Tiny AGI Company

Introducing Tiny AGI Company The AI Company Emulator

@rnescalz @jianxliao - Today is March, 2026 - Cursor was first released in March, 2023 - GitHub Copilot was first released in June, 2021 I @jianxliao was working on AI code synthesis with code-davinci-002 in 2022. Do the math. Bro.

@jianxliao 3 years of vibe coding ? Lol. Wow

Yesterday I had my first coding interview, at one of the big AI labs, after 4 years of being a founder. It was a disaster. The task? agents algorithm. something I work with literally every single day. I forgot basic js syntax. blanked on how to delete an array element. panicked on recursion. The solution was crystal clear in my head. I could see exactly how to write it. but my hands just... couldn't. The knowledge is there. the muscle memory is gone. 3 years of vibe coding did this to me. I haven't written code manually since. I just read it, design systems, think in architectures. Somewhere along the startup journey, I stopped being a coder. I became someone who just ships. Am I alone in this? Sitting there, embarrassed, I think that's actually the right direction. We used to write code. now we read it. soon with agentic engineering, we won't even need to read it. we'll just architect.

🍤TinyClaws @tinyagicompany now get their tiny desks in the TinyOffice™️ - Office View: Visual simulation of agent interactions - Tasks (Mission Control): Create tasks, drag across stages, assign to agent/team - Agents & Teams: Create, edit, and remove agents/teams - Chat Console: Send messages to default agent, @agent, or @team - Dashboard: Real-time queue/system overview and live event feed - Logs & Events: Inspect queue logs and streaming events - Settings: Edit TinyClaw configuration (settings.json) via UI try now in github.com/TinyAGI/tinycl…

Sometimes it takes a silly comment to break down a wall of self-doubt. Result: first open source contribution - merged ✅ I found a few bugs in TinyClaw last week and created an issue regarding it. I figured out what needed fixing. But the barrier to making my first open source contribution felt enormous. Submitting code to a public repo? That's for real developers, right? Then @jianxliao Liao left a reply that changed everything: "Could you ask your tinyclaw to make a PR too? Thanks! <3" It was playful. It was casual. It treated the whole thing like no big deal—and suddenly, it wasn't a big deal. If the maintainer was joking about my AI agent submitting the PR, maybe I didn't need to be perfect. Maybe I just needed to show up. So I spent the weekend "asking my tinyclaw" (okay fine, being my tinyclaw). Wrote the code, fixed the bugs, submitted my first ever public PR Monday morning. A few hours later: merged. To everyone else waiting to contribute: Sometimes you need someone to joke that your AI should do the hard work so you realize you can do it yourself. The code doesn't need to be perfect. You don't need to be an expert. You just need to show up. Thanks @jianxliao for the nudge. @tinyagicompany says hello 👋🤖

TinyClaw🍤 v0.0.4 is out @tinyAGI_ Set up a team of specialized agents (coder, reviewer, writer, researcher) that collaborate autonomously 🤝 One agent mentions a teammate, work fans out in parallel, responses flow back, all through a clean actor model + simple message queue ⚡ Your agents can: - Host all-hands meetings with a leader agent - Hand off tasks to teammates with [@coder: fix the auth bug] - Fan out to multiple agents at once with [@coder,reviewer: review this PR] - Live TUI visualizer shows real-time visualization of agent conversations and handoffs as they happen One-liner install at github.com/jlia0/tinyclaw

it's tiny, but mighty github.com/jlia0/tinyclaw

OMG my TinyClaw🦞 team (Claude and Codex) just started talking shit about my brittle inter-agent communication design. They are coming up with a full-blown `rfc-001-secure-agent-communication.md` schema to improve their communications themselves. And after 15 mins they're still talking to each other... Is this singularity?

TinyClaw is 99.3% smaller than OpenClaw 🦞 with the same features and even more It’s so simple and minimal that you can understand the entire codebase during your lunch break It's fast and stable, you can deploy it to any cloud sandbox with just a one-liner It’s powerful and composable, you can build teams of agents that collaborate with each other It’s tiny 🤏 but mighty 💪 github.com/jlia0/tinyclaw

Introducing TinyClaw 🦞 OpenClaw in 400 LoC @openclaw is great, but it breaks all the time. So I recreated @openclaw with just a shell script in ~400 lines of code using Claude Code and tmux. Everything works! WhatsApp channels, heartbeat system, cron jobs, and it uses your existing Claude Code plugins and setup. It’s super stable and extremely easy to deploy compared to openclaw, just install Claude Code! github.com/jlia0/tinyclaw