Making niche solutions is the point
Posted by evakhoury 2 days ago
Comments
Comment by jmathai 41 minutes ago
I attempted to productize a super niche idea I had for myself that I have ended up using multiple times. It's a simpler method of running wires without having to cut/patch drywall and drill through studs.
It started with trying to create a router-like device equipped with a motor (which was version 1 [1]).
But then iterated it to be a plate which can be attached to existing routers. That resulted in a drastically improved user experience which went from doing it all inside and making a mess to doing it in your garage with more precision. I've used the final design on 4+ projects at home and it has saved me so much time [2].
[1] https://www.dropbox.com/scl/fi/i5p330nqo0cfrfx3gxf05/Wiresha...
[2] https://trywireshark.com (includes diagrams and videos of what I 3d printed)
Comment by arjie 2 hours ago
But my wife used some base open source components to design a block that goes into the bits of a playpen that we used to have and transform it into something that docks with the wall instead of only with itself[0].
And I have designed with Claude a few small things like card holders for the board game power grid[1].
I wish there were better AI tools for interacting with modeling software. As it stands I use OpenSCAD with Claude and that seems as good as it can be. There are Solidworks AI startups but they’re like for professionals.
The Bambu P1S I have is quite low friction to set up. And I have an AMS2 Pro on top of it that feeds different kinds of filament (material and color) into the printer. I have just the one but now I wish I had more AMS hooked up.
0: https://wiki.roshangeorge.dev/w/Blog/2026-01-15/Modeling_Wit...
1: https://wiki.roshangeorge.dev/w/Blog/2026-01-11/Modeling_Wit...
Comment by ramboldio 3 hours ago
Unfortunately, it is still very hard to _design_ niche solutions. The usability of CAD tools did not really improve at all in the last 20 years..
Comment by Arcanum-XIII 2 hours ago
CAD is the easy part.
Comment by wat10000 1 hour ago
There's an engineering saying that anybody can design a bridge that won't fall down, but it takes an engineer to design a bridge that just barely won't fall down. Why do you want a bridge that just barely won't fall down? Because it's a lot cheaper to build. That's not much of a concern when you're printing little doodads at home. I waste some material by designing overly-strong structures, or getting it wrong and iterating. That's fine, the stuff is cheap.
Comment by barumrho 3 hours ago
Comment by direwolf20 4 hours ago
Comment by Bayart 3 hours ago
In truth every time an issue fit for 3D printing has come up in my life, I solved it easily with wood and cardboard. I'm starting to recognize I might be a craftsman at heart.
Comment by CyLith 1 hour ago
Comment by IncreasePosts 2 hours ago
Comment by CountHackulus 1 hour ago
Comment by ge96 3 hours ago
Comment by ge96 1 hour ago
Comment by nozzlegear 2 hours ago
Comment by jagged-chisel 1 hour ago
Comment by gurjeet 3 hours ago
From TFA:
> 1. I like to think that all printers are 3D, unless it's a printer in Flatland.
Comment by direwolf20 2 hours ago
Comment by pbronez 2 hours ago
Comment by uoaei 2 hours ago
Now the hype has seemed to shift to "do absolutely anything just barely well enough to get people to pay for it".
Comment by pixl97 2 hours ago
Comment by esseph 1 hour ago
"Unix Philosophy", which many of us wouldn't be here if it didn't exist, wasn't designed with any sort of money in mind.
That's like complaining that the company that picks up your residential trash is a shit company for not reducing your travel time to work.
Comment by 01HNNWZ0MV43FF 37 minutes ago
Proprietary / for-profit software is made that way, yes. There's no other way you can do it, given the incentive landscape. You take the huge base of FOSS code and spend as little effort as possible to build a thin little layer of moss on top. You respect the customers as little as possible to keep them around. You fix bugs as little as possible. You use other proprietary services that spy on your customers, like Sentry and Firebase, because privacy costs money.
Free / libre software has to compete on its merits. If a project isn't useful it doesn't grow. But not growing is okay. Some projects accrete over years, like a pile of stones forming a cairn. They don't need to squeeze money out of people to live because they aren't alive. They don't have to "eat".
I'm mixed on the Unix philosophy. It makes a lot of sense when you're building CLI tools that a hacker is going to plug together, because then your program is really a function in a programming environment that spans one or more entire computers. Tools like ripgrep, jq, curl, they're all great, I love them. A good function does one thing and does it well.
But just as often, I'm okay with huge software that does a lot. Web browser engines are evolving into universal GUI / IO frameworks and I'm trying to make peace with that. Systemd does a ton of stuff, but hell, so does the Linux kernel. I don't see an inherent problem with having an init system that acts like a monolith kernel for userspace. Microkernels are nifty but in the end we all ship our org charts. Maybe there's no need for microkernels if the real divide is between kernel programming and userspace programming. For the same reason, I haven't found any personal use for wasm, because wasm makes the most sense when you're connecting two pieces of code written by different teams at different times, like a GIMP plugin. I don't need wasm to plug my own code into my own code.
And for GUIs it's just been a fucking nightmare. 57 years since the Mother Of All Demos and it's still 10x easier to write `fn main()` and build a CLI program that runs on _every_ OS, rather than a GUI program that maybe runs on one OS, like it runs on Ubuntu 24.04 but not Ubuntu 22.04. What a fucking mess. GUIs don't compose, so every GUI project I've tried feels like I'm inventing the universe from scratch. It's fun but it's a stupid fucking waste of time.
Honestly browser engine frameworks like Electron and Tauri _are_ the Unix philosophy compromise. Making a GUI framework requires tens of thousands of lines of high-effort code made by experts. If a browser's one thing is "Be a GUI framework" then it allows your GUI app to just serve HTML and now it's a CLI app that runs anywhere without fucking with GTK 3 vs GTK 4 bullshit.
Sorry for the long rant. This stuff is all percolating in my head. I started with Visual Basic 6 and I still haven't seen GUIs improve since then. Phones have been a fucking step backwards, too. Everyone uses a phone but just like a mouse-and-keyboard player dominating gamepad players in a shooter game, I get a lot more done at a real desktop with a real pointing device.