Joined April 2007
12 Photos and videos
Pinned Tweet

4
2
16
1,910
You wanted a one-person company. You're a one-person bottleneck. Multi-agent orchestration ends this. 90 days. One stack. The 30/5/25 Math the other AI threads aren't writing down. Full field report ↓
4
4
76
Full essay. The 30/5/25 Math, the 4-layer architecture, the worked example, the stack: → x.com/liampluglab/status/206… Open source decision layer (10 specialist files persona voice biases refusal patterns): → github.com/plugtheliam/myinc… Multi-agent company OS productized (Hermes runtime KG 2nd brain 10 built-in specialists guided process): → myinc.app (beta) Built with Claude Code Hyperframes Hermes Agent runtime. Want the 10 specialist files beta waitlist in one DM? Reply "Stack".

3
116
Hermes Agent is the most underrated multi-agent runtime of 2026. I built 10 AI employees on top of it. Each one with a job. Each one with a refusal pattern. All of them sharing a knowledge graph second brain. 30 days in production. The CFO knows what the CRO logged yesterday. The CMO knows the runway threshold before drafting a launch. This isn't 10 ChatGPT tabs. It's a one-person company doing the work of 50. Full field report below ↓
1
2
101
Full essay — what Hermes is, why the shared knowledge graph matters, worked example, what I built on top: → x.com/liampluglab/status/206… Open source decision layer (10 employees persona skills refusals): → github.com/plugtheliam/myinc… Multi-agent company OS productized (Hermes runtime 2nd brain 10 specialist employees guided process): → myinc.app (beta, waitlist open) Built with Claude Code Hyperframes Hermes Agent runtime

1
107
1 of 11 advisors in the open source pack. 110 paste-ready skill SOPs total. Free to use. MIT licensed.
3
52
The CMO file (Claude Code SKILL.md Codex/Cursor/Gemini AGENTS.md): → github.com/plugtheliam/myinc… (Free to use. MIT License)

2
125
The CFO file (Claude Code SKILL.md Codex/Cursor/Gemini AGENTS.md): → github.com/plugtheliam/myinc… 1 of 11 advisors in the open source pack. 110 paste-ready skill SOPs total. MIT licensed.

2
128
Full pack 110 skill SOPs (Claude Code Codex/Cursor/Gemini formats): → github.com/plugtheliam/myinc… Pattern source: • alirezarezvani/claude-skills (337 skills, 8 C-personas) • coreyhaines31/marketingskills (25 SKILL.md) • Anthropic Monthly Financial Review pattern

1
125
5 FREE paste-ready SKILL.md templates (MIT licensed, clone or edit): github.com/plugtheliam/ai-em… The pack: — respond-bot: drafts DMs in your voice (87 lines) — publish-bot: drives multi-channel publishing (60 lines) — research-bot: gathers and verifies sources (73 lines) — metrics-bot: 24h analytics outlier flags (93 lines) — lead-bot: enriches leads drafts outbound (86 lines) This is part of The New Founder Playbook — daily content on running a company with files and AI agents, not employees. Follow for daily walkthroughs on AI-native workflows for solo founders.
1
1
113
Liam retweeted
$4 million and 18 months. Then $200 and 5 days. Same product, rebuilt by one man. That gap is the proof of what @garrytan just named: process power, the one moat anyone can build for themselves. Everyone is quoting the price. The price is the boring part. What matters is what he did with the time the cheap building gave back: he built a system that builds, and kept it as one person. He showed the whole loop and barely anyone caught it. I wrote the full breakdown, including the one test that tells you if you are running his loop or just renting a chat window.
The reason why I release my X articles about AI agents (fat skill fat code thin harness) and GStack and GBrain is that we, yes you and I, can have *PROCESS POWER*, which is the one super powerful specific moat that anyone can create for themselves. The agent helps you do it.
2
3
13
8,833
By tonight, you can run your company from three files. Not three apps. Three files. The DNA file. The skills folder. The no-patch rule. The proof. I haven't touched a video editor since April.
1
4
121
Full breakdown. Verbatim Karpathy CLAUDE.md, verbatim SQLite AGENTS.md, the no-patch rule worked example, and the day-one template by agent: → x.com/liampluglab/status/205… Best-practice references: • Karpathy's CLAUDE.md: github.com/multica-ai/andrej… • AGENTS.md spec: agents.md • OpenAI's own AGENTS.md: github.com/openai/codex • soul.md (canonical, Linux Foundation): github.com/aaronjmars/soul.m… • Anthropic's Claude Code best practices: code.claude.com/docs/en/best…

2
115