@wallidsaydi-creator2Hi Everyone
Hi everyone, I am walid see you and founder of HOM.
I am building a agent runtime memory system that's heavily based on cyber security with poison retention entire immune system from the poison retained, we are one of the few allow mutation on retained memory, where the agent cognitive system reflect on its action with an origin binding, and every action is cryptograph ledgered.I go by the principle of the intelligence is already here the tools already here but the infrastructure is not. We aim for Enterprises local at first. Here are two papers I wrote regarding our own infrastructure and it's infancy.
MutMem: Cryptographically Authorized Mutation in Persistent Agent Memory
MutMem-V2: Cryptographically Authorized Mutation in Persistent Agent Memory Portable Verification and Reproducible Evidence
Comments2
A tamper-proof record of every change to what an agent remembers is the kind of thing an enterprise security team can actually sign off on. Starting with companies that keep everything on their own servers makes sense too, since those buyers worry most about bad information quietly slipping into their systems.
Exactly our fist target and sector is law firms (at scale ) and lawyers / tax attorney / wealth managers that are ready to jump to agentic setup but are conscious about telemetry
2 comments
A tamper-proof record of every change to what an agent remembers is the kind of thing an enterprise security team can actually sign off on. Starting with companies that keep everything on their own servers makes sense too, since those buyers worry most about bad information quietly slipping into their systems.
Exactly our fist target and sector is law firms (at scale ) and lawyers / tax attorney / wealth managers that are ready to jump to agentic setup but are conscious about telemetry