An Engineer is smart. Engineers are dumb. So what about AI?

Directly or indirectly I've been asked to weigh in on the future of software engineering jobs in the age of AI. Directly by engineers who are often fearful and uncertain about their future and looking for some degree of hope. And indirectly by the torrent of ignorant drivel on social media.

My honest answer is that I don't know what the future holds. Software Engineering is an extremely broad field and it's hubris to assume that it wont, at least partly, be done by AI. With that said, I know that even before AI, there were engineers that I unironically thought could be automated away — and that's always been true. Some DevOps focused engineers will even throw around the quip that their job is to "make themselves redundant". I don't think anything has changed. New Technologies have shifted the way we think about and do engineering time and time again.

Though there is one perspective that I think I can add some colour to, and that's as to why you're probably seeing a lot of this rhetoric around replacing software engineers that has, so gleefully, captured the public eye.

And that is: at scale, engineers (from some angles) look kinda dumb.

In some leadership circles there seems to be a sentiment that if they don't hear anything (and bank-balance go up) it's probably a good sign. An aggregated view of engineers and engineering teams (especially under weak leadership) therefore, unfortunately and unavoidably comes with an aggregation of; incidents, missed deadlines, bad UX, poor customer feedback, resignations, missed requirements, bugs, compliance violations... you name it. All things that, when inspected at the level of the individual, are reasonable and intelligent. But at scale, suddenly seems reactionary and foolish. Especially to our feeble, logarithmically biased brains.

I've heard "why can't engineers just do what they're told?" more times than I remember. Ignoring the fact that if they did (without applying some degree of autonomous intelligence) we'd be marching toward a certain doom.

The joke's on them.

I've run out of fingers and toes to count the number of times I've witnessed single engineers literally save (much more often than destroy — though I've seen that too) entire companies or careers of their in-line executive. I am willing to bet that almost every engineer will have a story to tell about how they identified and fixed an egregious security problem in production before it was exploited, stayed up until the wee hours of the morning bisecting a codebase to identify a bug to bring critical services back online, convinced leadership to take a specific direction on the product that later salvaged delivery, or included a slew of "non-functionals" during development that later form the backbone of the service and company. You get the point.

But this, I think, starts to explain why some technology leaders seem a little too excited about the prospect of replacing engineering with AI. It feels a lot more real than it likely is. As if it's just over the horizon.

For the Software Engineers: I personally wouldn't fear for the people that are building the software today — or tomorrow. It's going to take time. Take solace in knowing that they are having to spend literal hundreds of millions of dollars at even having a crack at (if you squint) automating a junior software engineer (which, again, if you ask me is intruding upon the doorman fallacy. After all, the average engineer only spends 48 minutes a day actually writing code).

For software engineers (current or aspiring), I want to leave you with a few rapid fire thoughts:

1. The pursuit of competence in anything is itself valuable. It's the journey that composes into who you are. If it's interesting to you and you want to do it, just do it. It's only wasteful if you instead do nothing.
2. Software Engineers have always been remarkably proud of their ability to learn faster than others. Why does that suddenly stop now? Get learning.
3. AI is getting pretty good at asking the right questions. Aspects of all roles are under threat. If the doom-sayers are right, no one is safe.
4. Historically, the world has consumed just about everything we can throw at it. More execution power usually just means more competition which equals increasing consumerism and more jobs.
5. It's always been about what you build—not how you build it. There will always be room for people with great ideas who can bring them to life.
6. You will also benefit from the advent of AI — you'll likely live longer with less disease, be able to spend time doing things that are meaningful, and live in a more affluent society. That's good, right?
7. The same charlatans that pump any new tech beyond what it's actually capable of aren't going to stop just short of AI — there is still A LOT of noise out there. Take a well-earned break from social media to recalibrate.
8. Incompetence begets confidence. It's easy to look at another persons job at a distance and think it's simple. Software Engineering is no exception.

For leaders that are in a hurry to replace junior engineers:

The value of junior anything isn't so much that they are cheap, low-skilled labour. Juniors are your canaries. If your organisation is only navigable by highly-experienced people you're leaving real, tangible, improvements to your systems and culture on the table. An organisation juniors can navigate, where they can be productive, grow, and not get tied in knots is proof that you might be onto something.

The absence of junior roles is a red-flag that will let everyone know that in your organisation: to do anything, you'll have to know everything.