If you have been on Tech Twitter (I’ll probably never call it X) for the past 6 months, you’ve probably consumed enough dystopian level rhetoric to fuel a minor manic episode. Developers left and right are screaming to the sky about artificial intelligence replacing their livelihoods completely. Why? Let’s talk about it.
On one hand, I find it completely fair that people are skeptical about the future of their careers. It’s natural. Since that is how they provide for themselves and their loved ones, they’d be fools not to stay up to date with industry trends. And there are many, quite honestly, amazing tools out there that are free for use and can provide an unimaginable amount of value in the right hands.
On the other hand, why are so many software engineers making it seem like they are the ones in the most danger, and that this has never happened before? It’s a classic pattern throughout history: every time a “job killing” tool arrives, it actually ends up raising the floor of what a single person can achieve.
We move from being the engineer of the work to being the orchestrator of the work in many capacities. If the “AI is the end” crowd were around for the last two centuries, they’d have run out of breath a long time ago. Let’s explore some disruptions that were supposed to leave us all in the breadlines years ago.
The Carpenter vs the Saw
Back in the 1800s, building a house meant every piece of wood was to be cut by hand. It was backbreaking, slow, and prone to human error. You fast forward to circular and miter saws, and now carpenters stopped spending 40% of their days on tedious grunt work and instead spent more time building more complex, ambitious projects. Did carpenters disappear? No.
Architects vs CAD
Ask an older architect about “the good old days”, and you’ll hear about their world of tactile precision and physical gravity that today’s digital native designers might find tedious.
In the 1950s, drawing was a full body experience. The smell and feel of cedar pencils, ammonia from blueprint machines, the satisfying slide of a heavy parallel bar or the precision of a solid brass drafting compass. (That experience compared to the roars of an overheating Windows machine, does seem quite lackluster though)
They might miss the “think twice, ink once” mentality–because making changes was difficult, architects spent more time mentally resolving a building before a single line was drawn–today it’s easy to iterate 100 versions of a floor plan in an hour.
You may hear horror stories of an ink blot ruining weeks worth of work. Then CAD, Revit or Rhino arrives and offers unmatched speed and complexity, and architects lose the human printer aspect of their day.
Today one architect could design a skyscraper that would have required a room of 50 people in 1950. We didn’t get fewer buildings; we got better, more technical, and more artistic ones.
Programmers vs Compilers
In the early days of computing and programming, engineers didn’t code. Instead, they toggled switches or used punch cards to speak directly to the hardware in binary or Assembly. Then, Grace Hopper–the legend she is–struck gold and developed the first operational compiler, the A-0 System, which translates mathematical notation into machine code. Following that up with FLOW-MATIC in 1958, which is the first programming language to use English like syntax, making programming far more intuitive and efficient.
Her work fundamentally changed how programming was approached, and machines could now understand human language. Odds are, there were hardcore Assembly programmers still thinking that punch cards were the only viable option and that “real programming was dead”. While I do acknowledge the idea that there would’ve been some truth in saying that, the slow adoption and hesitancy as an industry would’ve unncessarily halted innovation.
Fast forward to the development of higher level languages like C, then Java, then Python and boy those Assembly and COBOL diehards were probably clutching whatever pearls they had left. But instead of the industry dying, the abstraction allowed us to build the internet, mobile apps, and train the AI itself that scares so many today.
By making coding “easier”, we made the world run on software. I’ve spoken with developers who were well into their careers when more and more no code tools were released. People were worried and saying the jobs were gone then too. Just think how ridiculous it would sound today if someone told you Wix was going to remove the need of having great web development teams, or a large amount of web developers at all.
The storm feels scary because these tools are new, but history doesn’t suggest we are being replaced, we are being upgraded. The core skills of a great engineer remain invaluable:
- logical and critical thinking
- problem solving
- passion for building
- attention to detail
AI will augment these skills, not replace them. If you are a builder, someone who uses whatever tools around them to solve a problem for themselves/their company/the world, this is your time to shine. Innovation always comes with a price, and we’ve seen it many times in the past.
The Safe Job Delusion
Let’s be honest. In this world where AI is capable of replacing a massive amount of jobs, software devs are probably amongst the last to actually see true extinction. It seems people are forgetting that enterprise is typically slow when it comes to implementing cutting edge technology, and I find it quite safe to say that the general public will be even slower than that in its’ adaptation.
In this imaginary land where all jobs are replaced by AGI and robotics, what jobs are actually safe? I challenge you to take the idea a bit further. If we continue this line of thinking about AI replacing dev jobs completely, it only makes sense that you also believe that eventually most careers are gone as well. If the people with the actual technical capacity to maintain, innovate, and build these systems are not safe, then everyone else is doomed.
Let’s say the infrastructure for robots was ready today, and we replaced all the essential careers the world has depended on, what’s to say that general society would accept it?
We could have fully replaced cashiers with self checkout machines years ago. Why haven’t we? You have to properly handle theft, system errors, or even the elderly not knowing how to operate the system. As we’ve known for years now, cutting edge software still needs to have practicality.
We already have the power today for AI assisted brain surgery, and if not yet then soon will have full AI powered robotic surgeries as well. Are you putting your child’s life in the hands of experimental technology? (It will take many years for the general notion of these systems to not be experimental)
Let’s say the boiler in someone’s home is malfunctioning. Do you think the average person is going to call the company who can have a AI robot technician come out in a full-self-driving Tesla (the majority of people are still cynical and skeptical about this too) at 4am, or would they wait the extra ~12 hours for the tech with 10 years of experience to come out?
I’d argue the average person would elect to wait for the human to come out to their home, because like I mentioned before, the convenience of the robot is great and all but one error is all it takes to make a possible simple fix, a huge fix. The amazement and convenience of a machine is great until a hallucination becomes a house fire, and you lose what’s most important to you.
Do people seriously think that as large of an industry as artificial intelligence is & will grow to be, that governments will not interfere with it? I could easily see a world where if companies choose to do massive layoffs and replace their engineering teams with AI, that there could be some serious regulatory friction by way of taxes or fines.
In lots of discussions on the impact of AI and its’ future, I don’t think the economy is touched on enough. Governments are not going to let AI be the downfall of life as we know it. More powerful models and tooling will be gatekept from the average consumer and companies, the average cost-of-use will eventually dramatically increase and the market itself will see a rise in demand of human content and interaction.
It is after all, in a countries’ best interest to keep unemployment low, and I believe there will always be a demand for human based creativity and innovation. I doubt the world will just openly and happily accept the fallout of society and rush to replace all careers ASAP. iRobot is not happening any time soon.
A Reality Check
2026 makes it the fourth year I have heard the idea that developer jobs are done. Please. Stop. Overreacting. Programming stretches far wider than web development, which is primarily where these current models excel anyway. While AI is a powerful tool for developers, it isn’t without it’s growing pains, and as it moves from novelty to infrastructure, several societal concerns arise.
There are issues with copyrighting and the data these models were trained on. As of today, the fuel for these models is human generated data–books, art, and the very code you write–and the fair use vs theft argument is a massive legal and ethical gray area. Is using copyrighted work for training “transformative” or just unlicensed copying? Depends on who you ask. Some could easily say yes, and others could say the models aren’t copying materials and instead are learning, referencing and understanding them like you or I would.
AI is incredibly thirsty. Training a large model isn’t just a software task, it’s a heavy industrial process too. By the closing of 2026, data centers are expected to consume over 1,050 terawatt-hours of electricity. That’s roughly the equivalent to the energy needs of Japan.
Keeping thousands and thousands of GPUs from overheating requires massive amounts of water. Estimates suggest AI could be responsible for withdrawing up to 6.6 billion cubic meters of water by 2027.
Dan Vega shared some thoughts about AI on Twitter recently:
I’ll be honest, as someone who creates videos and courses as part of my job and something I genuinely enjoy doing, I have real anxiety about the future of that too. Are developers still watching long form content when AI can answer questions on demand? Analytics don’t lie and the landscape is shifting.
Let’s be clear: the growth in capability and technical knowledge of artificial intelligence does not remove the human desire for connection. We are nowhere close to AI understanding humans better than humans. It means something to hear someones story, to see their journey, and to see their humanity. Humans are social creatures by nature, and that will not change within any of our lifetimes. While AI is a powerhouse at pattern recognition and understanding systems, it still lacks the accountability and contextual nuance that remains with humans.
So, how about we take a step back, take a deep breath, and acknowledge the reality of the current day and future. Instead, we should spend the brain power figuring out how to wield this incredible new power. The developers who learn to effectivelty partner with AI, who embrace it as a force multiplier, will be the ones who define the next era of innovation.