6 min read

Learning to Play with the Machines

When I was in high school, I first experimented with an Akai sampler, marveling at the multitude of ways to manipulate audio. I learned how to slice a waveform, stretch a note, layer sounds, and apply filters and EQ to replay and drastically change the tonal characteristics. Each tweak had the potential to generate something unique and captivating—a sound never heard before.

Using samplers to mangle audio reminds me of the creative possibilities available using AI today. Just as it was possible to process sounds into something new, it’s possible to combine data and algorithms using AI to generate images, video, text, and code that not only surprises but has its own distinct essence. However, while these possibilities excite me, for some others, venturing into the realm of AI isn’t as fun sounding.

Current Barriers to Adoption

Let’s start with the obvious: the world is heavy right now. Between political drama, economic instability, and an endless stream of global uncertainty, even people who are doing “fine” are often just trying to get through the day. Their focus is largely divided between managing daily responsibilities and seeking relaxation. Engaging with new technology demands cognitive and emotional resources that are often depleted.

The Emotional Weight

There’s a lot of emotional baggage that comes with AI. For some, it doesn’t feel exciting—it feels dystopian. Instead of being empowering, it feels like it’s eroding the very things that give people purpose, security, or even joy. It can feel like a threat—not just to people’s livelihoods and craft, but to their sense of identity and their legal rights. For others, playing with tech just isn’t that interesting—or how they ever wanted to spend their time.

When First Impressions Fail

Even when people do try out new AI tools, the initial experience often isn’t great. I’ve demoed tools like Cursor—which I imagine would be an absolute gift for many folks working at tech companies—and discovered some never opened it again. Using an IDE can be daunting if you’re not accustomed to spending your days in one, so it’s perfectly reasonable for people to move on if they don’t quickly connect with it.

Cursor—and many other AI tools—are evolving at breakneck speed. This rapid development not only introduces genuine innovation but also brings bugs, broken features, and unpredictable changes. Getting the most out of these tools requires frequent use and unconventional thinking, best achieved through playful tinkering. This approach can uncover new ways to utilize the technology that might not be immediately obvious.

For those accustomed to stability and who already excel in their current roles, these tools might initially seem like a step backward. Similarly, if you’ve previously struggled with learning technical skills, it’s easy to feel that these tools aren’t suitable—especially now, when time and energy to learn something new are particularly scarce. Fortunately, there is a wealth of tutorials available online, both free and paid, that can help bridge this gap quickly. These resources are designed to flatten the learning curve, offering step-by-step guidance that can accelerate your proficiency and confidence in using these tools.

Diverse Reactions

In my circles, I’ve noticed a divide: generalists tend to be more inclined to explore and experiment, while specialists tend to be more skeptical—often focused on where the tools fall short.

For example, most of the experienced engineers I know are used to picking up new technologies in a very deliberate way: reading the docs, diving into the language, working through tutorials. They’re thoughtful and methodical—and usually not interested in engaging with something in a half-baked manner. When AI spits out buggy code in a language they’ve mastered, it’s not just frustrating—it feels amateur and they don't want any part of it.

On the other hand, I’ve also seen folks who are newer to a field, or simply less technical overall, who just want to jump in and make things. For them, AI is a creative unlock. It allows them to accomplish in a short amount of time what used to take weeks or months—work they didn’t have the time, patience, or ambition to tackle before. They’re not chasing perfection—they’re chasing momentum. For them, AI unlocks capabilities they haven't developed on their own. Whether the code is well-written or not doesn’t matter—they don’t know, and frankly, don’t care.

Both of these reactions, and a spectrum in between, are valid. One person sees AI as a toy or a threat. Another may see it as the beginning of Something Big. General purpose technologies reshape entire industrial sectors, so it’s no surprise they spark strong, polarized responses.

For those with deep experience, choosing not to engage with tools that feel unpolished isn’t wrong—it’s often a sign of discernment and maturity. But this moment also calls for shoshin—a beginner’s mind. A mindset rooted in curiosity, openness, and a willingness to explore new ways of working, even when things are still rough around the edges.

How to Build Intuition

No matter where you stand, one thing is clear: these tools aren’t going anywhere. For those expecting plug-and-play solutions, it might be surprising to realize that their real power lies in unlocking entirely new ways of thinking and working.

Using AI tools effectively often requires reimagining familiar workflows, adapting to shortcuts, and embracing alternative paths to the same outcome—sometimes faster, sometimes messier. Doing your best work in this new landscape requires not just learning how and when to use these tools, but also being open to how they can reshape your creative or professional process entirely. As you navigate this new landscape, don’t be surprised if you find yourself chasing your tail at times.

That kind of adaptation takes trial and error: testing things out, hitting walls, stepping away, and coming back later. It’s less about judging a tool’s value in the moment, and more about building the muscle to understand where it fits, how it works, and when it truly shines. Have fun with it. Be playful.

Simon Willison’s ongoing experiments prompting vision models to generate images of a “pelican riding a bicycle” are a great example. He’s using a lighthearted, familiar prompt to consistently test what the latest models can do. This approach to monitoring advancements in state-of-the-art image generation models has been both delightful and engaging.

Consider staying open—even if your first experience wasn’t great. I’m not suggesting you throw the same problem at every new model or force AI into everything you do. But it’s worth returning with fresh challenges and noticing what works and what doesn’t. Over time, you’ll build intuition—for when to lean in, when to walk away, and which tools are the right fit for the job. More often than not, what you discover will surprise you.

Corporate Resistance

AI has the potential to reshape how we build software products. But embracing that potential isn’t just an individual challenge—it’s a cultural one.

As one experienced product manager recently put it (paraphrasing):

"AI can and should transform how we build software. But many companies—especially larger ones—aren’t embracing that shift. There’s still a strong preference for specialists, even though AI increasingly rewards generalists who can move quickly, adapt, and experiment. In this new era of software development, I find that especially striking. Over the past two years of interviewing and job searching, I’ve consistently seen employers leaning in the opposite direction—which, to me, feels short-sighted and like a missed opportunity."

I responded:

"Totally. Mature companies tend to suffer from continuity bias, and many employees are incentivized to protect their existing roles. In contrast, startups thrive on risk-taking—there’s a cultural push to embrace change, experiment, and bring others along. In bigger organizations, that kind of mindset often gets crushed by the existing culture before it can gain traction."

The current climate favors curiosity, boldness, and speed over caution and rigidity. Succeeding in this kind of environment isn’t just about mindset—it’s about making things happen. With access to tools that offer real leverage, the advantage goes to those who are willing to experiment, move fast, and turn possibilities into reality. It means getting out of your comfort zone, trying new approaches, taking smart risks, and staying open to change.

What Lies Ahead

I get that this perspective isn’t for everyone—and I don’t expect universal agreement. I generally lean towards optimism, even though I’m aware that I am likely underestimating the potential drawbacks. However, one thing is certain: the transition towards widespread use of AI is already in progress. It’s not slowing down, and I’m genuinely excited about what it enables me to do today and in the near future—even if not everyone feels the same.

Over the next 1–3 years, I expect coding models will outperform most top engineers when it comes to writing high-quality code. We’re nearing a point where a generalist can prompt their way to fully functional software products—front end, back end, design, marketing, and deployment—without needing deep expertise in any one domain.

That’s a significant shift. It doesn’t replace the need for experts or devalue craftsmanship—but it does raise the bar for what individuals can build on their own. And it’s likely to reshape how we work and where we choose to apply our time, energy, and human judgment.

For me, it’s an exciting time to adapt—technically, emotionally, and creatively—and I’ve chosen to lean in. I’m enjoying exploring how to use these tools not just to work smarter, but to play, experiment, and express myself in new ways.

That’s my goal—to enjoy the process, remain curious, and explore new ways to play with the machines.