As the global breed in artificial intelligence meets, the American startup of the Anthrope has presented the last and most powerful versions of its Claude model. The company says they can write computer code for themselves and play Pokémon for much longer than their predecessors. Yuka Royer de France 24 talks to the company’s product director Mike Krieger, about guaranteeing security, fighting defenders and reducing AI’s environmental footprint.
Enhancing AI safety: Anthropic releases Claude 4 with increased protections
