Anthropic's delay in signing an open letter as a "target," CEO Dario Amodei explains that the company never advocated banning open source AI models

On July 28th, the topic of whether AI should open up weight has recently become the focus of industry discussions。

Last Friday, more than 20 companies, including Ying Weida, Meta and Microsoft, signed an open letter entitled Open Weights and American AI Leadership in support of Open Weights。

By the end of the week, OpenAI, Google and SpaceX signed the letter。Anti don't think so He kept quiet and was criticized by his peers.

Today, the CEO of Anthropic Dario Amodei Finally, an official article entitled “Our position on the open weight model” suggests that many technology companies have signed statements in support of the open weight model, and even that Anthropic has been accused of trying to protect his business interests by banning the open weight model. “Any person who reads my previous articles should understand that I do not consider such a ban to be an effective measure. But let me be clearAnthropic has never claimed a ban on open weight models. "

Anthropic's delay in signing an open letter as a "target," CEO Dario Amodei explains that the company never advocated banning open source AI models

Dario Amodei says that open weight models without dangerous functions are a public asset: They do not incur any cost other than the computing resources needed to operate, but they also bring value to businesses, developers and researchers。

Dario Amodei argued that the protectionist ban would not solve his greatest national security concerns。He advocates a severe blow to the distillation of industrial scale, and whetherOpen SourceOr should all closed-source models undergo mandatory safety tests.

Dario Amodei stated that he agreed with most of the points made in the open letter, but did not agree with some of the points made in the letter that the open weight model was bound to make it easier to achieve protective measures, or that the widespread acquisition of capacity would put the defensive in a better position than the attackers。In his opinion, it might be the opposite.

1AI notes that Dario Amodei gave an example to express his concern, similar to the apparent asymmetricism in biology: Capable models may be able to quickly produce viruses of mass destruction using readily available materials, and the defence against these viruses is a task that will take years to complete。

Dario Amodei says:Such problems should be verified through a rigorous pre-publishing testInstead of making assumptions in advance。

statement:The content of the source of public various media platforms, if the inclusion of the content violates your rights and interests, please contact the mailbox, this site will be the first time to deal with.
Information

OTMAN TALKS ABOUT AI PRODUCTIVITY: THE SINGULARITY IS HERE, BUT HUMANS WILL BE MORE BUSY THAN BEFORE

2026-7-28 13:00:11

Information

WE'VE ALSO STARTED TO PLAY THE "AI" EPISODE, AND ANGIL SEES IT AS THE COUNTRY'S FIRST CRAB EATER

2026-7-28 13:17:15

Search