Robotics: The House, the Cost, and the Test Report

The house as a cluster, the cost of a capable machine, and the same audit questions that show up in software agents.

Robots are computers that can reach you. I write about the house as a cluster, the cost of a capable machine, and the same identity and audit questions that show up in software agents.

Start with the house, then the price. The connected-house posts walk room by room. The cost posts read the Anthropic robot report the way a test engineer reads a datasheet.

Start here

  1. The house is full of computers now. Anthropic just showed what an AI could do with them. Anthropic tested whether an AI could copy itself onto cheap IoT hardware. What the paper showed, what it did not, and why a house full of computers changes the risk.
  2. The Robot Is the Next Node in Your House Cluster Home humanoids are coming with cameras, network links and remote operators. Treat a house robot like the newest node in your cluster: inventory, network, data, and off.
  3. Capable Isn't Cheap: Reading Anthropic's Robot Report Like a Test Engineer Robots can do 74% of physical tasks but are cost-competitive on 0.3%. Why the gap is cost, not capability, read the way a test engineer reads a datasheet.

Every post in this hub

  1. Free Plumbing, Paid in Footage
    A company will pay for your plumber if the technician can wear a camera to train robots. Read it like a data deal. Here's the checklist I'd use before saying yes.
  2. Who Signs the Test Report for a Robot?
    A startup wants AI to red-team robots in simulation and hand the evidence to insurers and regulators. From a test lab view, the hard part is who certifies the tester.
  3. Changeover Time Is the Spec Nobody Quotes
    A factory robot startup says AI stations can retool a line in minutes, not weeks. Changeover is where uptime goes. How I'd measure a claim like that.
  4. Hardware Is AI's Teeth: Robots Are Building the AI Buildout
    Robotics leasing in SF passed 2020 to 2025 combined, a $450M HardTech fund closed, and robots now assemble AI servers. What the hardware side of AI looks like.
  5. The Robot Is the Next Node in Your House Cluster
    Home humanoids are coming with cameras, network links and remote operators. Treat a house robot like the newest node in your cluster: inventory, network, data, and off.
  6. Monitor a Humanoid Like a Production Service
    Robot ROI lives in uptime, interventions and repair time. Why I'd put a humanoid fleet on SLOs, dashboards and alerts the same way I'd run a production service.
  7. Two Percent a Year: The Slow Curve Behind the Robot Headlines
    Each year robots learn about 2% of the physical work they couldn't do before. Why that slow curve, plus 3% yearly price drops, sets the real robot timeline.
  8. The Last 20 Percent: What Robots and LLMs Still Can't Do
    Robots plus LLMs touch about 80% of work time, per Anthropic. The rest is hands-on, face-to-face and often regulated. What that last 20% tells us about jobs.
  9. The Hard Part Is the Hands: Why Sensing and Touch Hold Robots Back
    Drivers and warehouse workers are most exposed to robots. Nurses and mechanics aren't. Anthropic's data says the reason is touch, not intelligence.
  10. The 20-Hour Robot: Duty Cycle Decides Robot ROI
    Agility says Digit 5 will work 20+ hours a day on a 9-minute charge. Why duty cycle, batteries and uptime decide whether a robot pays for itself.
  11. Capable Isn't Cheap: Reading Anthropic's Robot Report Like a Test Engineer
    Robots can do 74% of physical tasks but are cost-competitive on 0.3%. Why the gap is cost, not capability, read the way a test engineer reads a datasheet.
  12. Room by room: what a self-copying AI could do in your house
    A walk through the connected house, device by device: what each one has, what has really happened to it, and what Anthropic's self-replication research says.
  13. The house is full of computers now. Anthropic just showed what an AI could do with them.
    Anthropic tested whether an AI could copy itself onto cheap IoT hardware. What the paper showed, what it did not, and why a house full of computers changes the risk.
  14. LLM Robot Fleet Governance with agentgateway
    Every inference, MCP tool call, and agent message in an LLM-driven robot fleet runs through agentgateway: tool authz, model failover, and latency cost.
  15. The Future of AI and Robotics: A Look at What's Next
    Exploring the rapidly advancing world of AI and robotics, from autonomous vehicles to humanoid robots, and what it means for humanity.

Frequently asked questions

Why write about home robots next to agent gateways?

A home robot is another computer on a network you already run, with motors. The identity, audit, and cost questions are the same ones that show up when an agent calls a tool.

Where should I start?

Start with the connected house if you want the threat model. Start with the cost column if you want the datasheet. Both threads are listed below.