@mattpocockuk I recently started using more often /diagnosing-bugs instead of /grill-me or /grill-with-docs like before for tasks which are not new features, but just bug fix. Its game changer 🔥
@peter_kow yes, all works happening on Mac Studio as headless runner, with Hermes Agent etc, but if i need to do anything in Claude/Codex i still run it on remote from my macbook via T3 Code :-)
Announcing Apex: agentic coding model for mobile and web, specialized in React.
We built Apex from a single idea: can an open-weight model deliver frontier-level performance at lower cost?
Yes. Apex is SOTA in React Native and Next.js.
Try it now: https://t.co/nKbBxcmRFX
Introducing Gemini 4 Argon – our new frontier model.
It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
@ferrannp maybe I did something wrong if that works for your OC instance. My Hermes for agents + few dockers + iOS/Android simulators + all remote harnesses were not performing well on that base mini. I think it mainly struggled with local LLM for Honcho memory for agents 🤖
Meet TraceSift ⚡️
An open-source tool that turns CPU and React profiles into actionable performance insights without flooding your agent’s context.
Early benchmark (GPT-5.6 Terra):
• Codex: 45K tokens
• TraceSift: 1.6K tokens
That’s ~28× fewer tokens 👀
We’re excited to share this one: the Agent Conf talks are live on YouTube.
A huge thank you to every speaker who helped us shape this conference.
Watch Agent Conf on YouTube 👉https://t.co/KnRgHHOJmO
Which talk are you starting with?