Extinction Rebellion: The Fall Of The Dev

There is a specific species of developer having a very public breakdown in 2026, and if you have spent five minutes in any AI-builder community you have already met him. He lurks in the Lovable subreddit. He patrols the Bolt Discord. He appears, unbidden, under every LinkedIn post where a non-technical founder posts a screenshot of an app they shipped over a weekend. His comment is always some variation of the same three sentences. 'This is a security nightmare.' 'Wait until you try to scale.' 'You have no idea what you have just built.'

He is not wrong about everything. He is wrong about almost everything. And the reason he keeps posting is not because he cares about your database. It is because his industry is being unbundled in public and he can feel it in his bones.

Meet the dinosaur

The dinosaur dev is not every developer. Good engineers are quietly using these tools, shipping faster, and keeping their mouths shut because their day rate does not depend on gatekeeping. The dinosaur is the specific subspecies whose entire professional identity is built on the difficulty of the craft. He learned React the hard way. He memorised the CSS box model. He has opinions about tabs. He gets paid, in large part, because the barrier to entry used to be enormous and he cleared it.

That barrier is being demolished by tools that let a florist in Denver ship a booking app in an afternoon. The dinosaur has two choices: adapt, or convince the florist she has just built a lawsuit. He is picking option two, and he is picking it loudly.

The greatest hits of the FUD playbook

The comments repeat because the anxiety is the same. Here are the reliable classics, and what is actually true about each.

'Your database is completely exposed.' What he means: you are using a Postgres instance behind a public API. What is actually true: so is every serious app on the internet, including the one his own employer runs. Supabase, the backend most of these platforms ship with, enforces row-level security at the database layer. If RLS is configured — and the templates increasingly force it — the exposure surface is smaller than the last WordPress install he shipped, which was running six unpatched plugins and an admin panel at /wp-admin. The correct answer is 'turn on RLS and write policies,' not 'delete your app.'

'You have no idea what you are shipping.' A curious accusation from a professional who ships npm modules he has never read. Every developer, at every level, ships code they did not fully write. The stack is turtles all the way down. The question was never 'do you understand every line' — it was 'does it do the thing, and can you fix it when it breaks.' AI-native founders can now answer both, which is what really stings.

'Wait until you have to scale.' The unspoken assumption is that the app will succeed enough to need scaling, which is already a better outcome than the dinosaur's last four consulting engagements. Ninety-five percent of apps never need to scale beyond a single modest server. The five percent that do will hire someone — possibly even him — at that point, with revenue. Optimising for a scaling problem you do not have is called 'never shipping,' and it is the disease he is trying to sell as a cure.

'It will cost you a fortune in tokens.' Said by a man charging four hundred pounds a day. The maths on this is embarrassing for him. A typical solo founder on Lovable spends between forty and a hundred pounds a month on tokens to ship features. His invoice for the same work starts at four figures and ends whenever the retainer runs out.

'GDPR will destroy you.' GDPR applies to how you handle personal data. It does not care whether the code was typed by a human, an AI, or a trained parrot. Every framework in existence is equally capable of violating GDPR. The dinosaur invoking it as an AI-specific risk is either lying or has not read the regulation, and both are disqualifying.

Why he does it

It is easy to assume malice. It is more accurate to assume terror. The dinosaur is watching his profession's pricing floor collapse in real time. The junior React developer whose salary was propped up by artificial scarcity is now competing with a marketing manager who shipped a working MVP on a Sunday. The freelancer whose retainer covered 'small fixes' is being replaced by a founder typing 'fix the header spacing' into a chat box. The agency that charged twenty thousand for a brochure site is quoting against a solo builder who did it in an afternoon for the cost of a takeaway curry.

None of that is the fault of the non-technical founder shipping her booking app. But she is the visible face of the thing that is eating his career, and the internet lets him yell at her for free. So he yells.

The specific tells

• →Vague authority claims: 'I have been doing this for fifteen years.' Cool. So has the WordPress ecosystem, and look how that turned out.
• →Escalating hypotheticals: starts with 'what if you get one user' and ends with 'what if you get ten million users on Black Friday during a DDOS while GDPR is being amended.'
• →The unrequested audit: he did not read your code. He read your screenshot. He is now telling you your architecture is wrong.
• →The Dunning-Kruger inversion: assumes the AI cannot possibly have done a competent job, because if it had, what has he been charging for.
• →The gatekeeper's sigh: 'this is why we cannot let non-developers ship software.' Said by a man whose last project was three months late.

What the good engineers are doing

The developers who are not panicking are doing three things. They are using these tools themselves, quietly, to ten-x their own output. They are moving up the stack — from 'writing the code' to 'reviewing the code, hardening the security, and shipping the thing.' And they are picking up work from every founder whose weekend build outgrew its britches and needs an actual professional to take it the next mile. That is a bigger market than the one they had before. The dinosaur cannot see it because he is too busy screaming at it.

The real risks, briefly, because they exist

In fairness — the dinosaur is loudest about the wrong risks because the real ones are less dramatic. If you are shipping with an AI builder in 2026, the actual things to check are: RLS policies on every table with user data; a proper auth flow with email verification, not a hardcoded admin flag on a profile; sensible cost limits on any AI feature you expose to the public before someone spins up a bot; a backup and restore plan you have actually tested; and a rate limit on anything that touches money or email. None of this requires a computer science degree. All of it is in the docs of every platform mentioned in this piece. The dinosaur will not tell you any of it because it undermines his position that the whole endeavour is impossible.

The prediction

The dinosaur dev has about eighteen months of runway on the FUD strategy. After that the market catches up. The florist's booking app will have been running for a year without incident. The marketing manager's SaaS will have paying customers. The receipts will be public. At that point the FUD stops landing, and the dinosaur has to either adapt — pick up the tools, move up the stack, learn to supervise instead of type — or watch his rate compress until it matches the value he actually adds, which by then will not be much.

Extinction is rarely dramatic. It is a slow salary decline, a shrinking pool of clients who still want it done the old way, a growing gap between his invoice and the market rate, and eventually a career pivot that gets described in his LinkedIn as 'strategic advisor.' The meteor has already hit. He is standing in the impact crater complaining about the taste of the dust.

“The barrier to entry was never the point. Shipping was the point. He forgot.”

If you are the founder being yelled at

Read the comment. Extract any specific technical claim. Check it against your platform's docs. If it is a real issue, fix it — it takes an hour and now your app is better. If it is a hypothetical about scale you do not have, or a vague sneer about 'real engineering,' ignore it and get back to shipping. The people building the future do not have time to argue with the people mourning the past.