This paper introduces Autogenesis, a self-evolving agent
Autogenesis的引入代表了智能体领域的一项创新,它可能对智能体的未来发展方向产生重大影响。
This paper introduces Autogenesis, a self-evolving agent
Autogenesis的引入代表了智能体领域的一项创新,它可能对智能体的未来发展方向产生重大影响。
Summary This design of the Cambria-Automerge integration has several useful properties: Multi-schema: The document can be read and written in any schema at any time. A document has no single “canonical” schema—just a log of writes from many schemas. Self-evolving: The logic for evolving formats is contained within the document itself. An older client can read newer versions without upgrading its code, since it can retrieve new lenses from the document itself. Future-proof: Edits to the document are resilient to the addition of future lenses. Since no evolution is done on write, old changes can be evolved using lenses that didn’t even exist at the time of the original change. One area for future research is further exploring the interaction between the guarantees provided by lenses and CRDTs. For example, Automerge has different logic for handling concurrent edits on a scalar value and an array. But what happens if the concurrent edits are applied to an array in one schema, and a scalar value in another?
“Doing what scientists do” is not doing science, and won’t deliver—just as “doing what a ground crew does” doesn’t bring planes. It’s just going through the motions.