Tool Calling Models with Personality

<!-- SC_OFF --><div class="md"><p>Beginner here, now I have a good GPU and ollama using docker. Pulled the Gemma4 weights and was able to add it to cursor using ngrok. </p>

<!-- SC_OFF --><div class="md"><p>I build context/harness optimization tooling, so provider-side serialization quirks actually matter to me. If you're optimizing over prompts, you

<!-- SC_OFF --><div class="md"><p>Getting local models to make use of tools properly requires that I produce a multi-turn synthetic dataset. I find this process often tedious as I

<!-- SC_OFF --><div class="md"><p>Long time Claude Opus user, but after the recent session limit changes by Anthropic, I am seriously considering trying Chinese models for coding.

<!-- SC_OFF --><div class="md"><p>I've noticed something about Claude from talking to it. It's very very distinct in its talking style, much more of an individual than some other

<!-- SC_OFF --><div class="md"><p>Tomorrow, I am getting delivery of a 13&quot; M5 macbook air with 32gb RAM and 1T SSD. Currently, I have pro subscriptions to Gemini, Perplexity

<!-- SC_OFF --><div class="md"><p>In LM Studio the new Qwen 3.5 models (4b 9b 27b 122b) when analyzing large (more than 50k tokens) texts start to output gibberish. It is not a

<!-- SC_OFF --><div class="md"><p>Most of my code is similar to agent-cpp from Mozilla. I create common_chat_templates_inputs Inputs from message history.</p> <p>auto params =