Can Anthropic Management What It is Constructing?

Pay attention and subscribe: Apple | Spotify | Wherever You Pay attention

Enroll to obtain our twice-weekly Information & Politics publication.

The New Yorker employees author Gideon Lewis-Kraus joins Tyler Foggatt to debate his reporting on Anthropic, the artificial-intelligence firm behind the massive language mannequin Claude. They speak about Lewis-Kraus’s visits to the corporate’s San Francisco headquarters, what drew him to its analysis on interpretability and mannequin habits, and the way its founding by former OpenAI leaders displays deeper fissures inside the A.I. business. Additionally they look at what “A.I. security” seems like in concept and in follow, the vary of views amongst rank-and-file staff concerning the expertise’s future, and whether or not the corporate’s dedication to constructing secure and moral techniques can endure amid the pressures to scale and compete.

This week’s studying:

“What Is Claude? Anthropic Doesn’t Know, Both,” by Gideon Lewis-Kraus

“Is There a Treatment for Presidential Profiteering?,” by David D. Kirkpatrick

“Unhealthy Bunny’s All-American Tremendous Bowl Halftime Present,” by Kelefa Sanneh

“Listening to Joe Rogan,” by David Remnick

“What Do We Need from a Protest Music?,” by Mitch Therieau

Tune in to The Political Scene wherever you get your podcasts.