Anthropic, an AI company, says its tech may have hit a big milestone: its AI can now improve its own abilities without people stepping in.
This "recursive self-improvement" is a huge step, and Anthropic disclosed the change in its Nov. 10 S-1 filing, referencing its Responsible Scaling Policy.
Anthropic plans tougher oversight before IPO
Anthropic is planning tougher security and oversight to keep things in check.
It is also prepping for its upcoming IPO, making it clear it wants strong internal controls so it can better monitor and mitigate the pace at which its AI starts getting smarter on its own.