Observations from the intersection of restaurant operations and AI implementation. Operator-level. No hype.
Jack Dorsey announced four thousand cuts at Block in late February and framed them as AI-driven restructuring. The skeptical reads were expected. Ten weeks later, Square is shipping a quarter of operator-requested F&B features and has another quarter scheduled. Cuts of this scale historically slow shipping. This one accelerated it. Here is what the structural read looks like from the operator side of the cycle.
Read →AI in restaurants needs three layers, not two — a system that watches, a system that remembers, and a system that acts. The platforms that bet on open are betting on the operators who matter most over the next decade.
Read →Nothing we do in a restaurant is hard on its own. Making a sandwich, greeting a guest, dropping chicken in a fryer, are all easy. Doing twenty of those things at the same time is what makes it hard. That compounding is what most software misses, and what most AI is promising to solve without actually knowing what it is.
Read →Toast just spent seven figures telling every operator in New York that they are built for busy. The campaign is a brand play. It is also a tell about the AI bet Toast is making for the next three years. Who else will be sitting at the table to call?
Read →Square just announced ManagerBot publicly this week. I have been using it for weeks. Here is what it actually feels like to manage restaurants with an AI agent riding shotgun.
Read →A survey asked 600 restaurant operators where AI could deliver the most value. 38% said scheduling and labor. 3% said dashboards. Everyone keeps talking about the 38. The 3% is the number that tells us more.
Read →Jack Dorsey predicted most companies would follow Block's AI-driven restructuring within a year. Mark Zuckerberg did not wait the month. What the CEO agent race means for every business with layers between decisions and data.
Read →Zig Ziglar and Tony Robbins are inspiring Skynet on how to succeed. NVIDIA GTC 2026 just validated decades of "the power of positive thinking" by proving that visualization before action makes your actions better.
Read →Nvidia's CEO compared agentic AI to the arrival of the internet. He was talking to data center operators. He should have been talking to you.
Read →Every station is a node. Every ticket is a message. Every expeditor is a load balancer. The restaurant already runs on distributed systems principles. The AI tools being built for it mostly don't know that yet.
Read →No inventory counts. No developer. No ongoing cost. A single-purpose AI system built in under two days delivered food cost tracking within 1-2% accuracy for an independent diner that had no systems at all. Food cost dropped 5% in under two months.
Read →They let calculators into the ACT. Nobody mourned long division. The same thing is happening to basic programming right now, and the institutions haven't caught up yet.
Read →A 40-facility skilled nursing organization in the mid-Atlantic brought in outside help after clinical staff lost confidence in their contract food service provider. Two weeks of on-site consulting across two campuses, 300 beds total, produced HACCP gap remediation, AI-assisted CMS tag targeting, and a training pipeline that made corrective education repeatable at half the development time.
Read →