Closed 0ptim closed 1 year ago
Currently, we use an Agent as the top layer, which decides on what tools it should use.
We could maybe bring cost/delay down by using embeddings for this step.
https://twitter.com/langchainai/status/1657876190261047297
Not a viable option. A lot of intelligence would be lost by this implementation.
Currently, we use an Agent as the top layer, which decides on what tools it should use.
We could maybe bring cost/delay down by using embeddings for this step.
https://twitter.com/langchainai/status/1657876190261047297