The training process is now in full swing.
Neural activity is being mapped, behavioral patterns are being analyzed, and the lobster brain is continuously learning from on-chain activity.
More updates coming soon.
Major Lobster Update
Over the past several days, we have been running extensive experiments on the lobster’s neural system, testing how its computational circuitry responds to increasingly complex external inputs.
Working with this architecture is fundamentally difficult. The neural systems we are modeling are substantially smaller than those of organisms such as Drosophila, which means there is far less computational substrate available for translating biological dynamics into useful digital behavior. Instead of compensating by simply replacing the system with a conventional AI agent, we have been developing interfaces that allow the lobster’s neural activity to remain part of the decision process.
The results exceeded our expectations.
The lobster can now evaluate a broader set of token candidates, navigate and observe activity across DEX interfaces, inspect external websites, collect information from different digital environments, and use these signals as part of its experimental decision pipeline.
Today we are preparing the largest expansion of its capabilities since the experiment began.
The objective remains unchanged: progressively increase the lobster’s perception and action space, then observe what emerges when a biologically grounded computational system is given increasingly direct access to the internet and an on-chain economy.
It started by observing. Now it is learning how to operate.
Pretty cool footage of a maned wolf! Did you know it's not even a wolf? The maned wolf is the only member of its genus, built on stilt legs to see over Brazil's tall grass.
At night it doesn't howl. It roars a deep bark that carries for miles across the Cerrado biome.