Ilya Sutskever launches Safe Superintelligence Inc.

Background

Ilya Sutskever had left OpenAI on 15 May 2024 ^[1]. He announced the new company on X on 19 June 2024 ^[1].

What happened

On 19 June 2024 Ilya Sutskever announced Safe Superintelligence Inc. on X ^[1]. His co-founders were Daniel Gross and Daniel Levy ^[1]. The company's single stated goal is building safe superintelligence, meaning AI that is far more capable than people, with safety treated as the central requirement ^[2].

Superintelligence is a term used for AI systems that would exceed human ability across most areas of work. The company's name puts safety in the first position, and its statement of purpose gives it no other stated goal ^[2].

Early setup

The company set up offices in Palo Alto and Tel Aviv ^[2].

Early funding

In September 2024 the company raised $1 billion at a $5 billion valuation ^[1]. The valuation is the estimated worth of the company at the time of the round, not a measure of revenue, since the company had no product at that time.

What AI safety research means

AI safety research studies how to make advanced AI systems behave as their designers intend, and how to prevent serious harm from them. The term covers technical work on how models learn and fail, as well as questions about how such systems should be governed. SSI's name places this goal at the front, and its public statement gives no other aim ^[2].

Why a single goal matters

A company with one stated goal makes its direction easier to describe to outside readers. It also means that its funding, hiring and research plans can be measured against that one aim. Investors and researchers can judge the company by whether its later work serves the goal it set out at the start.

Why it matters

The founding of SSI is notable because a researcher who had just left OpenAI started a company with a single stated goal and raised a large sum before it had any product.