Keep pulling the thread on Sebastian Malaby.
Reid Hoffman offered $1 billion to Demis Hassabis to finance a potential spin-out of DeepMind from Google.
Demis Hassabis has shifted his view from believing in a 'singleton scenario' for AGI development to now seeing AI safety as a collective action problem that only governments can solve.
Elon Musk co-founded OpenAI at the end of 2015, creating a competitor to DeepMind shortly after hosting DeepMind's AI safety summit.
Effective global AI safety regulation will ultimately require collaboration between the United States and China.
The UK's AI Safety Institute has successfully identified vulnerabilities in AI systems that the private labs which created them had failed to find.
In 2026, Anthropic was the first to market with its coding agent, Claude Code, followed closely by OpenAI with its Codex model.
DeepMind had a secret plan, codenamed 'Project Mario', to potentially spin out from its parent company, Google.
Demis Hassabis decided against spinning out DeepMind from Google to avoid a potential legal battle and to maintain access to Google's vast compute resources.
Demis Hassabis was awarded a Nobel Prize for his work on AlphaFold, a model that predicts protein folding.
The US government shifted from a laissez-faire position on AI to a more regulatory stance after a model from Anthropic, called Mythos, raised public concern.
Ilya Sutskever of OpenAI recognized the potential of the transformer architecture on the day the paper was released and immediately initiated a language model project with Alec Radford.
In a January New York Times article, Sebastian Malaby predicted that OpenAI had a 50% chance of going bankrupt or selling at a discount by the summer of the following year.