I'm two minds on this one, and I'm not going to pretend I'm not.
On one side: we should be making it dead easy for people to solve their own problems, deterministically, which means apps. Apps invoked like incantations into existence. Shipped to some org slop marketplace. Providing slices of value. That's a genuinely good world — software stops being a thing you buy and wait for, and becomes a thing you conjure for the specific, boring problem in front of you. More of that, please.
On the other side: the vibe noise is palpable, and it's making genuine gold hard to surface. Everyone on X and LinkedIn is a dev now. Apparently they shipped Spotify in a weekend. (They shipped a Spotify-shaped screenshot in a weekend. That's most of the difference.) When the cost of producing a thing falls to zero, the volume goes vertical and the gold gets buried under the slop. That's the tax on abundance, and the tax is signal.
There's an interesting overlap here with the "SaaS is dead" debate, which is a partly-true, partly-false narrative. Some products are easy to replace, bad, or stupid. Some of them just aren't fit for purpose in an AI age. Others are genuinely great and add real value. Some SaaS is dead — and it should have been dead without AI; the AI just pulled the curtain back and showed everyone. The rest got a turbo. "SaaS is dead" is a slop headline. The truth doesn't fit on a slide.
But here's the part I keep turning over, because I think it's the real shift and almost nobody's pricing it right.
The tables have turned. Ideas are cheap, execution is hard — execution is the moat was the old mantra. Drilled into every founder for twenty years.
It's inverted. Ideas are incredibly valuable now. The cost of execution has dropped to near zero. But the quality of execution is still a premium — as is shipping it, maintaining it, and continually adding value (new ideas) to it. Cheap execution is free. Good execution is rare, and staying good is rarer. The moat didn't disappear. It moved — from "can you build it at all" to "can you build it well, ship it, keep it standing, and keep making it better." Good SaaS didn't die. Good SaaS got the opportunity to be good faster.
And underneath all of it is the thing I come back to with almost everything: AI is only an accelerant.
It will let you build more broken, useless crap than you ever dared dream possible — or weave more magic in less time than you ever hoped or dreamed. The outcome is faster, but you were always going to get there. Good engineers and bad engineers alike just reach their inevitable outcome at a speed that wasn't possible before. The flood of slop and the flood of gold aren't two phenomena. They're one: everyone arriving where they were already headed, faster. The slop was always going to be slop. Now it's just louder, and there's more of it, sooner. Same for the gold.
So where do I actually land?
Right here: I am pro org slop. I am anti attention slop. And I am pro vibes all day long — right up until the moment the end result isn't working software.
Let me split the word "slop" in half, because it's doing two completely different jobs and people keep smudging them together.
Org slop is earnest vibe builders stealing back time from their enterprise and their work, extracting value from a tool that's just handed them something close to a superpower. It's everyday people using technology to be lazier — or faster, or better; same same but different. Productive but flawed. I am pro this all day. This is the velocity gain reaching the people who never got to have it before, and it is one of the best things happening in software right now.
Noise slop is the other thing entirely. Lazy posting wearing the costume of the earnest builder — the same "building in public" language — shipping broken toys to localhost or a UI that doesn't work and declaring "job done" for the clicks, the likes, the attention. Clicks and likes and attention are not an end game. They're noise, dumped into the only spaces genuine builders have to talk about the actual thing they built. AI gave a single person the power to act like a full dev team and build genuinely great things — and you will never be able to tell those things apart from the polished turds and the skid marks left by attention-stealing noise merchants proudly declaring "I built X in a weekend."
Until recently I assumed everyone was shipping. X and LinkedIn certainly sell that perception. Then the informal surveys started — "who's actually shipping versus who's just on localhost posting for the like?" — and I've now read dozens of them in a few weeks, because they aren't hard to find; it's a whole trend. Some have thousands of replies, an avalanche of people fessing up to vibe slop they were never going to release. Then you click their profile and you'd swear they were on the hunt, curing cancer with AI, building the next big thing. I genuinely don't know what the end game is for that user. But I know the result: noise. Friction for the people actually building, fatigue for an audience that's already overwhelmed. Try launching a real thing into that. It's impossible.
So here's my modest proposal. We're going to start labelling AI images — fine. Let's label this too. Three categories:
- Org slop — productive but flawed. Built to get something done, ships rough, earns its keep anyway.
- Noise slop — was never going to be a thing and never will be. The output of an attention-seeking roach.
- Software — vibed or otherwise, but built with actual intent to solve a problem.
Label honestly and the river clears. The org slop gets to be proudly, usefully rough. The software gets found. And the noise merchants have to wear the badge that says what they actually made, which is nothing.
Because here's the part I won't be coy about: this is the engineers' moment, and it's unashamedly theirs. For the first time the people who can actually build are positioned to sit at the top of the food chain and eat the rest of the business — not get eaten by it. But that only happens if they take the gift seriously and move up the difficulty curve, not down it. The play isn't to ship a thousand localhost toys for likes. It's to transition to the harder problems — to become product engineers who own the whole arc from "what's worth building" to "it's live, it works, and it keeps getting better." That's the moat now. That's the throne.
Ship dangerously. The danger was never that you'd build the wrong thing — you'll build it in an afternoon and bin it, no harm done. The danger is drowning the genuine article in your own noise and losing the ability to tell which of it was ever any good.
There's still gold down there. There's just a lot more dirt now. Label it, mine it, and go eat the business.