zlacker

[return to "AI agents are starting to eat SaaS"]
1. benzib+xN[view] [source] 2025-12-15 07:45:57
>>jnord+(OP)
I'm CTO at a vertical SaaS company, paired with a product-focused CEO with deep domain expertise. The thesis doesn't match my experience.

For one thing, the threat model assumes customers can build their own tools. Our end users can't. Their current "system" is Excel. The big enterprises that employ them have thousands of devs, but two of them explicitly cloned our product and tried to poach their own users onto it. One gave up. The other's users tell us it's crap. We've lost zero paying subscribers to free internal alternatives.

I believe that agents are a multiplier on existing velocity, not an equalizer. We use agents heavily and ship faster than ever. We get a lot of feedback from users as to what the internal tech teams are shipping and based on this there's little evidence of any increase in velocity from them.

The bottleneck is still knowing what to build, not building. A lot of the value in our product is in decisions users don't even know we made for them. Domain expertise + tight feedback loop with users can't be replicated by an internal developer in an afternoon.

◧◩
2. mikert+Vp1[view] [source] 2025-12-15 12:56:24
>>benzib+xN
A year or two from now it will be trivial to copy your product
◧◩◪
3. Bridge+gL1[view] [source] 2025-12-15 14:47:26
>>mikert+Vp1
Yes friend. That's what people said 2 years ago. Next?
◧◩◪◨
4. latent+kR1[view] [source] 2025-12-15 15:18:59
>>Bridge+gL1
There's been an obvious step change on the coding front from 2 years ago, and it feels obvious to me there's going to be another. The difference now is the people working on systems to clone SaaS at scale are likely starting to put real effort, sustained effort into now that agents are good enough to accomplish subsets of it, can be improved much further with the right techniques and orchestration, and themselves will get better over the next two years along with all the improvements and build up of tooling. Right now feels like one of those "skate to where the puck is going to be" moments in time.
◧◩◪◨⬒
5. Willia+rA4[view] [source] 2025-12-16 08:21:36
>>latent+kR1
Personally, hard disagree. There were improvements, yes, but the last year felt particularly stagnant to me
◧◩◪◨⬒⬓
6. latent+Rc5[view] [source] 2025-12-16 13:51:04
>>Willia+rA4
I've found assigning issues to GitHub Copilot on GitHub itself to be a real step-wise change from one off requests to a chat based interface that has no awareness of my codebase. Maybe it's just me, but I'm getting significantly more realworld value out of AI these days. I've been working through doing an implementation of a spec using GitHub Copilot mostly from my phone, and it has been a really instructive exercise in how to squeeze as much out such a narrow interface as possible. I still dip into a full desktop + IDE from time to time, but for like 90% of it I've been doing it in my phone. Slowly but surely working my way from "vibe coded in a weekend" level initial AI slop quality to nice, clean architecture, excellent dev tooling, production grade multitenant SaaS. I've still got a ways to go, but I'm able to make quite a lot of progress just using my phone during my commute, on my lunch breaks, in between meetings, and bathroom breaks etc.
[go to top]