Project Prometheus: How Jeff Bezos Plans To Use AI Beyond Chatbots

Amazon co-founder Jeff Bezos, who stepped down as company CEO in 2021, is returning to an operational role in tech. He will be the co-CEO of a new artificial intelligence startup called Project Prometheus.

The company, backed by $6.2 billion in funding, marks one of the largest early-stage bets on AI.

Bezos will share the role with Vik Bajaj, a physicist-chemist with a background at Google X and Alphabet's life-sciences unit, Verily. The venture has already hired around 100 employees, recruiting talent from top AI firms, including OpenAI, Google DeepMind, and Meta.

What is Project Prometheus?

Bezos has a long-term interest in space travel through his company Blue Origin, and Project Prometheus is working on technology that could eventually support his space ambitions, according to The NY Times.

Project Prometheus is part of a new trend of AI companies using artificial intelligence for real-world, hands-on tasks rather than just software.

The company is building artificial intelligence (AI) tools that can help engineers and manufacturers produce products in a number of fields, including computers, aerospace, and automobiles, which also connects to Jeff Bezos' interest in space travel.

“Is it machine learning? Is it agents? Or is it generative AI? They're sort of connected,” said Bajaj.

"But people talk about them as very different blocks. Your ability to be either influenced, supported or disintermediated by any of these technologies is what's causing the biggest concern or opportunity for businesses," he added, according to a tech website.

How will Project Prometheus work?

Unlike AI chatbots like OpenAI's ChatGPT, which learn by analyzing vast amounts of digital content from sources such as Wikipedia, news articles, and books, Project Prometheus aims to build AI that can interact with the physical world.

While large language models can mimic human language, solve math problems, or even write computer code, Prometheus' AI will learn from real experiments, data, and machinery, allowing it to make scientific discoveries and help in engineering and manufacturing.

The venture aims to push AI by enabling it to learn in more complex ways than chatbots. From cars to spacecraft, the venture hopes to revolutionise how products are designed, tested and built.