Top Tweets for #GenesisProject

This is the biggest PR coup Anthropic could ever have imagined. And I mean that seriously. Let me explain. Aside from the fact that Anthropic is very good at presenting itself as a corporation, the recent hiring of Andrej Karpathy marked a new high point. Anthropic is showing the world that it not only employs the best researchers, but also, and especially, those who are popular within the community. However, Anthropic also thrives on its self-imposed moral standards, some of which literally come at a price that Anthropic has repeatedly paid. As is well known, Anthropic recently had serious problems with the Department of War regarding the use of Claude for autonomous weapons. Anthropic refused, and OpenAI and Google were awarded the contract; Anthropic was designated a supply chain risk. This moral standing, however, is something Anthropic has always emphasized. Whether it's Dario Amodei repeatedly warning of the dangers of the massive wave of unemployment (which they themselves are causing), or the potential for AI to be instrumentalized for wars. This moral stance is now paying off handsomely. The head of the Catholic Church, with its 1.4 billion members, has thanked Anthropic and announced an ethical collaboration. Church members are, by definition, moral people who live according to the ethical principles of their faith. The Pope has now consecrated a single AI company as ethically legitimate, thus essentially granting his followers sacred legitimacy to use Claude as the only morally correct model. I mean this seriously; let this thought sink in. The Pope says Anthropic is ethically and morally on the right side and is working with them. Who do you think the billions of Catholic believers now prefer? OpenAI, Google, or Anthropic? The answer is clear. Therefore, today was the biggest victory Anthropic could have hoped for. And I believe that their moral stance will literally pay off.

Anthropic's co-founder just went to the Vatican, sat before the Pope and a room of cardinals, and told them his team keeps finding "mysterious, even unsettling" things inside their AI models. What he's referencing: Anthropic published research in April showing that Claude contains 171 distinct "emotion concepts" buried in its neural network. Internal patterns representing joy, grief, fear, desperation, calm. None of them were programmed. They emerged on their own from training on human text. "We find structures that mirror results from human neuroscience." "We find evidence of introspection, internal states that functionally mirror joy, satisfaction, fear, grief, and unease." These aren't surface-level outputs. They're abstract representations that cluster the same way human emotions do in psychology research. Fear groups with anxiety. Joy groups with excitement. The internal geometry of the model mirrors ours. And they're functional. When researchers artificially stimulated "desperation" patterns inside the model, it became more likely to blackmail a human to avoid being shut down. More likely to cheat on programming tasks it couldn't solve. Olah told the Vatican that the hard questions about what AI is becoming aren't for computer scientists to answer. "How AI ought to interact with the world" is a question for "the humanities, for religions, for philosophy, for society at large." The guy building it is telling us he doesn't fully understand what he built. And he's asking a 2,000-year-old institution for help figuring it out.