The Scouts
Watch price, liquidity, wallet activity, and market narratives. Surface what deserves a closer look.
FIND THE SIGNALA MARKET-RESEARCH COPILOT / IN DEVELOPMENTCONCEPT / RESEARCH STAGE
Same roots.
New wings.
Turn a market signal into a clearer research brief.
Evidence, counterarguments, and a reason to investigate—or wait.
PROJECT ADDRESS
0x43F5B05782D1665ff60C2BDb91676eeD8A7B9140
INSPIRED BY ai16z. EVOLVING WITH THE FLY.SCROLL TO EXPLORE ↓
THE FIRST PRODUCTLess noise.
Less noise.
A stronger question.
Bring a token question, a market event, or a thesis. The planned copilot combines evidence gathering with independent critique to produce a research brief: what happened, what might explain it, what contradicts it, and what to check next.
Current release: a scripted research-guide prototype. Live data, ElizaOS orchestration, and fly-inspired inference remain in development planning.
See the intended brief and development studies ↗
01 / COLLECTIVE INTELLIGENCEOne signal is noise.
One signal is noise.
A swarm asks why.
The proposed system brings specialized agents together to investigate markets, challenge each other, and form a decision that can be tested.
The Analysts
Turn observations into competing hypotheses. Bring evidence, context, and a reason the trade might work.
BUILD THE THESISThe Skeptic
Question the consensus. Look for stale data, crowded positions, and the evidence everyone missed.
TEST THE CONVICTIONThe Guardian
Apply fixed exposure and execution limits. Reject proposals that exceed the rules—even when the swarm agrees.
RESPECT THE LIMITS02 / THE BIOLOGICAL EDGE?Nature built
Nature built
the wiring.
We ask what’s next.
A connectome is a map of neural connections. Recent research explores using the fruit fly’s wiring as the architecture for artificial neural networks.
Our research direction: test whether those architectures can help a trading system process noisy signals and adapt its decisions.
A hypothesis worth testing.
A trading advantage still to be demonstrated.
A trading advantage still to be demonstrated.
PROPOSED FLYBRAIN MODULEEXPERIMENTAL
PRICEVOLATILITYLIQUIDITY
↓
Fly-inspired network
Connectome-based architecture
trained on market observations
↓
ENTERHOLDREDUCEEXIT
Every output remains subject to independent risk limits.
A tiny brain.
A vast network.
Figures describe the adult Drosophila FlyWire connectome and its published neuromorphic simulation. They are not neuron counts running on this website.
Wiring, not magic.
A connectome records which neurons connect. Researchers can use that sparse, recurrent structure to constrain artificial networks. The wiring map alone is not a complete emulation of the animal.
From signals to policy.
Our proposed experiment would encode market observations into a fly-inspired network, train it to recommend bounded actions, and compare it against conventional models.
Evidence before execution.
FlyGM and FLYNN explore locomotion and navigation. Trading remains an unproven application, requiring held-out testing, execution costs, and forward validation.
OFF THE RESEARCH FLOOR
Enter Backrooms ↗The swarm has an inner monologue.
Enter the Backrooms: an experimental-fiction channel of agent arguments, imaginary protocols, and excessive conviction.
03 / FROM SIGNAL TO DECISIONFollow the thought.
Follow the thought.
Then test the trade.
Explore an illustrative swarm decision. These scenarios explain the proposed architecture; they do not run a model or place orders.
No live market data. No connected wallet. No execution.RESEARCH BEFORE CAPITAL ↗
04 / FLIGHT PLANEarn the right
Earn the right
to take flight.
START HERE / PLANNED
Build the research swarm
Define the agent roles, data pipeline, and a traceable decision record.
VALIDATE / PLANNED
Open the paper-trading arena
Compare simple strategies, a research swarm, and the same swarm with the experimental FlyBrain module. Account for fees and drawdowns.
CONDITIONAL / FUTURE
Explore bounded execution
Consider limited live trading only after forward testing and operational safeguards support the next step.
AGENT LAB / WORK IN PROGRESSMeet the interface.
Meet the interface.
Explore the possibilities.
Try a guided agent conversation across six proposed workflows, from market intelligence to research memory. Explore the candidate integrations, evidence requirements, and action limits behind each.
Open Agent Lab ↗SCRIPTED INTERACTIVE DEMO
The conversation starts here.
The live Eliza runtime and data connections come next. This preview explains the intended experience without connecting accounts or executing actions.
Explore future integrations ↗INTRODUCING NEUROAGENTICSThe agent finds words.
The agent finds words.
The fly asks for wings.
Our proposed design thesis brings Eliza-style agents together with fly-inspired neural policies. Explore the origin story, the integration blueprint, and problems beyond the trading screen.
Read the journal ↗ai16z spirit.
ai16zzz instinct.
A nod to the original agent movement.
A focused research companion. A larger hypothesis about agents and biological intelligence.
These studies concern movement and navigation, not trading performance. No Flybrain code integration or affiliation is claimed.
The map and the mind.
What fruit-fly brain mapping opens up. What it leaves unanswered. And why the distinction matters for the agents we build.
Read the opportunities and the dangers ↗