SCOOP: Anthropic wants to scale AI by training whole teams of Claudes
Two research listings describe teaching Claudes to cooperate and preparing the company's "largest-ever agent run," with salaries reaching $850,000.
By Ryan Merket · Published
Primary source: Anthropic
Why it matters
If Anthropic can train Claudes to delegate, communicate and correct one another, coordination becomes part of the model customers buy rather than work handled mainly by surrounding agent software. That could change Claude's cost per completed task and shift value from agent frameworks toward the model provider. The listings prove the research ambition. They do not prove that a trained team can outperform a larger single model at an acceptable price.
Two research listings describe teaching Claudes to cooperate and preparing the company's "largest-ever agent run," with salaries reaching $850,000.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.