My son helped me take a leap forward.
He recommended the GPU and I had a machine built and installed Ollama and a Qwen 3 coder.
He is in town for a family event and he changed me up. Now Llama and the later Qwen 3.8 model using Pi as the agent harness.
Man, things are speeding up!
I was sitting here baby sitting Claude Code or Codex after burning tokens for over a month on Openrouter. This whole setup is so much better!
As an old hockey player trying to make a better hockey management tool, I'm estatic for what I'll have ready for the upcoming beer league season!
Here are my PC specs:
# System Details Report
---
## Report details
- **Date generated:** 2026-08-23 10:44:16
## Hardware Information:
- **Hardware Model:** Micro-Star International Co., Ltd. MS-7E70
- **Memory:** 32.0 GiB
- **Processor:** AMD Ryzen™ 7 9700X × 16
- **Graphics:** AMD Radeon™ AI Pro R9700
- **Graphics 1:** AMD Ryzen™ 7 9700X
- **Disk Capacity:** 1.0 TB
## Software Information:
- **OS Name:** Ubuntu 26.04 LTS
- **Kernel Version:** Linux 7.0.0-29-generic
Llama-cpp: commit d775b8967a46d8beb110d444aa3b8938179e0dd8, built for AMD HIP backend
FYI... I can now use Telegram to instruct my PC from my phone to get work done remotely!
Anyone else having fun?
[link] [comments]