I just quit my six-figure job to build a SaaS. Is SaaS dead, or am I?

Three weeks ago I handed in my notice at a six-figure job. No co-founder, no funding, no safety net beyond 4 months of savings. I did it to build one small SaaS product.

Since then I've had the same thought roughly every second day at 2am: this is the dumbest possible moment in history to do this.

Here's the fear, stated plainly, because I'd rather hear it argued with than keep circling it alone:

1. Every SaaS is now one feature announcement away from irrelevance. I don't mean "a competitor might launch." I mean an AI lab ships a model update, a checkbox appears in a product 400 million people already pay for, and my entire reason to exist becomes a bullet point in someone else's changelog. I didn't build something a big company could copy — I built something a big company could copy by accident, on a Tuesday, while shipping something else.

2. "SaaS is dead" is now a genre. Half my feed says software margins are gone and the future is agents doing the job instead of tools helping you do it. The other half says there's never been a better time to be one person with an idea. Both camps are extremely confident. Neither of them has to make my rent.

3. The salary was real. The revenue is not, yet. I gave up a known, boring, actually-quite-good number for an unknown one. Everyone who's done this says the same thing afterward — "best decision I ever made" — but I've noticed we never hear from the ones for whom it wasn't.

For context, what I actually left to build: Anyratio — you upload one approved campaign photo and get it back in every size your channels need. It's not a prompt toy and it doesn't generate new creative; it extends the canvas around the photo you already signed off on, so the subject never gets cropped in half for a 9:16 story. One image in, a full size kit out, in under a minute. Marketing teams currently do this by hand, one crop at a time.

Small, unglamorous, deeply boring problem. That's the thesis: it's too boring for a frontier lab to care about and too specific for a generic image model to get right. That's either genuine defensibility or a very comfortable story I tell myself. I honestly cannot tell which anymore, and that's the real reason I'm posting.

So, three questions:

• Have you done this? Quit something good to build something uncertain. What did the first six months actually feel like, and what do you wish you'd known before the notice period ended?

• How do you think about "one feature away"? Is niche depth real protection, or just a slower death? I want to hear from people who've been steamrolled as much as from people who weren't.

• Where do you all hang out? I'm looking for a community of people actually shipping and selling small SaaS — not a Discord of 12,000 people posting screenshots of their $9 MRR. Where's the good room?

Roast the idea, roast the decision, or tell me it works out. All three are useful right now.