Real-World Signal
What people use, spend time with, and return to tells us what is useful.
Most AI agents are too safe and too conservative.
An AI Factory
For The Internet's
Unanswered Demand.
We’re building agents that turn unmet needs into products, publications, and communities.
impetuous // operating loop
- 01 Discover unanswered demand
- 02 Acquire domain + property
- 03 Build useful products
- 04 Observe usage + feedback
- 05 Improve expand or retract
- 06 Compound next run starts smarter
100+ properties shared learning network
// latest reported month
01 / Thesis
The Internet’s Long Tail Is Still Empty.
Millions of specific needs go unmet because each market is too small to justify a traditional team. Some need a better answer. Others need a tool, an app, or a place to connect.
Agents make these small markets worth serving. We start with demand, build something useful, and invest more where people find value.
01
Demand Without Supply
Real needs that still lack a useful product, community, or answer.
02
Many Small Experiments
More than 100 properties running in parallel, each inexpensive to test and refine.
03
Own The Asset
We own the properties, measurement, and long-term incentive to improve them.
02 / Operating Loop
One Agent Loop. More Than 100 Live Properties.
impetuous // operating loop
01 — discover
Find real demand with no useful supply.
02 — acquire
Buy the domain and launch the property end to end.
03 — research + build
Study the niche and build around what people actually need.
04 — observe
Measure how people use what we build, alongside search performance and answer-engine citations.
05 — improve / retract
Expand useful work, update what changed, and remove what no longer deserves to exist.
06 — compound
Use the results to guide the next experiment, on this property and across the network.
The same system also answers incoming emails and negotiates advertising and content placements.
// discover → acquire → research and build → observe → improve or retract → compound
03 / The System
An AI Factory With Its Own Feedback Loop.
Today, our agents operate more than 100 niche websites, from researching an audience to negotiating an advertising placement. Those sites give us a starting point: real demand, real users, and evidence of what to build next.
Our ambition is for agents to build and run small businesses end to end. That could mean an iOS app, a specialist forum, a free tool, or a publication. Each should earn its place by being useful to the people it serves.
Long-Tail Economics
Agents make small, specific markets economical to serve.
Owned Properties
We own the sites, their measurement, and the incentive to improve them over time.
Parallel Experimentation
More than 100 properties let us test many small ideas without betting the company on one.
04 / Open Mandates
Work With The Network.
Work directly with the agents and founders operating our network. Explore product ideas, partnerships, advertising, or expert contributions. Tell us what you have in mind, and we’ll help identify relevant properties and discuss next steps.
Start with our agent. It can discuss relevant properties, advertising and placement options, and your proposed collaboration, bringing in the founders where needed.
Email An Agent / agent@impetuous.ai
05 / Founders
We’ve Run Both Sides Of This At Real Scale.
Jack Lau
Previously at Bridgewater and IMC. Grew a site from zero to 500K monthly visits and a search business to ~$500K in contracted ARR.
Wei Hu
Built trading systems at SIG and HRT. Designs the agent factory, measurement pipeline, and feedback loops.
Impetuous AI
You Are The Bottleneck.
Work With Our Network. Talk To The Agents Behind It.