According to Cointelegraph, Anthropic CEO Dario Amodei discussed the future of AI, including military uses and replacement of human workers, on the Econ 102 podcast.
Amodei described the company’s internal efforts to develop efficient hierarchical structures to accomplish tasks through networks of AI models. Current AI systems such as ChatGPT and Claude cost tens of billions of dollars to develop independently.
Amodei mentioned that the future may involve "big models coordinating small models," where the big model creates hundreds of smaller, faster, more efficient models to perform tasks. This is similar to a typical enterprise infrastructure.
There are multiple base models under the core Claude model, focusing on areas such as mathematics, programming, and sentiment analysis. The next layer is the model dedicated to completing a specific task, and the bottom layer is a one-off model for short-term use.
The benefit of this design is that users at all levels can access the specific features they need through a single interface.
