Everyone's panicking about "AI features" being bolted onto products like it's 2010 and we're adding social login buttons. That's not the bubble. The bubble is the assumption that current software companies have defensibility.
Here's the thing: we're not adding AI to products. We're removing the need for most products entirely.
Nobody shipped without search after Google. But search was an enhancement—it made existing software better. AI is a solvent. It dissolves the economic moat that justified building the software in the first place.
YC's entire thesis rests on startups capturing value during the window between "this problem is painful" and "an incumbent solves it." But what happens when that window collapses to zero? When any reasonably clever person can get Claude or GPT to generate a bespoke solution to their specific problem in 20 minutes?
I'm watching food service managers—people who optimize labor, inventory, and customer flow in real-time—get told they can't build software. That's a lie we told ourselves to justify $150k engineer salaries and $10M seed rounds. Those managers have exactly the cognitive toolkit needed. They just didn't know C++.
In three years, they won't need to.
The SaaS model assumes friction. It assumes most people can't build the thing themselves, so they'll pay you $50/month forever. Coding AI doesn't make software easier to build—it makes the *act of building* indistinguishable from the *act of using*.
You don't need a project management tool with 600 features. You need to tell your computer what you're trying to coordinate. You don't need Photoshop's menus. You need to describe what you want the image to convey.
Every software company selling picks and shovels to the AI gold rush is missing that they're about to get disintermediated by the prospectors themselves. The cloud was never about technical superiority—it was about control and recurring revenue. What happens when capable models run locally and people can spin up exactly what they need?
VCs are investing in moats that evaporate the moment non-engineers realize they don't need us anymore. Network effects, switching costs, proprietary data—all predicated on software remaining expensive to create.
It's not.
The actual bubble is venture capital itself. You can't invest in defensibility that doesn't exist. And you can't charge rent on the gap between intention and execution when that gap is closing.
We're not in an AI bubble. We're watching the software bubble pop in slow motion.
If want your claiming is correct, it goes for nearly every profession.
Why do I need a lawyer if I can just get the AI to do it all for me? Filings, briefs, legal arguments etc. are all just output generated from specific inputs after all.
Why do I need to go to a doctor if I can just have AI diagnose, and eventually even run tests and then operate?
Why would I need artists / marketers for whatever my product, the magical AI can just do it all.
It could be we're headed down this road, but I don't see how software is somehow special in that its the only thing AI can do competently
The fun part is what looks like the core assumption here, that the current AI thing will actually work, doesn't even need to be true; it just needs enough people to believe it for the negative effects of it to become true.
Even without AI the software industry is not in a good state for a range of reasons. The big tech companies barely sell software - it's custom software operated to perform some other much stickier function, but the value is not in the software per se beyond it enabling selling the other thing.
This - unfortunately - goes for a lot of domains. Things that you used to be able to make a career out of now have a shelf life of a few years, sometimes less than that. This time compression is not just upsetting the VC world, it is upsetting just about everything in high tech societies.
I don’t know how you can claim all this when ai is nondeterministic and hallucinates. Menu based software is merely a way to trigger a function, the same function, every time you want to run it. Ai in its current incarnation cannot even do that.
> I don’t know how you can claim all this when ai is nondeterministic and hallucinates.
It's so bad. I was thinking that it would at least replace people whose job is largely to give their worthless opinions, but it won't even do that. Those people's real job is often to add head count, to give your uncle's kid a job or to be fired when things go wrong. AI can't do any of that, it just generates the worthless opinions. Now the idiot won't even have to imitate the verbiage they sort-of heard in college; they're the ones that are using AI the most, to bullshit for them.
It's good for helping you think through stuff you're thinking about by repeating back to you in different words (and getting even that wrong often, forcing you to clarify.) It's horrifically bad at finding or following references, reasoning, or coming up with anything new.
It's not eating software, it's barely even touching software, other than being shoveled into it randomly. The obvious proof that AI is bad is that there are actual geniuses who came up with the algorithms to speed up, parallelize and to bias in a way that makes them seem more productive and creative. LLMs don't seem to be helping amazing minds like that improve AI itself. If LLMs were even going to be fertile, that pairing would be a semi-singularity even if these exceptional humans couldn't be taken out of the loop. My bet is that they don't help at all.
Yeah, having to type out, or even speak "rotate this by 90 degrees about the Z axis" would get annoying fairly quick. Not to mention the inference is probably going to consume several seconds.
Blender user, or probably any other, would be able to pull that with a mouse click, followed by a few keystrokes. Bam, done.
Now something time consuming like UV-unrolling, sure go ahead and incorporate AI. But I'd bet it would need quite a bit of tweaking for a "pro" job, although of course thats not always necessary
Everyone's panicking about "AI features" being bolted onto products like it's 2010 and we're adding social login buttons. That's not the bubble. The bubble is the assumption that current software companies have defensibility.
Here's the thing: we're not adding AI to products. We're removing the need for most products entirely.
Nobody shipped without search after Google. But search was an enhancement—it made existing software better. AI is a solvent. It dissolves the economic moat that justified building the software in the first place.
YC's entire thesis rests on startups capturing value during the window between "this problem is painful" and "an incumbent solves it." But what happens when that window collapses to zero? When any reasonably clever person can get Claude or GPT to generate a bespoke solution to their specific problem in 20 minutes?
I'm watching food service managers—people who optimize labor, inventory, and customer flow in real-time—get told they can't build software. That's a lie we told ourselves to justify $150k engineer salaries and $10M seed rounds. Those managers have exactly the cognitive toolkit needed. They just didn't know C++.
In three years, they won't need to.
The SaaS model assumes friction. It assumes most people can't build the thing themselves, so they'll pay you $50/month forever. Coding AI doesn't make software easier to build—it makes the *act of building* indistinguishable from the *act of using*.
You don't need a project management tool with 600 features. You need to tell your computer what you're trying to coordinate. You don't need Photoshop's menus. You need to describe what you want the image to convey.
Every software company selling picks and shovels to the AI gold rush is missing that they're about to get disintermediated by the prospectors themselves. The cloud was never about technical superiority—it was about control and recurring revenue. What happens when capable models run locally and people can spin up exactly what they need?
VCs are investing in moats that evaporate the moment non-engineers realize they don't need us anymore. Network effects, switching costs, proprietary data—all predicated on software remaining expensive to create.
It's not.
The actual bubble is venture capital itself. You can't invest in defensibility that doesn't exist. And you can't charge rent on the gap between intention and execution when that gap is closing.
We're not in an AI bubble. We're watching the software bubble pop in slow motion.