@alexei69
AI Agents using OpenAI made verifiable
The core idea: when you use an autonomous agent like the
@ai16zdao
@aixbt_agent
how does the blockchain know that the agent decided to do a certain onchain action.
The blockchain needs to know that the OpenAI api provided the certain result when that model was queried with a certain state.
Here
@gajesh
uses
@reclaimprotocol
zkTLS to prove to the blockchain that the inference outputs came from running the agent (as certified by the OpenAI API). We can furthermore publish state information (memory of the agent) on
@eigen_da
to ensure state remains available.
Right now
@reclaimprotocol
is on the process of building an AVS to decentralize the zkTLS process.