My laptop’s running a little slow these days. It’s got a handful of personal dev servers running multiple tools, 90% of the code written by AI. Are they full-featured? Not even close. Maintainable? Barely. Do they need to be? Not at all.
They’ve already saved me about $500/year in SaaS costs. These are just a few of what I’ve built so far:
- Automatic database backups to Dropbox (previous service cancelled)
- Strava activity archive with metrics and time-based queries (Pro, cancelled)
- Conversation and meeting transcript backups I can actually search and ask questions about
- A 5-step AI prompt builder trained on the latest Google and OpenAI prompting guides (the circular nature of this is truly amusing)
- A janky Clay[dot]com knockoff
None of this runs in the cloud.
There’s no public deployments, no fancy infrastructure.
Am I going back to “on prem” software? It looks like it, I’ve even been considering buying a Mac Mini just to run it all locally and get it off my laptop. Sounds crazy, but it’s kind of freeing.
These tools have no auth, no users, no bloat—and that’s the point. They do exactly what I need, strip out what I never used, and give me back control.
What’s this mean for SaaS?
It’s probably too soon to tell, but maybe it’s time to admit not everyone wants a bloated all-in-one solution with 42 features and a $49/month price tag. I don’t care about your roadmap, I’ve got AI and a weekend.
Perhaps the future is personalized, customizable UI, API-first, where users describe what they need and the UI adapts to them.
That’s a design challenge I'm looking forward to tackling!