The Developer’s Dilemma in the Age of AI
Like many developers, I’ve felt a creeping sense of gloom about our profession’s future. Everywhere I look, headlines boast about AI writing code and companies slashing engineering jobs. It’s not just hype – even top tech firms admit a significant chunk of their code is now machine-generated. Microsoft’s CEO recently revealed that 20–30% of code in the company’s repositories is produced by AI, and Google’s CEO said AI is generating over 30% of Google’s code. These figures are staggering. While AI-assisted coding promises productivity, it also sparks a frightening question: if AI can do a third of the coding, will a third of us lose our jobs?
That question feels alarmingly real amid the ongoing waves of tech layoffs. Over 150,000 tech workers lost their jobs in 2023 alone, and tens of thousands more have been cut in 2024 and 2025. Even highly experienced engineers – people with 15+ years in the field – are finding themselves unexpectedly out of work. In many cases, companies are explicitly citing AI and automation as reasons for these cuts. A recent report noted that some firms loudly tout “AI-first” strategies to justify layoffs, and tech luminaries have fueled anxiety by predicting AI could eliminate a huge number of jobs in the coming years. No wonder morale is low. It’s a sad state of affairs when developers who once felt secure in their career now worry their livelihood is being eroded by the very tools we’ve created.
But is our fate truly sealed? Or is there a way for developers to thrive alongside AI? I recently found a hint of hope – surprisingly – in the concept of play.
A Glimpse of Hope: Learning to Play with AI
I recently attended an AI meetup in Manchester where a panel of experts from various industries discussed how AI is being adopted. One speaker, a representative from Bank of New York Mellon (BNY Mellon), said something that profoundly changed my perspective. He described how his organisation approached AI not as a top-down mandate, but as an opportunity for empowered experimentation. They run internal programs – essentially competitive hackathons – where employees across departments get time and guidance to play with AI and build solutions for their specific problems. In other words, they create a structured space for people to explore AI capabilities on their own tasks, almost like enforced play at work.
What he said next stuck with me:
We found that people need to find their own way of working with AI – their own relationship with it. Everyone’s mind works differently, and you need to work out what works best for you. And the only way to do that is through experience and play.
Hearing this was a revelation. It cut through my despair and made me rethink how we, as developers, might coexist with AI. At first, I wondered if he was anthropomorphising AI by talking about a “relationship.” But I soon realised he was right in a very practical sense. AI interfaces – whether a coding assistant, a chatbot, or an agent – is the same for everyone, it always involves prompting and it’s a neutral tool. The variable in the equation is what happens between the chair and the keyboard – us, the users. Two people can use the exact same AI tool and get very different results. Why? Because we each communicate with it differently. The secret sauce is in how we prompt, instruct, and collaborate with the AI. That is a skill – almost an art – that we each have to develop for ourselves. Yes, AI can create prompts itself but the starting point always has to come from somewhere.
Band Of New York’s approach actively encourages people to “play” with AI to discover what it can do and create agents from that play. They even extend this mindset to the C-suite: the speaker mentioned they are guiding board-level executives through the same playful process of building and using AI agents. This idea of play as a learning method resonated with me deeply. It dawned on me that my own successes and failures with AI largely came down to how much I had experimented with it.
Why Some Developers Get Better Results from AI
I started reflecting on why I often get more useful answers or code from AI than some of my peers, even when using the same tools like ChatGPT, Cursor or Claude (Desktop or Code). I’ve met developers who tried an AI coding assistant briefly, got mediocre results, and concluded it’s useless and even, dangerously, not a threat. Others complain the AI writes poor code or isn’t reliable, so they give up. Meanwhile, I’ve been using these tools to great effect – not because I’m smarter, but because I’ve spent hours playing with them, tweaking my prompts, learning their quirks, and figuring out how to coax the best output. In the process, I’ve essentially learned how to “talk” to AI.
It turns out this experience-driven gap is common. Surveys show that while over 70% of developers are now using AI tools in some capacity, a much smaller subset feel they have unlocked substantial productivity gains from them. For example, one large poll found 44% of developers use AI coding assistants daily, yet only 31% say these tools make them more productive (dev.to). That’s a huge difference between usage and actual benefit. Why the disconnect? I suspect it’s because effectively using AI requires new skills and a willingness to experiment, which not everyone has embraced yet.
Research and industry experience back this up. A recent study of nearly 5,000 developers using AI tools found a modest productivity boost on average – about 13% more code written per week (ksred.com). But interestingly, the gains varied widely depending on the developer’s experience and how they used the tool. The real insights came from those who dove in and learned to integrate AI into their workflow. As one AI engineering expert noted, “the learning curve is real.” It takes at least a week or two of struggling and fiddling to understand how an AI coding assistant fits into your process. It’s not a plug-and-play magic solution; you have to learn prompt engineering, provide context, and review the outputs carefully. Most developers aren’t initially willing to invest that time, so they conclude the AI isn’t worth it (ksred.com). In other words, many give up before they’ve truly learned how to use the tool – akin to picking up a musical instrument once and abandoning it because you can’t immediately play a song.
By contrast, developers (and teams) who do invest the time to play and learn can reap major rewards. They figure out the right way to phrase requests, when to trust the AI and when to double-check, and how to incorporate AI into bigger tasks. Some companies have reported eye-opening improvements – for instance, OCBC Bank saw a 30–35% productivity boost after training their developers to properly use AI coding tools (ksred.com). And tellingly, non-technical stakeholders are now actively looking to hire engineers who are adept at using AI, viewing those who don’t as being at a disadvantage in the modern workplace (ksred.com).
All of this reinforces that the critical difference isn’t the AI itself, but the human using it. If you approach AI in a rigid, one-dimensional way, you’ll get lackluster results. But if you approach it playfully – with curiosity, persistence, and a willingness to try different angles – you unlock its potential.
How can you “play” with AI to become better at working with it? Here are a few strategies that have worked for me and others:
Treat the AI as a collaborator, not a code monkey: Ask the AI to explain its solutions or engage it in a dialogue. For example, I’ll prompt, “Here’s what I’m trying to do… how would you approach it?” This turns the interaction into a back-and-forth exploration rather than a one-shot query.
Experiment with prompt phrasing: Don’t hesitate to reword your request or provide more context if the first answer isn’t good. Think of prompts as levers you can adjust – be specific, set constraints, or even try a creative role-play scenario if it might elicit a better response. Over time, you’ll learn which approaches work best for you.
Build small projects or challenges for yourself: A great way to play is to give yourself a mini-hackathon. Pick a toy problem or a task at work and see how far you can get using an AI assistant. Allow yourself to fail and learn. The goal is to familiarise yourself with the AI’s capabilities and limitations in a low-stakes setting.
Learn from others: Just like playing a game together, watching how other developers use AI can teach you new “moves.” Many developers share prompt tips and workflows online. Seeing an example of how someone gets a tricky unit test written by an AI, or how they use AI to generate boilerplate, can inspire your own experimentation.
The key is to cultivate a playful mindset: approach AI with the same inquisitiveness you might have had tinkering with a new gadget or playing a strategy game. In fact, psychological research suggests that this kind of playful experimentation can unlock creativity and resilience in problem-solving (nifplay.org). When we engage in play, we enter a state of mind that’s open to new ideas and not afraid of failure. That is exactly the mindset needed to harness AI effectively.
Embracing Play: A Developer’s Lifelong Advantage
This discussion of play made me reflect on my own journey into computing. I realised that play has been the driving force of my learning since childhood. I was the kid who got hooked on a computer at age 7 and never stopped tinkering. I didn’t learn BASIC on my ZX Spectrum because someone told me to it for a job – I did it because it was fun. I needed to know how this thing worked, so I pushed every button, tried writing little programs, and gleefully broke and fixed things. That playful curiosity carried through to my adult life (for better or worse – spending late nights experimenting with tech isn’t exactly a mainstream social activity!). It gave me an intuitive comfort with computers and now, with AI. When a new API or AI model comes out, my first instinct is to play around with it.
Many great developers share this trait – they learn through play and maintain a childlike curiosity. Sadly, not everyone does. It’s often said that somewhere on the road to adulthood, most people lose that natural inquisitiveness and playfulness they had as kids. We become focused on “doing it right,” sticking to routines, and avoiding mistakes, and we shy away from the open-ended play that once came naturally. In the professional world, play is sometimes stigmatised as frivolous, when in fact it can be incredibly productive. As Dr. Stuart Brown of the National Institute for Play famously said, “The opposite of play is not work — it’s depression.”. In other words, a lack of play and joy in what we do can drain meaning and creativity from our lives. I think many of the burnt-out, demoralised developers (myself included, at times) have simply lost the play in our work.
The Bank of New York expert’s insight was that every person’s mind works differently, so each of us must find our own relationship with AI. That is effectively an invitation to play – to discover through trial and exploration how AI can augment our individual strengths and compensate for our weaknesses. This idea aligns with research on adult play, which emphasises that identifying your own “play style” and integrating it into your work is not only possible but indeed necessary for creative, fulfilling work. We each have different activities that put us in a state of flow or spark joy in learning; for some it might be building toy apps, for others it might be competitive programming challenges, or collaborative brainstorming. Tapping into that personal play style can transform how we adapt to new technologies.
For developers, playing with technology is second nature – it’s why many of us got into this field. So in a strange way, the rise of AI might rekindle that spirit across our industry. Maybe we'll finally have more time for play. Instead of being the generation of developers that is rendered obsolete by AI, perhaps we will be the generation that embraces playful co-creation with AI. We’ll still be there, debugging and tinkering and innovating, but now alongside a powerful new partner. Our value won’t just be in the raw code we produce, but in knowing what to build, why to build it, and how to guide our AI tools in doing it. Those are things that come with experience and curiosity – things born of play.
The Next Generation: A World That Never Stops Playing
All this leads me to think about the next generation of programmers and workers. Today’s kids, my daughter included, are growing up with AI as a given – perhaps their first “computer” friend will be an AI chatbot, and their first programs will be written with the help of AI copilots. Will they ever lose the playful mindset that we tend to shed in adulthood? Maybe not, because they won’t really have a choice. The world they’ll inherit will be changing even faster, and continuous learning will be more of a way of life. To keep up, they’ll have to keep that curiosity alive. In a sense, they might never stop playing, because interacting with AI and the new technology and products created with it will constantly require experimentation and adaptation.
What kind of world would that be? I’m hopeful it will be a better one. Imagine if everyone managed to maintain the natural curiosity and playfulness of their childhood self into adult life. We might have far more people in every field who are adaptable, innovative, and unafraid to learn new things. Instead of dreading new technology, they would approach it like a new playground – something to explore and incorporate into their skills. AI would not be a threat, but a playground for creativity and problem-solving.
For us current developers, standing at this crossroads of AI advancement, the message is clear. We shouldn’t despair that “AI is coming for our jobs” and give up. Instead, we should do what we’ve always done at our best – roll up our sleeves and play with the technology until we master it. By doing so, we reclaim some control over our future. Yes, the game is changing, but it’s not game over. In fact, it might just be game on. And those of us who remember how to play – how to learn for the sheer joy of discovery – will lead the way in this new era.
In the end, the developers who thrive will be the ones who forge their own relationships with AI, built through curiosity, experimentation, and yes, play. That gives me hope that our skills won’t be erased at all – they’ll be amplified. And perhaps, as we mentor the next generation, we’ll ensure they never lose that spark of inquisitiveness. A world where adults never forget how to play and continuously learn? That’s a world I very much want to see – and one that we as lifelong “players” can help create.
Top comments (0)