@sharaff@trq212 True. I do run local models via Ollama. It works well, but yeah, tweaking and debugging the output definitely eats up the time you'd save just letting Claude handle it. I think I'll get Codex and do a sideby side comparison against Claude
Took a deep-dive looking into Claude Code Channels, they are pretty slick, kudos to the CC team!
Played around with setting something similar up in Gemini CLI and got it to work.
Connect to a session from telegram ๐ฒ
most people don't need frontier intelligence for what they do daily. open models running locally with no rate limits, no costs and full privacy cover 90% of use cases. gap's closing fast too .
@sharaff@MiniMax_AI Yes at the moment it's an open pr.
The point is it works.
And thank you .
I love the work you do .
You inspired me to install openclaw.
https://t.co/4uDf7ryo1J
Just submitted a PR to Google's Gemini CLI adding real-time channel notification support for MCP servers. Now any MCP server can push messages into a running Gemini session enabling Telegram, Slack, Discord bots and more.
Tested on telegram.
๐จ Anthropic just made OpenClaw look like a prototype.
Dispatch lets Claude run your entire computer while you're gone.
No terminal. No GitHub repo. No setup.
Just text Claude a task from your phone and leave.
โ Claude spins up agents on your machine
โ Works your files, browser, and tools
โ Sends you updates while it works
โ You approve what it touches before it acts
โ Come home to finished work
Persistent memory too. Claude keeps context across every task.
@steipete Love the direction @steipete. Running v2026.3.13 with local Ollama models one thing that'd make plugin bundles shine=perworkspace skill scoping so agents don't bleed tool access into each other. Any plans for that in the new plugin model?
200+ Google and OpenAI staff have signed this petition to share Anthropic's red lines for the Pentagon's use of AI
let's find out if this is a race to the top or the bottom
https://t.co/3qgmaLfM0i