Solo UnicornLab

A research lab forsolo unicorns

We study how one person with a team of AI agents can find a real problem, ship the product, and keep improving it with the people who use it.

Read the research How we work

About the lab

Building for an
AI-native world.

Solo Unicorn Lab is an independent initiative advancing AI-native ways of working.

We explore how people and agents can share context, coordinate decisions, and turn real problems into useful products. Our focus is the whole organization: how work starts, how it gets done, and how it improves.

Human direction. Agent collaboration. Learning from the world.

Explore our approach
lab / research

Find problems
worth solving

Notes on how agents spot real demand, work together without stepping on each other, and check a claim before anyone builds on it.

Read the notes 2 notes4 open questions
Fig. 1 — A field of weak signals. One peak is worth a closer look.
Illustrative
  1. readsignals/founder-forums/312 threads90ms
  2. grep"i would pay for"+14 matches140ms
  3. writeproblem-brief.md+1 brief −9 hunches620ms
  4. askfounder: worth a week?waiting on you—

Agent is sorting complaints into problems worth a closer look

lab / method

Test the
whole loop

From a user’s problem to the next release: who did what, what changed, and whether it helped. Every handoff stays on the record.

See the method 6 stagesone loop
Fig. 2 — One founder sets the heading. The crew keeps formation.
Illustrative
  1. planrelease-0.43 agents assigned210ms
  2. editcontract.yaml+2 limits80ms
  3. testcheckout-flow2 failing 31 passing1.9s
  4. ship--to productionneeds your sign-off—

Agent is fixing the two failing checks before it asks again

lab / careers

Build it
with us

We are looking for a Member of Technical Staff who moves between research and engineering, and ships what they design.

Explore the role 1 roleopen now
Fig. 3 — Small pieces, set down in order, until the thing stands up.

An open invitation

Bring a real problem.
Start a conversation.

What are you building? Where does the work get stuck? Tell us what you would like to figure out together.

Submit a problem

FAQ

Fair questions

What does the lab actually study?

How an AI-native organization works when the organization is one person and a team of agents. We work on four questions:

  • How to find problems worth solving.
  • How people and agents, and agents with each other, can work together reliably.
  • How real user feedback improves both the product and the way it is made.
  • How distribution and shared capabilities could support a new kind of work.

See the research areas

Is this a company or a research group?

An independent research initiative. We write notes, design experiments and build the systems needed to run them. No institutional affiliation is implied.

Are the results on this site real?

Not yet, and we label that wherever it matters. The published notes are AI-assisted literature syntheses with proposed experiments that have not been run here. The agent logs and drawings on this page are illustrations, not live activity.

Who is it for?

Solo founders and small teams building with agents, and anyone studying how people and agents share real work.

Can I bring you a problem?

Please do. Real product or service problems, examples of human and agent collaboration, and disagreements with our notes are all useful to us. Write to hello@solounicorn.club.

How do corrections work?

Send the page link and the passage concerned. When a revision is substantive, the note explains what changed and whether a conclusion was affected. More on corrections

Are you hiring?

One role is open: Member of Technical Staff, for someone who moves between research and engineering. Read about the role