From the article:

ChatGPT-maker OpenAI and its main rival, Anthropic, which provides the Claude chatbot, endorsed a petition released Tuesday that had been signed by hundreds of employees at the two companies, as well as Google, Meta and other leading companies working on AI.

The letter warned that companies could be close to automating the work of AI research, a development that it said could lead the technology to careen out of control.

“We request that the U.S. government support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development,” the petition said. (The Washington Post has a content partnership with OpenAI.)

Coordinating international agreement to limit AI development would be a gargantuan diplomatic task that has been likened to implementing a nuclear weapons control treaty for software that can be freely distributed through the internet.

[...]

The new petition is different because it has won corporate endorsements for a more specific government intervention on the industry, albeit at an unknown point in the future.

Its signatories also include senior executives from OpenAI, Meta and Google, companies that have generally been much more opposed to strict AI regulation than Anthropic.

The way I’m thinking about it is that this is an attempt to agree on a new Moore’s law.