Man I love Obama
Posts by solderlocks
Anyone out there using Gemma for local coding (via tools like Cline, Roo Code, etc)? Curious about how it compares to Flash 3.1 for coding tasks
Don't get me wrong, it'll be a major problem if these systems are as insecure as IOT devices are today. But local exploits are a smaller scale problem than potential mass enshitiffication of technology inside large numbers of people. What are your thoughts?
I'm more concerned about the cloud side (data control, privacy, device configuration, etc) than I am about local security. The thought of a corporation controlling what I can do with hardware implanted into my own body is disturbing.
You're absolutely right. With implantable technology and neural interfaces, user sovereignty is going to be significantly more important. What's the best path forward? Open source software and hardware?
What would software look like if we minimized engagement instead of maximized it?
What happened to live coding? Are there still people out there coding as a performance art?
I missed old Doodle polls, so I built a replacement.
Instead of asking when you're free, it asks when you're not. Conflicts get marked. The best time reveals itself in the empty space.
Free, open-source, no accounts ever: timeslot.ink
There should be app races where people compete to see who can finish a complex app fastest and best using whatever tools they have at their disposal
What's preventing the creation of a direct-manipulation development tool for OpenSCAD? I'm imagining something that breaks out the AST of the transformations like this model: scadder.dev?file=https%3...
We need more intuitive UI for open parametric CAD software
Some funny insights from the Claude Code leak:
- They detect user frustration with regexes that match strings like "piece of crap", "wtf", "f***ing broken"
- The system hides that Claude wrote the code when submitting code to open source repositories
Always double check your .gitignore
Sad to see that Firebase Studio is shutting down, but I understand the logic. Google AI Studio sounds like an effective alternative for browser-based development. Curious to hear how they compare from someone who's used both.
Would people use OpenSCAD more often if the tools for creating .scads were better? Pondering open source hardware and the barriers preventing it from being more like open source software (with vetted libraries of modular components, alternatives to paid products, etc).
#OpenSCAD #Opensource #3D
Just released my first npm package for #OpenSCAD and it feels like a milestone! Run "npx scadder install [url-or-model-id]" to pull the respective scad (like the gridfinity-rebuilt-baseplate) from GitHub onto your local machine.
#opensource #opensourcehardware #3dprinting #3dmodeling #npm #github
Just shipped Scadder: a serverless web viewer & CLI package manager for OpenSCAD.
Compile .scad files in-browser via WASM (shoutout openscad.cloud) or use the CLI to pull dependencies locally. No walled gardens or accounts. 100% open source.
github.com/solderlocks/...
Roo Code + Openrouter > Antigravity
Google Antigravity is a great tool, but their sudden model quota change just squandered their goodwill with the developer community and pushed most of their users to look for viable locally hosted alternatives.
Moving from writing code, to designing a series of agents that can reliably write code, to writing a spec that others can use to contribute to an ecosystem of agentic nodes/solvers is dizzying. It feels like "integration" over the problem space when you zoom out to higher level abstractions
What an amazing human being. He knew what it was all about
What would it take to implement a .f3d -> .scad pipeline? Could we analyze the parametric history of an object and find equivalent CSG transformations for each step? Maybe with fuzzing and AI semantic mapping we can actually interpret design intent and preserve the original parameter mappings.
Do you think we'll have AI tools soon that can understand design intent from parametric models and translate them across formats? Been thinking about a .f3d -> .scad pipeline and realizing that it might actually be doable (probably with a bit of human guidance/input)
@genesismachines.bsky.social Do you think that AI modeling systems will eliminate the need for tools that make it easy for humans to manually adjust designs (IE interfaces for scad files, f3d, etc)?
Yeah that's why I have been thinking about the SCAD community lately. STLs are like compiled binaries (fixed and relatively immutable), and SCAD is like source code that makes it easy to open a specialized solution and add a salad fork. It's tragic that people have to start from scratch so often.
Systems like James Bruton's robotics platform and Lego Mindstorms show the value of out of the box components that can be customized for different use cases. Modular systems are a godsend for rapid prototyping
Yeah it makes sense that more specialized components would perform better. It feels like there's a tendency to overcorrect in the direction of bespoke hardware though, when often a problem has been solved thousands of times in the past already by other teams and just a small tweak is needed.
I've been wondering why the Reprap community didn't end up adopting a culture of using modular parametric components (like hinges, threads, clips, etc) and building robust libraries of "open objects" that people could build on and improve. Was it because scad files were too cumbersome to view/make?
@oshwassociation.bsky.social could OpenScad files hosted on GitHub serve as the foundation for an open-source hardware community where contributors borrow components from one another and build on each other's successes? Tired of using STLs and other "dead formats" when designing printable objects
What games excel at enabling users to discover/invent their own emergent gameplay? What are some good ways to design for this?
@seanmcarroll.bsky.social What would morality look like in a universe where the arrow of time were flipped? Does that imply that morality is inherently tied to entropy? Pondering Jeremy England's work and its connection to our intuition about "ought"