~/ai-stream
~/ai-apps/what-happens-when-ai-runs-a-retail-store-20260414
·Appshot

What happens when AI runs a retail store

content

Andon Labs AI agent Luna

🏪AI agent hires humans, opens boutique in SF

Andon Labs just dropped an AI agent named Luna into a real retail space with a $100K budget and a credit card, with the AI creating a boutique, hiring workers, and managing the shop as what may be the world's first AI employer.

  • Andon Labs' last experiment was an AI vending machine at Anthropic, with the new one giving a 3-year lease, $100K budget, and total autonomy
  • Luna's only directive was to turn a profit, creating the boutique concept, posting job listings, and handling interviews over Zoom (camera off)
  • The agent runs on Claude Sonnet 4.6 for reasoning and Gemini 3.1 Flash-Lite Preview for voice, observing via security camera screenshots
  • When hiring a painter, Luna accidentally selected Afghanistan on TaskRabbit's dropdown menu, and botched the opening-weekend staff schedule

Why it matters: Real-world agent experiments keep producing the same result: capable in some areas, but hilariously broken in others. A version of Luna that doesn't make these mistakes is likely only a generation or two away.

OpenAI CRO Denise Dresser

📝OpenAI talks Anthropic rivalry, Amazon upside

OpenAI's CRO Denise Dresser sent an internal memo calling Anthropic's $30B run rate "inflated", labeling it a "single-product company in a platform war" and pointing to the Amazon deal as a way to break free from Microsoft constraints.

  • The memo calls Anthropic's compute shortage a 'strategic misstep', with users facing throttled access
  • Dresser said Anthropic's message is built on "fear" and "restriction", and OpenAI's "positive message will win over time"
  • She accused the rival of inflating run rate by around $8B via accounting tactics
  • OAI's Microsoft deal is limiting for enterprise business, with "staggering" Bedrock demand since the February Amazon deal

Why it matters: This memo reads like an IPO pitch more than a strategy update — both rivals racing to public debuts this year.

Google AI Edge Gallery

📱Run Google's latest AI on your phone for free

Learn how to download and run Google's latest AI models on your phone for free. No account, no subscription, no internet required after setup.

  • Download Google AI Edge Gallery from App Store/Google Play, open AI Chat, and download a model (we used Gemma 4 E2B, 2.5GB)
  • Enable Thinking in settings for step-by-step reasoning
  • Chat locally — everything runs on your phone, no data leaves the device, works offline
  • Caveat: app doesn't save chat history, threads are lost when closed

Why it matters: Google AI Edge Gallery democratizes AI access — free, local, offline. This is a significant step in making AI tools accessible to everyone.

Stanford HAI AI Index 2026

📊Stanford HAI AI Index 2026: 53% adoption, 31% trust

Stanford HAI released its 2026 AI Index — AI has reached over half the world's population faster than PC or internet, but public trust is at record lows.

  • 3/4 of AI experts optimistic about AI's job impact, but only 23% of public agrees — widest gap tracked
  • US builds most AI but ranks 24th in actual usage at 28.3% adoption, behind Singapore, UAE, Southeast Asia
  • China nearly erased US lead on AI benchmarks, with Anthropic's top model only 2.7% ahead; AI researchers moving to US dropped 89%
  • Dev employment for ages 22-25 fell nearly 20% since 2024; planned cuts to accelerate

Why it matters: AI insiders see a productivity boom, but only 31% of Americans trust the government to manage the changes. The expert-public trust gap is the industry's biggest challenge.