Ask HN: What Are You Working On? (July 2026)

[flagged]

@rambleraptor, I haven't actually tried to do anything with your solution. It looks like you've put a good amount of work into it, enough that it works for you and you're willing to share. That's more than I can say about my similar ideas. So don't take the following as anything negative. It's actually more a self-critique than anything else...

Ok, so, how many of us see something like this, and instantly just want to do it ourselves in our own way? I've seen at least 5 different versions of a "self-hosting platform" make their way on HN, sometimes more than once, but every time I really look at them, they just don't fit quite right, and often end up abandoned a while later. I've tried to make my own, and never have gotten anywhere worth sharing, and the couple times I've tried what others have made, I haven't been able to do what I want.

I mean the idea is really good. Make it easy to host apps for yourself and your family. But it seems like we're all more interested in building the platform than actually picking one, or two, platforms and turning them into something sustainable...

Anyway, what would it take for you to actually use something like this that someone else built?

Or am I just dealing with a bad case of not-invented-here?

I was making a project for more than last 7 months to solve and issue that i had while i was applying for the GSOC that when i went for this huge codebase it was so overwhelming i took that problem tried solving it by myself , i solved it you can view it on github- https://github.com/RajX-dev/N3MO its solves the problem by solving by providing the graph of that particular function you searched

You may (or may not) find Paul McKenney's free book on parallel programming interesting - it has an unusual tone for a Computer Science textbook that made it less dry for me. There's a substantial section covering lock free structures.

https://mirrors.edge.kernel.org/pub/linux/kernel/people/paul...

Outside of my day job, I've been building a few small SaaS tools. The biggest change AI has brought for me isn't that it does the work for me—it has dramatically reduced the implementation cost. The hard part is still figuring out what to build.

Most of these projects come directly from problems I run into.

One example is cloud cost management. There are plenty of dashboards that show where your cloud bill goes, but I haven't found one that actively points out the small, easy-to-miss services that quietly charge you month after month. Each charge is tiny, so nobody notices, but over a long period they add up.

Since I work in the data infrastructure space, I also built a tool that helps people install and configure databases. Setting up a database still involves a surprising amount of manual work, and I wanted something that made it much more straightforward.

Another project comes from my appreciation for GitHub Actions. I like the workflow, but much of our code isn't hosted on GitHub, and our deployment environments are all different. So I started building an offline CI/CD system. It can package code, copy artifacts, deploy them, run validation, and execute AI-generated deployment scripts. As long as your code lives in a Git repository and the target machine is reachable over the network, it works regardless of where the Git server is hosted.

All of these tools run on top of a backend platform I call *MotherBoard*, which provides the common pieces: authentication and SSO, metering and billing, products, subscription plans, and other shared services.

I've noticed that almost everything I build starts with a problem I personally encounter. In the past, I'd think, "That would be nice to have," and stop there because building it would take too much time. Now the implementation cost is low enough that I can actually build it. And if it turns out other people have the same problem, maybe it'll make a little money too.