Trending
Things have been blissfully “quiet” from my side. Today though, I need to “write to you guys about something”. I’ve been putting it off in the hopes that either my perspective would change or I would be able to put it in a way that would land best, and quite frankly, neither the former nor the latter have transpired-so here goes: the tech industry’s response to “vibe coding” is wrong. I’ll pad this and explain myself, but ultimately, I believe it is acting as a disservice to the software industry to have this reaction to what’s been happening of late.
The AI Storm Is Here. We’re All Already in It.
No one reading this will be blind to the devastating effects AI can -and now does indeed- have on creativity and creation. Be it images, as in the latest Ghibli-style scandal of last week; writings, as in the Meta stealing our books scandal (BTW, that got two of my books-and countless articles about banks for some reason -so if you would like to support despite how Zuck robbed me, please buy them for real from Amazon: People Before Tech is here and Tech-Led Culture is here); or code, when it comes to having inserted itself into tech business strategy, practices, and jargon; or any of the many other ways in which human toil and spark is at risk, AI’s presence is undeniable and we need to reckon with its arrival and think ahead.
Hopefully, there is a well-thought-out ahead we can design, where the ethics of our collaboration allow for protecting existing and future creation with a valuable “by human.” Hopefully, we get to collectively, as a species, sit down and carefully consider what we hold dear and valuable-and hopefully, we design and quickly execute a path where the human-machine interaction is balanced and moral. All this hope is a whole other conversation, though. What we’re here to talk about is the software industry’s reaction to AI’s admittedly dangerous and chaotic arrival.
Vibe Coding Isn’t the Problem-The Gatekeeping Response May Be
IT, Business, Product, Leadership, HR, Ops-you name it, there is no area this AI foray into our status quo doesn’t affect in today’s tech businesses, and it is only going to intensify at warp speed but the technology community’s response has been abuzz with near-unanimous condemnation.
In a nutshell, there are two hypotheses that most seem to be thrown around as cause for debate in a myriad of forms: the “Will AI replace programmers?” and the “AI-generated code is garbage” one. The former is interesting and remains to be seen, but the latter is the crux of the issue we discuss today and the backdrop of tens of articles around LinkedIn.
The tech community is appalled at the number of applications and features created by neophytes that enjoy the low barrier to coding entry that the explosion of AI tools brings to every Joe and Jill. They deplore the lack of reliability and sustainability. They already saw apps with no genuine back-end, appalling security, lack of pipeline and ability to maintain or upgrade, and a complete lack of understanding of software architecture principles reflected in a few loud disasters in the past year. Some of the vibe code has already famously backfired, and AI is still 70% “hululu” (hallucinatory), so the outrage is natural, the mistrust is natural, the disdain in the face of broken pipes and clear future mounting risk of vibe-code making it to production and infecting every backend is a natural reaction.
What We’re Really (Rightfully) Afraid Of and What Excites
Me? I’m not a techie per se, as I can barely write a few lines of code-I would have said “without Googling” in the day, but that feels superfluous now!-but after 25 years at the intersection of tech and business, I see all the many facets of these separate perspectives from a uniquely personal vantage point, as I’ve had each of those hats on at one point or another. You could say I know why vibe coding blows and why, but I know why it rocks and why that is too.
As a creator who is losing income to AI impersonation and content assimilation every day and sees what it does to the creative industry, I loathe the current state of affairs with a passion, of course. As someone who doesn’t think AI will replace us in the software industry-or anywhere else completely, any time soon-I still think we need a valid “ethical contract” in place ASAP. As an intrapreneur and founder of software companies with DevOps fetishes and SRE magic, of course I am sharing much of the same concerns and am appalled by the ever-growing mass of applications churned out with no concern for the bigger picture or how sustainable or scalable any of this instant code a machine once wrote live in front of our marvelling eyes really is, once we deploy it out there.
But also, you know what? Ever the ableist, self-tormenting ADHDer, my career never stayed firmly on either the business or the tech side of things-or I would have joined into the stream of protesting and accusing pieces. But I can’t because I’m all Product at heart. And as a non-programmer CPO who has led multiple software product teams, I am feeling freed, I admit. Yes, I’m churning out prototypes (and “prompt-based-project-initialisations”™?!), but beyond that -which is magic Invision would have lent my team as far back as pre-pandemic-, I’m in love with the possibilities, the speed of shared visibility, the capability of common language, and ultimately yes, the capability to break past the IT gatekeeping we have been suffering for years, where we met more “computer says no”s than we can count. As that Product Owner? It was about time!
There I said it.
The Role Developers Should Play in the AI Age
As a maker of software for people, I’m ecstatic about the possibilities we have to co-create with the damn thing-once we get our heads, hearts, and regulatory systems in place, that is. But beyond that, I’m even more excited about how quickly and suddenly humanity will have to realign under its truly human attributes, as the rest of them have ceased to be their apanage. How clear as day and uncontested this need is, even to the blindest of tech leaders of yesterday. How even those that still think they can afford to lag on the human work, as it’s a “fluffy nice-to-have but not a priority,” have started to realise increasing the wellbeing and emotional intelligence of the humans they employ alongside AI is the only way to move forward.
But I get it. Who doesn’t hate luddites and petulant new entrants to an industry? Who wants to feel like they have spent years perfecting their craft only to have something-not even someone, the impunity!-come in and churn out seemingly comparable results to the naked eye of the casual observer, unable to see the dangers and lack of quality? Of course it’s infuriating and painful. More and more ways in which that is true are coming out by the day, and if you are interacting with OpenAI routinely on a daily basis (and let’s face it: We. All. Are.), you may have your own strong-held beliefs about its uses and its need, but you can’t help but feel the overall sense of spent creative energy and extinguished human spark that its influence has on your respective industry.
Let’s Elevate the Conversation Not Just Gatekeep the Craft
That said, let me spell it out to the techies -at the risk they hate me for it: when it comes to this pivotal moment in the life of your craft, you can do better than this outrage. You have to elevate your position in this dynamic from Critic to Orchestrator. Down here in the trenches, where you point out all the dangers down the line, you are placing yourselves in a mere admin, config and maintenance role-which you may be free to corner (should we ignore the bevvy of tools and frameworks that successfully do that today and could adapt to include monitoring or fixing these lawless projects it generates, the Zero-Security, the AI-disasters and vibe-code victims)-but between the cornering or the tools, we can all deal with the “maybe it’s fast but they break immediately down the line, deploy to prod with an All-AIgen product and you’re doomed” argument easily if we’re honest.
And then what? No offence to that seminal side of things that needs sorting, undoubtedly, but that’s ultimately a minor and temporary objection-and one that, again, doesn’t place software crafters into the position they should assume with AI: architecture. That’s where the invaluable “experience” shows. That’s where the “gut” matters. That’s why the machine will never imagine the system down the line as the human will. That’s why it will never have the moments of sheer genius that come when writing (code or prose!) whilst in flow, in the zone. That’s where the spark is in software development, and that’s what we must preserve!
Stop making it about how everything AI makes is automatic tech debt. Firstly, that’s untrue. And secondly, when have we started measuring how much of the code any of us ever wrote amounted to value or created tech debt? When have we even started genuinely auditing our tech debt routinely, anyways, with the exception of some of the DORA champions? Is producing clean, tested, and never-debt-making code being produced by humans only really the hill you want to die on, or do you want to take it further and-instead of pointing at it saying “this may increase our overall tech debt”-say: "Let’s see how to mitigate that, software crafters are architects, and you’ll need them to vibe-code.”
So if the IT community’s response would have been “Stop going at it blindly, you can vibe code all you like if you’ve had a software crafter help think some of the architecture through, and if you can have a pair of human eyes double-check any AI decision point or much will break,” then I would have understood. Perhaps there would have been an argument for only “assisted-vibe-coding” (or even “pair-vibe-coding” where a dev would sit next to the vibe-code-writer and they would together converse with the machine to build-why is that not a thing?!?), but that hasn’t happened, and this rejection from your side now feels more like a grumpy-old-man’s stomp that says: “What you got from Cursor or made on Replit is not solid and stable sweetie, it’s not a production-grade product with high reliability, this will create untold more tech debt, stop touching this stuff, you should buzz off because you’re not a developer and you aren’t really coding when you have your little chats with your ChatGPT over morning coffee.”
We can and should do better than that.
Look, I think we have to hold on to the ever-lovin’ core of who we are as individuals and teams and seek ways to better that core right away, if we want to become and remain the AI sherpas and the creators we deserve to be in this long-longed-for future economy-where currency and value will shift to get us paid for being uniquely human and doing only what we love, while AI does what we loathe - but meanwhile, wishing “vibe-coding,” “insta-authors,” or “self-professed-designers” away, or hoping AI will confine itself to empowering these “hobbyists” is not the way to go.
So, my advice tech community for the little that it may be worth? Start vibing with the times and ascend to your elevated position of co-creators, guides, and architects of the coding-not its sole-makers or gatekeepers.
Related: How We May Have Killed Agile – And How to Revive It in 2025