Every person leads a team of ten agents

Homebot One Inc describes itself as a superintelligence-native company. In practice, the company says, every person on the team leads a team of ten AI agents that work across research, code, design, and operations.

That changes what one person can take on. An engineer can have agents drafting test plans, reviewing code, and summarizing results while they focus on the hardware in front of them; a marketer can have agents preparing drafts while they talk to families and partners.

Managing agents like a team

Running ten agents is a management job. Homebot One Inc says it hires people who can split the work, review what comes back, and keep every thread moving, the same skills it takes to lead a team of people.

Review matters most. Agents can produce a lot quickly, and someone with judgment has to decide what is right, what needs another pass, and what should never ship.

Pointing superintelligence at the hardest problems

The company’s Careers page says it wants people with the nerve to point superintelligence at the hardest problems in the world. Building a robot that is genuinely helpful in a real home is one of them: homes are messy, every family is different, and safety and privacy come first.

Open roles reflect that range, from hardware engineering, world models, assembly, and supply chain to marketing and sales, all based in Fremont, California.

What the phrase does not claim

Superintelligence-native describes how the team works. It is not a claim that KOKO itself is superintelligent. Homebot One Inc describes KOKO as running on a vision-language-action model whose skills are still growing, and says it will share honest progress updates as each new task becomes reliable.

For families, the product claims to rely on are the written ones: what KOKO does today, its safety design, and its privacy commitments. The way the company works is context for how quickly it can learn and improve, not a feature of the robot.

Why it matters for families and partners

A small team that works this way can test more ideas, write more documentation, and respond to feedback faster than its headcount suggests. For families and care partners, the practical effect should be clearer answers and steadier progress updates.

If you are evaluating KOKO, ask the team how agents fit into testing and support, and which decisions always stay with people. The answers say a lot about how a superintelligence-native company keeps humans responsible.

Frequently asked questions

Is KOKO superintelligent?

No. Superintelligence-native describes how the Homebot One Inc team works. The company describes KOKO as running on a vision-language-action model whose skills are still growing.

What does it mean that everyone leads ten agents?

Homebot One Inc says each person on the team runs a team of ten AI agents across research, code, design, and operations, and is responsible for reviewing their work.

Where is the team based?

Homebot One Inc is based in Fremont, California, where it builds KOKO.

Sources & further reading

  1. Careers at Homebot One Inc (opens in a new tab)
  2. About Homebot One Inc (opens in a new tab)
  3. KOKO Robot on Kickstarter (Homebot One, Inc.) (opens in a new tab)

From Homebot One Inc, the team building KOKO in Fremont, California.