deeper into CS fundamentals.

My bet is that these skills will be even more valuable in the future. Topics like computer architecture, networking, operating systems, databases and compilers- they’re Lindy skills. Fundamental concepts that were developed decades ago and that will still be here decades from now.

To be fair, I was going to learn this stuff anyway but the impending AI doom has given me even more impetus.

First on the list is computer architecture- I’ll be working through Computer Systems: A Programmer’s Perspective.

See you on the slopes!

Ps: if you’d like to follow along, I’ll be posting regular updates of my curriculum and journey. You can subscribe below or follow me on Twitter/X (link below).