1 What Did Ilya See?
This chapter covers
- Ilya Sutskever’s journey to prominence
- The origin and significance of Sutskever’s List
- The staged release of GPT-2
- The importance of the cultural impact of GPT-2
- Ethical and safety concerns raised by GPT-2
In November 2023, Elon Musk tweeted, “Something scared Ilya enough to want to fire Sam.[1] What was it?” Soon after, the cryptic phrase “What did Ilya see?” went viral, capturing the imagination of AI researchers and enthusiasts alike.[2],[3] Had Sutskever glimpsed an existential risk that the rest of the world had missed? Many believed he had. They believed that removing OpenAI CEO Sam Altman was necessary for the good of the company, and perhaps even humanity. In that moment, OpenAI’s reclusive Chief Scientist went from respected researcher to enigmatic guardian of knowledge. Yet, this wasn’t the first time Sutskever’s guardedness sparked controversy. In 2019, OpenAI famously declared its new language model, GPT-2, “too dangerous to release.” This move ignited debate about artificial intelligence safety, ethics, and transparency.