I had a conversation at the airport recently with a few older travellers. It started casually enough, as these things sometimes do, but it quickly moved into AI, work, and what all of this means for people trying to build a life and a career.
They asked me how I felt about AI.
That is a much harder question than it sounds.
The honest answer is that I feel excited, unsettled, grateful and worried, often all at the same time.
AI as an accessibility tool
AI has become part of my daily life. Not in a distant, theoretical way. I use it. I rely on it. I value it.
As someone who is dyslexic and ADHD/neurodivergent, it gives me a way to work with my brain rather than constantly fight against it.
I can write short, sharp bullet points. I can get the facts down. I can capture the shape of a thought before it disappears. Then I can use an AI co-pilot to help expand, structure and refine it.
That does not mean outsourcing my thinking. I still approve it. I still shape it. I still know when something sounds like me and when it does not.
But it removes a layer of friction that has been with me for most of my life.
Growing up with dyslexia was hard. I struggled. I was bullied. That had a real effect on my mental health and confidence. Like a lot of people who find school difficult for reasons that are not always understood, I carried a sense that I was behind, or broken, or not built for the systems around me.
Technology became the place where I could breathe.
IT gave me an identity. It gave me a language. It gave me problems that made sense, even when they were difficult. It gave me a path where persistence mattered, curiosity mattered, and where being able to see patterns differently could actually become a strength.
That is part of why AI feels so personal to me. It is not just a productivity tool. It is an accessibility tool. It is a thinking partner. It helps me express things that are already there, but sometimes hard to pull out in a clean line.
The everyday and the uneasy
At the same time, the conversation at the airport stayed with me because there is another side to this.
We talked about how normal AI has already become. People are using it every day now, sometimes for useful work, sometimes for research, sometimes to write, plan, code, summarise or learn.
And sometimes just to generate silly pictures.
There is something oddly human about that. Give people a powerful new technology and, almost immediately, we use it to solve problems, save time and make ridiculous images for no reason other than it made us laugh.
I do not think that is a bad thing. Play is often how people get comfortable with new tools.
But there is also the uncomfortable part.
The same tools that can help someone communicate more clearly, learn faster, or create something they could not have made before can also be used for the wrong reasons. Scams, misinformation, impersonation, manipulation, low-effort content at massive scale. The technology itself may be neutral in theory, but the way people use it is not.
That is where the excitement gets complicated.
What happens to early-career learning?
We talked about entry-level work. Junior work. The kind of work that, historically, helped people learn.
A lot of the tasks that used to be given to someone starting out can now be delegated to AI. Draft this. Investigate that. Summarise this error. Write a first version. Debug this obvious issue. Generate a quick script. Explain this log.
That can be brilliant for speed.
But I keep thinking about what gets lost when the early struggle disappears.
Some of the most important learning in my career came from the painful bits. The hours, sometimes days, spent chasing a bug through an application or platform during an incident. Reading logs until they stopped looking like noise. Following a thread through systems I did not fully understand yet. Getting things wrong. Asking better questions. Slowly building a mental model of how everything actually worked.
At the time, that work could feel frustrating, stressful and endless.
Looking back, it was foundational.
It taught me more than the answer. It taught me how to think. It taught me how systems fail. It taught me how small assumptions become big incidents. It taught me patience, humility and the ability to stay calm when something important is broken.
That kind of learning is hard to compress.
AI can tell you what might be wrong. It can suggest likely causes. It can explain a stack trace. It can even produce a fix that looks reasonable.
But if you have never built the muscle of investigation yourself, how do you know when the answer is shallow? How do you know what context is missing? How do you develop the instinct that says, “This explanation is neat, but it does not quite fit”?
That is the bit I worry about for people starting out now.
I am not anti-AI. Far from it. I use it every day, and I think the best engineers will be the ones who learn how to use it well. But I also think we need to be honest about the learning curve we are creating.
The five-year plan problem
For years, people talked about a five-year plan. Where do you want to be? What skills do you want to build? What role are you aiming for next?
Right now, that feels harder than ever to answer.
The ground is moving quickly. Almost everyone seems to be using AI in some form, and the pace is only increasing. Every week or two there seems to be another company using AI to solve X, Y, or Z. Some of it is genuinely useful. Some of it is noise. All of it adds to the feeling that the map is being redrawn while we are walking it.
My old five-year plan feels almost non-existent now. Not because I have lost ambition, but because the shape of the industry is changing too quickly to pretend certainty.
Maybe the skill now is not having a fixed plan. Maybe it is learning how to adapt without losing yourself.
For me, that means using AI in a way that supports how I think, while still protecting the parts of the craft that made me an engineer in the first place.
It means letting AI help with expression, structure and speed.
It also means still doing the hard thinking. Still reading the logs. Still understanding the system. Still asking why. Still caring about the difference between a generated answer and a correct one.
The cost of all this intelligence
There is another part of the conversation that should not be ignored either: the environmental impact.
AI does not exist in the cloud in some magical, weightless way. It runs somewhere. It needs data centres, hardware, cooling, power, water, networks and constant expansion.
And this is happening at the same time as the world is already becoming more dependent on electricity.
We are moving towards AI-driven data centres. Electric cars. Electrified homes. More devices, more automation, more connected systems, more demand. Our reliance on electricity is only going to increase.
That does not mean we should stop building or stop innovating. But it does mean we should be honest about the cost.
When we talk about AI changing the world, we should talk about the infrastructure behind it too. The power it consumes. The resources it depends on. The trade-offs that are easy to miss when the user experience feels instant and effortless.
Progress still has a footprint.
Both things can be true
The conversation at the airport reminded me that AI is not just a technical shift. It is a human one.
It changes how people work. It changes how people learn. It changes who gets access to certain kinds of expression. It changes what juniors are asked to do, and what seniors need to teach.
For someone like me, it can be liberating.
For someone just starting out, it could be overwhelming.
For society, it could be transformative.
For the planet, it could be another growing pressure point.
All of those things can be true at once.
I do not think we should respond to that with fear. But I do think we need to respond with care.
We need to make space for people to learn deeply, not just move quickly. We need to teach judgement, not just tool usage. We need to remember that the messy, frustrating, slow parts of engineering are often where the real understanding is built.
We also need to ask harder questions about how these tools are used, who benefits from them, who is harmed by them, and what they cost beyond the screen in front of us.
AI may change the work.
It should not take away the learning.
And it should not stop us from thinking carefully about the world we are building with it.