Machines Don’t Care. Humans Do.

In a few years, everything will be vibe-coded. That’s the prophecy.

Yeah, sure.

For now, we just need to ride it out. It’s the age of slapping AI labels on everything, inflated job titles packed with AI-flavoured bullshit, and vibe-coding sold as the next revolution.

Managers are getting starry-eyed at the thought they no longer need those weirdos with beards and shorts, sipping coffee, avoiding eye contact, and smelling a bit odd.

The core problem with all these companies thinking they can replace developers with specs written by product folks, fed into an LLM to magically produce implementation plans, is that they’re still thinking in waterfall terms.

Surprised? Not really.

They still see engineers as the physical arm of product delivery, not as strategic minds who bring back insight from the technical frontier and help shape product direction with it.

Just like me, there are engineers in every company who care, who’ve been speaking up for years and gone unheard. People treated as measurable, replaceable “resources”, not recognised as valuable co-creators. We’re not short of wisdom or initiative. We’re buried under layers of hierarchy and control fantasies.

It’s arrogance, plain and simple. The arrogance of thinking the problem is already understood and the only missing piece is someone to execute the brilliant solution. The arrogance of ignoring feedback loops. As long as engineers are seen as instruction-followers rather than co-authors of intent, companies will save a lot of money and still fail in the same old ways.

I’ve got 30 years in this field. I’m not just a coder. I build software products. I know how to look at a messy problem and shape a solution that fits within the company’s constraints. I know how to structure work so that it serves both the business and the customer.

And I can do that because I care. An LLM doesn’t. An LLM can’t.

An LLM is a statistical model trained on oceans of text. When it gets a prompt, it doesn’t think or reason or feel. It calculates probabilities. It chops up the input into tokens, maps them into a high-dimensional vector space, and runs them through stacked layers of mathematical operations like self-attention, linear projections, and non-linear activations to guess what token should come next.

It doesn’t understand the task. It doesn’t grasp the user, the stakes, or the context. It can generate text that looks plausible, based on what it has seen before, but it doesn’t know what it’s doing. It doesn’t know why. It has no compass. No grasp of trade-offs. No sense of ownership. No ability to care if what it produces solves a real problem or creates a bigger one.

It’s not lazy. It’s not evil. It’s just indifferent. And that’s what not caring looks like.

“Everything will be vibe-coded”, yes, and every customer will be damaged.

Originally posted on LinkedIn.