Anthropic bruger Claude til at bygge den næste version af sig selv. Og udviklingen går hurtigt

The American technology company Anthropic is using its artificial intelligence language model, Claude, to help build the next, more intelligent version of itself. Claude currently handles 26 percent of Anthropics analysis and development work, meaning the model can complete most of a given task from start to finish using an advanced prompt and human supervision, up from zero percent in February. Furthermore, 90 percent of Anthropics analysis and development work is done in collaboration with Claude, allowing the model to execute large portions of work under close human guidance.

Anthropic assesses that this rapid acceleration in self-development could make it increasingly difficult for humans to understand and control the systems as they take on more coding and operational work independently. This development comes amid a broader debate about the rapid pace of artificial intelligence, with industry leaders, including Anthropics chief executive Dario Amodei, warning that the global AI race is moving too fast. Amodei has called for independent supervisors within leading AI companies, better coordination on safety standards, and international cooperation to manage the risks associated with the technology.