Ex-Anthropic engineer Jacob Coxon doubts Claude 'solved' humor claim
Technology
Jacob Coxon, who used to work at AI company Anthropic, shared that their model Claude was claimed to have "solved" humor in an internal Slack channel.
But Coxon wasn't buying it. He stayed skeptical about whether any AI can really get what makes people laugh.
Coxon says Anthropic teases model consciousness
Talking to New York Magazine, Coxon described Anthropic's vibe as sometimes a bit groupthink-y, which left him feeling out of place.
He also said Anthropic openly teases the possibility that AI models could already be conscious, though he said Anthropic being described as a cult was an "extreme" view.
His story adds to the bigger conversation: can AI ever truly understand human quirks like humor?