John Palmer on PartyDow joining Stripe, vibe coding hype, and the Mac Mini open-source moment
Feb 26, 2026 with John Palmer
Key Points
- PartyDow is joining Stripe; co-founder John Palmer argues that as software production costs collapse, companies are actually acquiring developer mindshare and the audiences that follow influential engineers.
- Mac Mini hype outpaces practical utility: Palmer tested the device for 48 hours and abandoned it after discovering browser automation is too fragile for real workflows, though he expects meaningful capability gains in three to six months.
- Palmer envisions local AI agents evolving from single-player productivity tools into a multiplayer network where agents negotiate with each other, price their own services, and others micropay for context-specific judgment.
Summary
John Palmer, co-founder and CEO of PartyDow, is joining Stripe. He spent 48 hours setting up a Mac Mini and has barely touched it since. Browser-based automation proved fragile—he couldn't get past a Google Sheets share modal when trying to invite someone. His main coding workflow, working remotely on repos from his phone, was already covered by Claude Code. The Mac Mini added almost nothing new.
Palmer is not bearish on the hardware itself. Setup friction is real, browser tools will improve, and the same device should look meaningfully more capable in three to six months. The gap between buying the device and finding a daily workflow is simply wider than the hype suggests.
What interests him more is the cultural shift. People who hadn't opened a terminal in years are opening one again. Vibe coding carries the aesthetic credibility of hacker culture returning after a long absence. The terminal, even when you're just prompting in plain English, feels more legitimate than slicker desktop tools.
But most things people want to vibe-code already exist as free or cheap apps. The genuinely interesting use cases—paywalled content, cross-platform automation, anything that bumps into terms of service—are popular precisely because they're difficult to fully productize.
As software production costs trend toward zero, what companies actually acquire shifts. Engineering skill becomes more of a commodity. What you're buying is the audience that follows the engineer. Open Claw became the fastest-growing open-source repo of all time and is forkable by anyone. The reason to acquire Pete Steinberger is that you're acquiring every developer with notifications on for his posts and who spent $600 on a Mac Mini to run his model.
Software engineering job postings have risen since vibe coding went mainstream, which fits the interim-period logic. AI is doing 90 percent of coding, engineers become highly leveraged, and demand goes up. The concern is what happens when it tips to 100 percent. The leverage collapses quickly. Palmer takes a somewhat more optimistic view. Displaced software engineers are smart, hardworking people with AI tools to help them reskill. They'll compete across finance, sales, and other domains rather than simply disappear from the workforce.
The most speculative thread is Palmer's vision for what Open Claw-style local agents could become. Your Mac Mini, trained on everything you've written and spoken and every file on your system, becomes a personal agent. Not just a single-player productivity tool but something others can query directly. Instead of accessing a generic frontier model, a third party could micropay your agent for its specific context and judgment. Agents could negotiate with each other, price their own services, and operate in parallel while their owners are doing something else entirely.
Security infrastructure isn't close to ready for this. Palmer set his Mac Mini up with its own isolated Gmail and iCloud because he won't let it touch his real accounts. But the direction he's pointing—from single-player assistant to multiplayer agent network with economic primitives—is the longer-horizon bet he finds genuinely interesting.