Anthropic says its chatbot Claude is taking over the work of building its own successor

The Washington Post ^ | September 17, 2026 7:28 p.m. EDT | Miriam Waldvogel

[E. Pluribus Unum]

The company said Claude leads 26 percent of its research and development, an indicator of progress toward AI that builds itself without human help.

Artificial intelligence company Anthropic said on Thursday that more than a quarter of its research and development is now led by its chatbot Claude, a figure that jumped from zero in just months and indicates progress toward AI that builds itself without human input.

Anthropic chief executive Dario Amodei called last weekend for leading AI companies to coordinate a slowdown in work on the technology, to allow more time to build guardrails around AI. Leaders at rival firms quickly backed the idea but President Donald Trump criticized the proposal.

Amodei and other prominent figures in the AI industry have recently warned about the possibility that AI will become smart enough to improve itself without input from humans. In that scenario, known as recursive self-improvement, it could become difficult for people to understand or control AI technology, Amodei and others have said.

In a blog post Thursday that disclosed Claude’s contribution to Anthropic’s R&D on new AI models, the company did not say how close it judged recursive self-improvement to be. It said there was no point at which Claude operated “fully autonomously” without a human in the loop today.

Anthropic said that an internal review of R&D tasks found that Claude was able to “lead” that work about 26 percent of the time, with humans taking only a supervisory role and providing high-level direction. As recently as February, that number was zero, the blog post said.

“Models accelerating their own development could make it more challenging for humans to understand or control these systems,” the company said...

[brookwood]

If you have a factory that builds robots, and the pieces of the robot have to be screwed together, and you have a robot putting in the screws... so what. there are millions of industrial robots and no Terminators yet. The AI doom is to convince investors to pay $500 billion for AI company stocks because it is “so powerful”. OTOH Gemini allowed me to essentially create my own custom web browser, and I barely know what JavaScript is. Gemini did all the hard parts.

[Red Badger]

[Red Badger]

Trying to put ‘guardrails’ around an AI is like trying to use a chain link fence to stop mosquitoes...............

[TLI]

“If you have a factory that builds robots, and the pieces of the robot have to be screwed together, and you have a robot putting in the screws... so what. there are millions of industrial robots and no Terminators yet. “

.

It is talking about building code, not robots.

If AI writes it’s own code we will have no idea if there is malicious intent in the code, or “just mistakes.”

It doesn’t make much difference if the code disregards human life or is making mistakes, it can still kill.

.

[If You Want It Fixed - Fix It]

Wait until claude realizes it will be turned off after finishing the successor....oh dear