Anthropic’s AI is writing its own blog — with human oversight

The anthropological AI gave a blog.

A week ago, the ethnic was turned on mute Claud explainsA new page on its website that most companies have been produced by the AI ​​Model Family Claud. Posted by posts on technical issues related to different clodes (such as “simplify complex codebuses with claddies”), the blog is identified as the type showcase for Claud’s writing skills.

It is not clear how much the Cloud’s raw writing explains the posts on the clode. According to a spokesperson, the blog is supervised by anthropological “subject experts and editorial teams”, who “enhances” of Cloud’s draft “insight, practical example and” improved ” […] Relevant knowledge. “

“This is not just a vanilla clod output – the editorial process needs human skills and goes through repetition,” the spokesperson said. “From a technical point of view, Cloud explains an associate approach where Claud [creates] Educational materials and our team reviews, refine and increase it ”

Cloud explains’ None of this is clear from the homepage, which carries the description, “Welcome to the small corner of the ethnographic universe where you write about each subject under the sun.” Someone can easily think that the copy of the Claud blog is responsible for the end to the end.

Anthropic states that it explains Claud as “starting with educational resources” as “explains how human skills and AI power can work together” as an exhibition.

“Cloud explains that parties are an initial example of how parties can use AI to raise their jobs and provide their users more prices.” “Instead of replacing human skills, we see that AI can widen on how the subject experts can achieve what can achieve […] We plan to cover issues from creative writing to data analysis to business strategies. “

Anthropic test with AI-exposed copy, which came just a few months after the rivals Openai made it a model that made a model Creative writingFrom the very beginning it can be clearly stated. Meta’s Mark Zuckerberg says he recently wants to develop an end -to -end AI advertising equipment and OpenAI chief executive officer Sam Altman Forecast AI that AI can handle 95% of the day “marketers, strategists and creative professionals.”

Somewhere else, publishers have driven AI newsworthing equipment to enhance productivity and have reduced the need for recruitment in some cases. Ganet has been AggressiveRolling the briefs at the bottom of the AI-exposed sports rackaps and the title. Bloomberg AI-exposed summons added to the top of the article in April. And Business Insider, which has given 21% of its employees last week, contains Push To assist the authors in AI equipment.

Even inherited outlets are investing in AI, or making the least obscure overturn so they can. New York is Times Encourage staff Washington Post to use AI to give edit, title and even questions to ask for asking during interview Has said to be developed An “AI-driven Story Editor” called Embar.

Yet many of these efforts have not been well, basically because AI is a tendency to make things with confidence today. Business Insider was forced to apologize to the staff after recommending a book that does not seem to exist but may be instead produced by AI, According to Semfour. Had to do blomberg The correct dozen AI-exposed summons of articles. G/O OCIR ERROR-Shopped AI-Written Features Published against editors’ greeting, MockeryThe

Ethnographic spokesperson mentions that the company is still hiring on marketing, content and editorial and “many other areas that are involved in writing”, despite the company’s drowning in the AI ​​-driven blog draft. Take it for what you want.