Anthropic CEO Dario Amodei has one of the biggest contradictions in tech: he genuinely believes AI could end the world, and yet he founded one of the most aggressive frontier AI labs in the business.
Even back at OpenAI, he was intense. Sensitive memos never touched Google Docs—he wrote them on an air-gapped computer at home and printed them out for colleagues. Some say he refused to travel to China entirely, worried he'd be kidnapped.
Before GPT-3 even started training, he worried it might already be close to AGI. That's why Dario's safety team got Microsoft to...