cloudking

Benchmarks are one thing, but I've yet to see anything substantial actually built with local models. Most of the posts on X are about benchmarks and tok/s. Has anyone actually built a quality production app or game with real users?

How do you create repeatable tests in a non-deterministic system? Every time you send the same prompt you get a different answer.

Those simple prompts produce nearly the exact same layout in the 2 different models?

Pretty much doing everything that takes more time doing manually on computers

It's painfully obvious these decisions were not made by people who play Magic

They are also beholden to enterprise pricing and can't use the subsidized consumer max plans.

Definitely worth following for ideas.

How did you solve memory in your system?

I ended up using QMD + custom skills to keep it up to date as new data comes in. Essentially a cross-referenced set of markdown files forming a knowledge base: https://github.com/tobi/qmd

Tried it for a few weeks, was really buggy and unpredictable.

Kudos for the concept though, I ended up rolling my own agentic system with Claude Code from scratch that works much more reliably for my use cases.

You should open the issue with your IT department and see if they can optimize the performance. You can gather proof that GoGuardian is causing slow downs by looking at it's memory usage in the Task Manager (right click empty area in Chrome tab bar).

I'd be happy to meet and discuss. I will reach out.

The knowledge graph is well done. I think what's missing from all coworking apps is the UX.

Prompting is a very specialized skill, average users just don't know what to ask for to get the most out of the LLMs.

Ideally the UX should organize and surface information to the user that is important automatically, without needing to be prompted.

/sandbox https://code.claude.com/docs/en/sandboxing

We use https://devin.ai for this and it works very well. Devin has it's own virtual environment, IDE, terminal and browser. You can configure it to run your application and connect to whatever it needs. Devin can modify the app, test changes in the browser and send you a screen recording of the working feature with a PR.

Why is OpenCode better than Claude Code?

Nice solution, I had some success with https://www.task-master.dev/ it's a bit more complex though

For connecting agents to third-party systems I prefer CLI tools, less context bloat and faster. You can define the CLI usage in your agent instructions. If the MCP you're using doesn't exist as a CLI, build one with your agent.

Well kind of. They use the same data center infrastructure, with internal versions of the external GCP services.

I'm fairly new to Docker. Do folks really rely on public images and registries for production systems? Seems like a brittle strategy.

Built something similar and it worked quite well. Gemini Flash models are very performant and low cost for these types of solutions.

For me it's about the least amount of hand/wrist movement. Trackpad + Swish feels natural and reduces strain.

Pretty cool, but nothing beats Swish if you're using a Macbook trackpad https://highlyopinionated.co/swish/

I took it for a spin spin to build a simple task manager. It worked for 45 minutes, built half the solution and then asked me for money. Compared to the other options in this space, it seems like an expensive solution.

Another simple solution works on LAN https://pairdrop.net/

Most people in the world only have a phone, no computer. Their window to software is the App Store or Play Store.

Built entire businesses on Google Sheets and Apps Script. Powerful combo.

What languages do you use?

- HTML, JavaScript, Python, PHP, Rust

What kind of projects?

- Web apps (consumer and enterprise), games

Do you maintain these projects or is this for greenfield development?

- Both, I have my own projects and client projects

Could you fix any bugs without Claude?

- Yes, I have decades of software development experience

Are these projects tested, who writes the tests. If it's Claude how do you know these tests actually test something sensible?

- For serious projects yes, I will define the test cases and have Claude build them out along with any additional cases it identifies. I use planning mode heavily before any code gets written.

Is anybody using these projects and what do users think of using these projects?

- Yes, these are real projects in production with real users :) They love them

I've tried them for some more complex projects, but I've found as long as I am working on isolated features/bugs, with good component breakdowns the agents won't cross paths. Planning mode also helps because you know exactly which files will be modified before going into execution mode.

If you want to keep features in separate Git worktrees, https://conductor.build/ is pretty nice.