Andrew Ng came up with a theory almost 15 years ago: What if Google, the internet giant, thought of machines like the brain and trained neural networks on mass amounts of data using computational power in an attempt to lead to artificial general intelligence?
Google bit on Ng’s line of thinking, and in 2011 he helped launch Google Brain, one of the most ambitious artificial intelligence programs at the time. Flash forward: Ng has cemented his status as one of the most respected figures in computer science.
Earlier this year, Ng’s AI Fund, which backs small teams of experts using AI to solve key problems, said it planned to raise over $120 million for its second fund. A June filing with the SEC showed that the fund had amassed $69.75 million from 13 partners.
Ng launched the AI Fund in 2018 with $175 million, and initial backers included Sequoia and Softbank Group. It’s made 43 investments to date, according to Crunchbase data.
In April, Amazon also named Ng to its corporate board of directors, a further sign the tech giant wants to advance its profile in the space. In addition to his time at Google, Ng helped build out Baidu’s AI development.
Part of Ng’s efforts have been getting more people to understand the field of artificial intelligence.
Ng has been working for years to “democratize deep learning” by teaching more than 8 million students through online courses through projects like Coursera and DeepLearning.AI. (“More than one in 1,000 persons on the planet have taken an AI course from me,” Ng tells Fast Company.)
Ng’s name has come up again and again in the past year as I’ve spoken with people in the tech world. He is not only, as Fast Company put it in 2017, “one of the most important people in AI” but an “AI superstar.”
This is all to say, Ng can be considered central to AI’s future. Now, Ng’s thinking about how agentic AI could shift the intelligence space.
Earlier this year, Ng put out a prediction through DeepLearning.AI saying that agentic AI—essentially, AI-powered agents that can autonomously solve tasks—was going to be a big focus in the tech space in the coming years as it would contribute even more to progress in AI than scaling up large language models would.
“It was a slightly controversial prediction,” he says. To be sure, agentic workflows are starting to cement their status as the next keyword.
Google, for example, recently announced a slew of agentic prototypes. Microsoft has “Copilots” to help businesses automate tasks. Anthropic’s Claude released a feature that lets users create their own AI assistant.
“Agentic workflows have really taken off and are driving meaningful business results,” Ng says. “What I’m seeing is that it feels like an early and accelerating phase of development of agentic workflows where more and more technology companies are trying to build platforms to better support agentic workflows and more and more applications are also being built using it.
This story is part of AI 20, our monthlong series of profiles spotlighting the most interesting technologists, entrepreneurs, corporate leaders, and creative thinkers shaping the world of artificial intelligence.
