Great question. I don't have enough testing yet. So far just some basic, "is it working" on a small repo.

*Require*, not let.

Thanks for the advice!

What I did like about OpenClaw: 1. Simple, guided install. 2. Relatively simple UI and config UI. Disclaimer: I admit I am new to the local llm world, and was learning as I went.

3. After uninstalling the store version and installing the website version, reconfiguring everything from scratch. I finally gave the OpenClaw agent a name and tried out a prompt, to which it promptly crashed. Upon restart, the gateway was hopelessly borked again. (more)

2. While trying to troubleshoot/identify aforementioned defect, I clicked the toggle on the tray icon, rendering the gateway hopelessly broken. (more)

Yes. 1. The store version runs a scheduled task (gateway cleanup) every minute that opens a console host window that is supposed to be invisible, but is in fact visible. The binary from the OC website does not have this defect (more)

I dont have enough data yet. So far just installed qwen3.8 and finally got everything wired after some trial and error. Will definitely report back!

3. Hermes is impressive. Can configure it to run it's own model or even piggy back off the Ollama instance. The installer is impressive for Windows. So far, running qwen3.8. GPU topping out around 61C under load. Curiosity to hear others' experiences.

Here's my take after a week and a half with my @surface Ultra 64GB, trying to config and use local models for coding and agentic AI" 1. OpenClaw is a dumpster fire and not worth time installing. 2. Install Ollama and the Continue ext. in VS code to code against a local model 1/

? air be and what now?

holy cow! In addition to this - for those that are on the receiving end of screenshots and want to copy out the error code, in Windows 11 you can press WIN + Shift + T to extract text on screen

what a moron

Kid, we're talking about assets, not imaginary assets. Take a seat.

Level 53+. 2 weeks experience.