Ollama
4 storiesOllama is a tool for self-hosting and running large language models and local LLM development. Recent coverage highlights gotchas migrating preprompts from Opus and inference privacy risks when using frontier providers, a Zima Board 2 + RTX 2000 ADA setup to run Qwen‑3.8, and a Mac how‑to using OpenCode and sbx that recommends Qwen 3.8 and Gemma 4.
Related topics
Pizza Bot is a cross-platform self-hosted desktop app (Mac/Windows/Linux) that manages background AI agents via an email-like UI. It’s Apache 2.0-licensed, requires no signup or telemetry, and supports Anthropic, Amazon Bedrock, Google Gemini, OpenAI, OpenRouter and local models via Ollama. The project originated from an internal Amazon team and is released as open source.
Hacker News covers gotchas migrating preprompts from Opus to self-hosted Ollama and warns that running inference on frontier providers can leak session metadata and be used for training; the thread references the Navier–Stokes episode and distrust of OpenAI and Anthropic.
A Reddit thread discusses using a Zima Board 2 (~$411) plus an RTX 2000 ADA (~$700) to run Qwen-3.8 27b, citing a Luke’s Dev Lab demo that achieved good token speed on Ollama powered by the Zima supply. The poster compares this setup to a Mac Mini M5 24GB and asks if cheaper new self-contained alternatives offer similar token performance.
A how-to describing setting up a local LLM development environment on Mac using Ollama, OpenCode, and Docker sandboxes (sbx), recommending Qwen 3.8 and Gemma 4 and providing install/config steps.
That is everything