Is The Vibe Killing SaaS?

In my spare time, I occasionally love to create stuff with code, I say create, it’s more “hack around with“, and it could be anything from a weekend with COBOL to becoming a certified HubSpot developer (for a laugh), I feel it’s creative, like why I like to write. It’s the quiet pleasure of creating something from nothing.

I just dabble now, but was once classically trained; I have certificates, I knew my object orientation from my procedural block-structures, and am a faded Perl slinger.

And of course, like writing, AI has arrived to either assist or replace us in this pursuit, and now, according to the numerous stories I am reading on LinkedIn, anyone can do it, and commentators claim this will kill SaaS software.

But… will it?

In the early days of web development, and the rise of web content management systems, there was the build vs buy debate, should you build your own web applications, or implement something off the shelf and build on that?

Long story short: buy won, whole software market categories were formed, and in due course, on-prem computing escaped and took flight to the cloud, becoming SaaS.

Why?

Because writing applications was relatively easy, even back then. The fucking hard bit is security, performant/scalable hosting, documentation, technical debt, integrations, data modelling, etc. - the basic ball ache of application adulting that adds no value to the application or the business, it’s commoditised hygiene, better shared across all the vendors' customers than handcrafted bespoke for your implementation.

If we look at web applications, that extends to the ability to add pages, security, metadata, user logins, permissions, WYSIWYG editing, workflow, scale for traffic spikes, etc. etc.

Develop it yourself, and you'll wish someone else had done the adulting. In fact, in the early days of web application platforms, that was part of the vendor sales playbook.

“Good luck with rolling your own, we’ll be back in 6 months when this shit gets real”.

I’m not sure if that argument changes, whether the code is written by a developer, hacked by someone reliving their technical youth, outsourced to an exotic land of developers, no-coded, or vibed into existence from the big giant AI head: There's the application adulting.

The most common vibe examples I am seeing are some version of a basic CRM, but still, once they get over the wonder and delight of delivering a simple use case, a single-user system, some forms, a database, what happens next?

Now you need to add multiple users, with permissions, audit, data ownership, ability to set and change passwords, and…. all… that…. is vibing that easier and cheaper than tossing $20 a month at something out of the box?

The first afternoon of vibe coding might seem cheap, but the next three years might not be.

Now, of course, this whole argument assumes a level of application complexity. If your SaaS proposition is a simple workflow app with forms and a dashboard (or a really simple CRM), then your moat just went dry.

A creative business user with a basic grasp of app development will vibe the shit out of that.

We love to say stuff is dead, this week it’s SaaS, but spreadsheets didn’t kill accounting software, they just enabled business users to do some of the simpler use cases themselves.

Which brings me to the phrase “enable business users”, in my experience, do they want to be enabled? Really? They just want to get their shit done, not do someone else’s job, or run a cottage software company on the side.

Maybe I can trot out my old free puppies and free beer analogy I’ve used before. Vibe coding looks a lot like free beer, effortless and immediately rewarding until it turns into a Great Dane, needs feeding and training before it breaks stuff.

Anyhow, making AI predictions is a perilous game, but rather than AI killing SaaS, the real question is something like:

What kind of SaaS survives in a world where building is democratised and cheap?

And of course, those SaaS companies will be built by AI-augmented developers.

The question is: what value can they add with their AI-augmented dev team compared to the AI DIY tools offered to the tech-savvy business user?

I think it’s still a classic build-versus-buy argument; it's just that AI has lowered the barrier to entry for build, making it easier and cheaper than ever, changing the game.

But that applies to both sides of the argument.

Will the vibe kill SaaS as some are predicting?

Or will the reduced cost of development mean that a SaaS still wins, build versus buy?

Whatever it is, I don't think the vibe has killed SaaS quite yet.