Building a startup from an idea can be an exciting adventure that tests resilience, passion, and commitment. It’s about creating something entirely new and adopting the right mindset to turn it into a successful reality.
With “Becoming Crono” we’ll revisit our starting point through an honest and captivating memoir written by our CTO, Marco Maddiona.
You’ll have the opportunity to witness firsthand the steps that propelled Crono into a startup with +150 customers all over the world. This will encompass the obstacles we overcame and the choices we made. Have you ever wondered what it’s like to develop software while building a business?
Enjoy our new episode of Becoming Crono!
Ep. 13 - Reflections on a Summer of Software Development
Hi Croners, I’m Marco Maddiona, CTO & Co-Founder of Crono, welcome back! If you missed the previous chapter of this story, I suggest you catch up here: The Story Behind Crono 2.0: Growing Through Change (Part 12) before moving on.
Let’s go!
It’s been a while since the last chapter. I broke my every-three-month rhythm of publishing an update. Pretty sure the world didn’t suffer from it, but personally, the idea of being “late” to this self-imposed appointment really bugged me.
As is tradition at this time of year, I’m writing another entry in the Reflections on a Summer of Software Development series (I recommend catching up on the earlier episodes if you haven’t read them: 2023, 2024, 2025).
Late August, the long tail of one of the hottest summers ever. A few summer storms have made the temperatures a bit gentler, though we’re still above 30°C during the day. The end of summer is marked mostly by people coming back from vacation and by companies slowly, and without much enthusiasm, resuming their activities.
There are also the late vacationers, who can’t wait to leave after an entire summer spent suffering indoors with the air conditioning on. They’ve been training for months, studying the best angles for selfies and Instagram stories, and they’re itching for their turn to hit the share button in front of a breathtaking sunset filmed from a restaurant with a sea view.
Amid all this, there’s me, CTO of Crono, who has spent another summer of software development together with my team of dragons, and now I find myself here taking stock of these past months.
First of all, I want to immediately reassure the reader about the most important things: I did go on vacation too. A beautiful seaside vacation with my partner Giulia and our one-year-old son, Massimo. We went in mid-June (like good new parents), a time when the beaches of western Sicily are a little less hot and decidedly less crowded.
In recent months Massimo has started actively interacting with the world around him, and since he’s gained momentum crawling and pulling himself up to stand, taking him on vacation has become a bit more demanding. Fortunately, like me, he loves being at the beach and swimming in the sea, so we still found ways to relax.
It’s well known that the main problem with a vacation is coming back from it. That’s why I don’t go often. Returning to daily life after days of sea, food, and relaxation is a torture we don’t deserve.
Fortunately I like my job, developing software, so the rest of the summer went by quickly: most clients scale back their activity and many colleagues go on vacation, so there’s much more time to focus and complete important work that requires more time and calm, and so gets pushed to this period.
Truth be told, it’s been several months since I last wrote a line of code myself.
Maybe three, maybe six, I don’t even remember the last time I had to write code alone, the way I always did for the last 20 years. This might sound strange to people outside the field. Giulia might roll her eyes and think I now just sit in front of the computer all day doing nothing.
And yet that’s how it is. Today, to develop software, you can give more or less precise instructions to artificial intelligence, and it can handle a significant part of the work you would have had to do yourself until recently. The generative AI revolution started a few years ago, but the progress made in my field over the last year has been incredible.
Beyond the models becoming ever smarter, the tools and their integrations have improved too. Claude Code (and its equivalents) has access to your entire platform’s codebase. It can analyze any line of code and assess the impact of a change in a short time, or suggest the best solution to fix a bug, or even implement a new feature from scratch.
Its ability isn’t limited to code; it can interact with your work tools to gain more context: read specs, consult documentation, search the web for how to integrate with an external platform, and lay out a work plan to do it. Finally, it can send a message to colleagues to share the work you did. Or, well, that it did.
The evolution has been so fast and disruptive that I almost struggle to remember how I used to work before, even though I did it that way for at least ten years professionally.
I do remember, though, the stress and effort of solving complex bugs: hunting for logs, retrieving tracked events that were often missing. Cross-referencing code and trying to reconstruct what had happened. Sometimes it wasn’t a single user command, but several concurrent commands from different sources causing the problem. The human mind isn’t built to cross-reference multiple events in parallel and spot potential conflicts. You can get there, but it takes deep system knowledge, experience, ingenuity, and time.
AI, on the other hand, can access all the logs, the infrastructure, and the code, and is able to cross-reference events in seconds or minutes. It has already seen all these cases before you have. It doesn’t need to study your system in depth. It’s able to recognize patterns and apply them.
Or I remember the complexity of having to develop a big feature that touches various parts of the platform. I’ve done it many times; it used to take several days or even a few weeks. But more than the complexity of pulling off something like that, what mattered was dedicating all that time to a single task without having to context-switch between other requests.
This summer, together with my team, we developed several important new features, working on them in parallel with other smaller jobs. After giving the AI documentation, specs, and context, we could move on to something else and come back later to check the work done.
So I spent a good part of the summer managing several things in parallel, clicking continue, confirm, and allow until the work was finished.
Meanwhile, my home office is increasingly turning into Massimo’s room, and I’m starting to sense a subtle threat: my desk here is on borrowed time. His toys have by now taken over the entire room, and soon we’ll need to make space for a crib now that he’ll need to start sleeping in his own room.
Sometimes, when I’m absorbed in work, without noticing, I find him standing next to me, looking at me with his wonderful smile. He tries to open the desk drawers to get my attention while watching me in silence. He lets me know he wants to be picked up by insistently touching my elbow, and I can’t help but give in.
Once I pick him up, that’s not enough for him. He’s watched me for a long time (time is relative, especially for a one-year-old :)) and now he knows what to do. He starts touching the keyboard and mouse. With a few shortcuts I can quickly open a clean notepad where he can write his own first lines of code too, without causing any damage.
He’s thrilled watching the numbers and letters he’s typing appear on the screen. The spacebar and the enter key move the cursor right and down, and he follows it with his eyes, aware that the computer reacts to his commands. Every time he manages to switch windows with the mouse, he lets out a cry of astonishment and looks at me as if to say: “Dad, did you see what I can do!”
The biological circuits of my heart skip a beat. I’ve never thought about imposing my passions on my son, even though I’m convinced every father can’t wait to share them. I like to imagine that, somehow, they might be passed on through the bits that make up genetic code.
It’s definitely too early to draw these conclusions, but I can’t help wondering what the software developer’s job will look like in twenty years, when, and if, it’s his turn.
We’re living through a revolution with an impact similar to that of the Industrial Revolution in the 1800s. One of the striking differences lies in how quickly this change is happening.
For example, automated textile machines transformed, within a few decades, the way clothes were woven,something that had existed for millennia.
Software developer, as a profession, has existed for about half a century, and in recent years we’ve started asking ourselves how deeply it might be transformed, or replaced, by the machine and all this thanks to the ingenuity of other developers who are shaping and empowering this technology.
I, and many others like me, are actively living through this revolution. We’re part of it, we’re the first to benefit from it and the first to be subject to it.
To do the work I’ve done so far, I studied computer science for 8 years, between high school and university. After a few years of work experience, I felt ready to start from scratch and develop software with the goal of selling it.
Today my skills and experience help me do my job effectively and with sound judgment. I know how to recognize problems, make architectural decisions, organize the work of various people on large-scale projects, and I have enough experience to ask AI to develop the appropriate software.
Artificial intelligence has a sneaky way of infiltrating our work. First it shows you what it’s capable of, creating the WOW effect. It’s an easy realization that to develop a new API, all it takes is giving it the right input and getting the finished code, rather than writing it by hand. Then it makes you sense it can do something more: it can access logs to fix bugs, prepare code for colleagues’ review, write documentation to share with the rest of the team, and it can automatically send a message to communicate the work done.
All you have to do is give it confirmation and permission to do it again next time, without you having to ask.
It’s clear that every choice here is obvious. Less manual work, the ability to handle more things in parallel, why wouldn’t I give it my consent? Ultimately, what matters to me is the final result, not the way I get there. In short, less stress and more productivity.
A bit like little Massimo, who sits next to me and, without making too much noise, watches and learns, AI does the same thing, but with a capability and a scale of power beyond human possibility.
By now, to build software it’s no longer strictly necessary to have studied as a developer. Just give commands to AI, and an application can come to life. Sure, having understanding and knowledge of what you’re doing still helps, especially if you want to build an application meant for external users.
But for how much longer will these skills be necessary? What will be the impact on our society once building software is within anyone’s reach, and years of study and experience are no longer needed?
We still can’t know. I tried to explore some of these questions during my last talk at Brescia Startup (“Building startups in the age of AI”). Four months have already passed, and some of the themes may already feel outdated given how fast the technology is evolving. But if you’re interested in digging deeper into the topic, I touched on several points I find worth reflecting on.
Lately you often hear the boiling frog metaphor cited. The frog is in a pot where the water is slowly heating up. At first it swims happily, then slowly, without it noticing, the water reaches a temperature that’s lethal for it.
Even aware of the unknowns regarding this technology’s impact on our future, who are we today to decide not to take part in it? The advantages are too obvious to just stand by and watch, and the questions it raises concern all of us.
At Crono, beyond supporting software development, we’re introducing AI agents to automate other business processes. The goal is clear: to spare people manual, repetitive tasks, make more complex ones easier to carry out, provide up-to-date data and context exactly when needed, in order to free up time for more strategic activities, where AI (still) can’t reach.
I watch AI advance on its path, increasingly covering, in our work and in our lives, activities that until recently we considered the exclusive expression of human creativity and ingenuity.
I still can’t tell whether all of this will make us freer, more productive, or simply more dependent on tools we can’t fully understand.
Another summer is ending as I write these last lines. This morning another storm cooled the air. Massimo is crying because his mom just scolded him for trying, yet again, to climb down head-first from the big bed.
I go back to the computer to develop software, or to explain to AI how to develop it for me. Meanwhile, Massimo keeps watching me from right there, still with tears in his eyes after the scolding, marveling at how a computer can respond to his commands.
Who knows if, twenty years from now, he’ll still remember that it all started this way.
Not sure where to start?
Start with the Playbook.
Download our free Sales Playbook 2026 and learn how top B2B teams structure their outbound process.
BANT, SPICED & MEDDPICC frameworks- 30+ ready-to-use sales templates
- 2026 outbound & pipeline benchmarks
Don't sell the product. Sell the change the product makes possible.
Build the process before the pipeline.
