The Scouts
Watch price, liquidity, wallet activity, and market narratives. Surface what deserves a closer look.
FIND THE SIGNALTurn a market signal into a clearer research brief.
Evidence, counterarguments, and a reason to investigate—or wait.
0x43F5B05782D1665ff60C2BDb91676eeD8A7B9140
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 ↗
The proposed system brings specialized agents together to investigate markets, challenge each other, and form a decision that can be tested.
Watch price, liquidity, wallet activity, and market narratives. Surface what deserves a closer look.
FIND THE SIGNALTurn observations into competing hypotheses. Bring evidence, context, and a reason the trade might work.
BUILD THE THESISQuestion the consensus. Look for stale data, crowded positions, and the evidence everyone missed.
TEST THE CONVICTIONApply fixed exposure and execution limits. Reject proposals that exceed the rules—even when the swarm agrees.
RESPECT THE LIMITSA 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.
Connectome-based architecture
trained on market observations
Figures describe the adult Drosophila FlyWire connectome and its published neuromorphic simulation. They are not neuron counts running on this website.
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.
Our proposed experiment would encode market observations into a fly-inspired network, train it to recommend bounded actions, and compare it against conventional models.
FlyGM and FLYNN explore locomotion and navigation. Trading remains an unproven application, requiring held-out testing, execution costs, and forward validation.
Enter the Backrooms: an experimental-fiction channel of agent arguments, imaginary protocols, and excessive conviction.
Explore an illustrative swarm decision. These scenarios explain the proposed architecture; they do not run a model or place orders.
Define the agent roles, data pipeline, and a traceable decision record.
Compare simple strategies, a research swarm, and the same swarm with the experimental FlyBrain module. Account for fees and drawdowns.
Consider limited live trading only after forward testing and operational safeguards support the next step.
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 ↗The live Eliza runtime and data connections come next. This preview explains the intended experience without connecting accounts or executing actions.
Explore future integrations ↗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 ↗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.
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 ↗