Dev builds local AI GTM workflow, argues the next platform entry point is hardware-bound
dotey · x · 2026-09-25
Countering the "Muse as the next WeChat" thesis, the author argues no super app can be the next-era entry point—the entry will be hardware-bound, sold at the OS level.
The argument comes from a GTM workflow he built himself:
- Standard Ubuntu/macOS VM images with browsers and IM apps preinstalled;
- A custom collector scrapes IM groups and social media to find prospects matching user profiles;
- Chrome profiles stored outside the VM avoid re-login; non-web apps have workarounds;
- AI drafts outreach materials, but a human sends them.
Core logic: any IM with a GUI can be run in a VM and mined by AI, so "my AI + my intent" builds connections across all platforms—meaning IMs can't gatekeep the next entry point. He notes the workflow has a data flywheel owned by the GTMer: everything runs on local Qwen. His setup: M4 Pro Mac mini 64GB + 2× DGX Spark, costing over ¥100k, though he says the M5 Ultra Mac Studio 256GB is now the better value, and hardware prices will only fall.
More from coding & agent
- Five mistakes people make with agentic factories: micromanaging, no quality bar, and 'puppet shows' — hugobowne · 2026-09-25
- Open Claude Code skills catalog ships 31 testing skills, from metamorphic testing to E2E — doodlestein · 2026-09-25
- Researcher proposes agent-suggests-human-executes loop for real-world science experiments — suragnair · 2026-09-25
- CrawlRaven hits 700 users in 60 days, targets 2,000 more with MCP-powered SEO audits — ayushtweetshere · 2026-09-25
- Human-in-the-loop or machine-executed: verifiable tasks turn agent traces into RL rewards — suragnair · 2026-09-25
- Formal methods hackathon on Nov 1: build games and AI agents with formal verification — xennygrimmato_ · 2026-09-25