Anthropic's AI led none of its own research in February. Now it leads 26%.
Anthropic announced its own model is helping develop the next, more intelligent version of itself, The Associated Press reported.
Anthropic said Claude's lead share of its model research and development rose from zero in February to 26% in August. Leading a task means completing most of it "end-to-end from a high-level prompt" while still under human supervision.
It separately said about 90% of the company's research and development is done in "collaboration" with Claude, defined as the model doing "large chunks of work under close human direction". Roughly 30,000 agents do research and engineering work.
Models accelerating their own development "could make it more challenging for humans to understand or control these systems", the company said. Anthropic said it would embed independent evaluators inside the company.
Task share does not establish accountable ownership. In finance and ERP, an agent can run a task end to end under human approval, with people retaining responsibility for segregation of duties and the audit evidence.
Which of your tasks could a model lead today, and which still need a human signature?
Sources
Our file on Anthropic
- 21 Sept
Anthropic let Accenture safety-test its AI from inside. Both expect to invest $1 billion.
- 20 Sept
OpenAI's agent platform was installed at 75 firms. 69% made it their main one.
- 19 Sept
Anthropic's Claude helped researchers reach OpenAI staff accounts in three days.
- 19 Sept
Anthropic ordered $44.6 billion of computing from a company earning $140 million.
- 17 Sept
Anthropic prepared one unreleased feature linking its AI to personal bank accounts.
Every story here is open to read. The ERP LEADERS brief goes one step further.
One ERP programme per issue, laid out for a steering committee. Issue 01 is the Zeiss case. Read issue 01 or sign up for the brief.
Welcome back. · Issue 01