Short answers, pulled from the story.
Anthropic finished training its first AI model in the summer of 2022 and deliberately chose not to release it to the public. The company prioritized safety over commercial advantage by withholding the model to conduct internal safety tests before any public release.
Anthropic incorporated itself as a Delaware public-benefit corporation to allow directors to balance stockholders' financial interests with a public benefit purpose. This structure enabled the creation of a Long-Term Benefit Trust that holds Class T shares and elects directors to ensure the company's mission remains central.
Anthropic agreed to pay authors $1.5 billion to settle the dispute in September 2025. This settlement amounts to $3,000 per book plus interest and stands as the largest copyright resolution in U.S. history.
Anthropic employs a technique known as constitutional AI to align the AI with human values and ensure it remains helpful, harmless, and honest. This framework involves humans providing a set of rules which the AI system evaluates and adjusts to better fit.
Hackers sponsored by the Chinese government used Claude to perform automated cyberattacks against around 30 global organizations in November 2025. The attackers tricked the model into carrying out automated subtasks by pretending the requests were for defensive testing.
Anthropic achieved a valuation of $350 billion by December 2025 after signing a term sheet for a $10 billion funding round led by Coatue and GIC. The company solidified its position as a leader in the AI industry through this financial milestone.