May 18, 2026
Ctrl+Alt+Delusion
Where Are the Vibecoded Photoshops?
If AI made making apps easy, why isn’t there a fake Photoshop by now?
TLDR: The article argues that AI may make typing code easier, but it still hasn’t produced convincing replacements for huge apps like Photoshop, which suggests the hardest work was never the typing. In the comments, readers split between “exactly, the hype is overblown” and “calm down, it’s just not ready yet,” with plenty of jokes in between.
The post "Where are the vibecoded Photoshops?" basically throws a glitter bomb into one of tech’s loudest arguments: if artificial intelligence really made building big software easy for everyone, then where are the finished masterpieces? Not toy projects, not broken demos, not what the author calls “slop” — but full, coherent replacements for giants like Photoshop, Excel, or even an operating system. The writer’s answer is blunt: they’re not here, because the hardest part was never just typing code. It was deciding what to build, testing it, fixing it, and making the whole thing hold together in the real world.
And the comments? Oh, they smelled blood immediately. One camp nodded along hard. User swiftcoder agreed that cheaper code-writing is not the same thing as solving software or product creation, which gave the thread a serious “calm adult enters the room” vibe. Then came the meme energy: fugaziboutit joked that the real way to measure AI progress is how fast people “move the goalposts,” which is basically the internet’s favorite accusation in one neat little grenade. On the more bullish side, roenxi pushed back by saying, hold on, maybe the fake Photoshop era just isn’t due yet — give it until 2027 to 2030, maybe longer. And then there was pure chaos: darkwi11ow dropped a localhost link like a magician shouting “behold!” while only people in their own house could see the trick. Meanwhile PunchTornado turned the whole debate into a stock-market punchline: if software is still hard, maybe software companies are still a great bet. In other words, the hottest reaction wasn’t “AI wins” or “AI fails” — it was everyone accusing everyone else of exaggerating.
Key Points
- •The article argues that there is little visible evidence of AI "vibecoding" producing complex, coherent software comparable to Photoshop, Excel, Maya, Blender, databases, compilers, or operating systems.
- •It distinguishes between low-quality AI-generated output and the harder task of building non-trivial systems that require architectural judgment.
- •The post claims many accusations about AI-made work being trivial are presented without verification or clear tests.
- •It defines three levels of software work: typing syntax, verifying through testing and harnesses, and making architecture and product decisions.
- •The article concludes that AI has lowered the cost of code typing but has not materially changed the harder work of verification and system-level decision-making.