# Action Shift by Never86’d

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Myke](https://cerebralvalley.ai/u/mykemueller)
- **GitHub:** https://github.com/mykemueller1-ctrl/action-shift-meta-hackathon
- **Demo video:** https://mykemueller1-ctrl.github.io/action-shift-meta-hackathon/
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2190

I’m Myke Mueller, founder of Never86’d and an operator at Community Tap & Pizza in Fort Dodge, Iowa. The question behind Action Shift is simple: the guest changed the order, so which version is the kitchen making? The food and labor still cost money when the crew makes the wrong version.

The demo follows a fictional crew meal. Muse reads a revised request and gets twelve pizzas and two pickup windows right, but leaves the named pickup location blank. The operator reads the original message, fills that one detail and records why. The original model failure stays visible. The completed draft is still unapproved. A person separately checks the $198 quote, timing and capacity before two simulated kitchen receipts acknowledge five and seven pizzas.

One added pizza puts that approval on hold. A new $215 review and acknowledgment produce current receipts for five and eight, while the earlier receipts stay in history. A separate November 4 example shows a reviewed restaurant rule surviving a real restart, requested delivery staying unchanged until explicit guest pickup consent, and an operator reply edit remaining unsent.

The source also includes a separately reviewed 11-pizza request using the improved Spark instructions. The video keeps the earlier human-completion example.

This standalone front desk implementation and its revisions were built November 4 during the event with AI coding assistance. My restaurant experience and Never86’d Mini Brain concepts predate the event. No earlier product code or private restaurant records are included. All operational records are authored fiction. Saved Muse results and local simulations are labeled. No live guest channel, POS, payment or kitchen integration is claimed.

## More from Meta Global AI Developer Hackathon

- [Cygnus-A²](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2187)
- [Ayushi Chadha](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2188)
- [Captain Planet and the Planeteers](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2189)
- [Jarvis Team](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2191)
- [SettleProof](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2192)
- [Synteron](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2193)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2190. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt