wired.com
wired.com
 
Did You Know There’s a “Godmother of AI”, and She Wants You to Build Worlds?
User avatar
Curated by
cdteliot
6 min read
4,722
24
Fei-Fei Li, known as the "Godmother of AI" for her pioneering work in computer vision and image recognition, is making waves with her new startup World Labs, which aims to develop spatial intelligence models that could revolutionize how AI interacts with the physical world. Despite her discomfort with the nickname, Li's contributions to the field of artificial intelligence, including the creation of the influential ImageNet database, have earned her widespread recognition and respect in the tech industry.

Pioneering AI's Next Frontier

hai.stanford.edu
hai.stanford.edu
If you aren't familiar with Fei-Fei Li and claim to be an AI fan, it's about time you get to know who she is, what she's contributed to tech, and how she plans to literally rock everyone's world. She works at Stanford and is obsessed with cognitively inspired AI, machine learning, deep learning, robotic learning, and both cognitive and computational neuroscience
1
.
Additionally, she has published 300+ scientific articles that have been published in leading journals in computer science, engineering, science, and more
1
.
Dr. Li is also the founder of the ImageNet and ImageNet Challenge, which has led to some of the top developments in deep learning and AI thus far
2
.
In recent days, there has been a discussion of a potential AI "winter" following the explosion of AI over the past few years. This does not sit well with Dr. Li, so she has come forward with a bold idea, to launch World Labs, a new startup that will build AI systems that will be able to understand and recreate the physical world
3
.
World Labs will take a much more profound role than generative AI tools that have captivated the world, making processes easier for individuals and organizations alike. She is seeking to build AIs that live in environments of their creation. This means that they will go beyond the confines of data, which has some pretty wild implications. This translates to the development of entire digital worlds, where the objects within them can interact with other objects while also interacting with things like physics, logic, so much more. Investors are so taken with her vision that so far she has already raised $230 million for World Labs, and the startup already has a valuation of $1 billion. In Dr. Li's mind this level of investment will finally put to rest some of the doubts skeptics have had, and she can work to unlock the next frontier of AI.
engineering.stanford.edu favicon
en.wikipedia.org favicon
linkedin.com favicon
3 sources

 

What Will Go Into Building AI-Driven Worlds?

To understand what's next, we have to understand what generative AI systems like OpenAI's GPT and Google Bard are all about. They are language-based so that content can be created with ease, conversations held via chatbot, and more. While their capabilities are quite amazing, they stop short of understanding the physical world. A perfect example: you can write all day about a pineapple falling from a palm tree, but AI can't currently grasp how the fall would impact the pineapple.
1
2
Dr. Li's vision for World Labs is to completely change this dynamic. AI would go far beyond data and into the law of physics and what it means for people, animals, plants, the planets, and so much more. AI will know via learning and practical experiences that a pineapple would fall from a tree because of gravity and how it interacts with the surrounding environment.
3
4
This is called "embodied intelligence" and it leaves far behind the current AI model that we're experiencing. When this is possible, AI will be able to not only enjoy the realm of cognition, it can interact with and build real environments. Soon, it will not only describe what a pineapple is, it can pick it up, measure its weight, and experience what it feels like to hold it.
1
5
youtube.com favicon
ei.csail.mit.edu favicon
radical.vc favicon
5 sources

 

Why AI Understanding the Physical World Matter?

ft.com
The answer to this question is quite profound. Imagine you have a robot that helps you around the house. Today it requires programming that enables it to pick up things, navigate around obstacles, and carefully clean things. But, what if it could have human-like awareness of the environment it works and lives in?
1
2
Then, it will be able to understand and work with different objects and materials more sensitively, be able to easier navigate around your house, and react to changes in your home without the need to continually program the robot.
3
Can we say Jetson's anyone? What kid who watched this beloved cartoon didn't wish that Rosie the robot couldn't come live at home with them? She was intuitive, kind, and an important part of the family, after all. If this is possible, the world of robotics will be forever changed within the car industry, healthcare system, and world of logistics. Imagine them knowing how to interact with patients in the hospitals and helping nurses by being careful with patients while turning them over. Or, in surgery, they could help with precision and insight.
4
We could even move beyond robotics and move into virtual reality in incredible ways. Today's VR world is a bit clunky and awkward, and interactions can feel quite hollow. But, imagine going into an AR world where AI has prepped an environment that adapts to the real physical properties that are set in stone in the real world. Wildly enough, VR environments may become indistinguishable from the real world.
5
This could help in everything from training astronauts to soldiers, doctors, pilots, and other positions that could greatly benefit from more realistic environments. Dr. Li's World Labs may be able to create a world where virtual and physical spaces can merge in ways that we never imagined. While these implications may sound scary to some, if done with care, may be able to help us all live and breathe easier.
ncbi.nlm.nih.gov favicon
scirp.org favicon
livescience.com favicon
5 sources

 

Who’s Backing World Labs?

techcrunch.com
techcrunch.com
As mentioned earlier, World Labs has already gathered $230 million in funds
1
2
3
, which has concerned Dr. Li's critics. Some feel that there is way too much hype around the project and with the AI bubble potentially about to burst, it may be a wasted investment. Dr. Li doesn't buy limited thinking surrounding her vision, which has been enough assurance for investors like Andreessen Horowitz, NEA, Radical Ventures, Marc Benioff, Ashton Kutcher, Adobe Ventures, Databricks Ventures, NVentures, Nvidia, and more
2
3
.
It's safe to say that two factors are keeping investors all in Dr. Li's extensive resume and the wildly exciting scope of World Labs' roadmap and vision. Inventors are betting on the importance of AI moving beyond theoretical models and into practical applications that can impact us all. It's a bold gamble, but one that just may pay off.
pymnts.com favicon
nasdaq.com favicon
bloomberg.com favicon
3 sources

 

The Challenges Explained

Before we can explain what the challenges are, let's take a look at what's first on World Labs' roadmap. Well, we can't yet! They don't seem to know or are at least not telling the world yet. There are numerous questions to be answered. We can all imagine the potential, but the reality of making it all happen is what has critics questioning Dr. Li. The breakthroughs in AI models are going to have to be monumental. Not only does it require AI to evolve beyond our wildest imaginations, but robotics, hardware, and software development are going to have to evolve big time…and fast. One of the biggest hurdles is going to be AI systems understanding the causal relationship between objects. We human beings learn this principle from birth. For example, a growing child learns that softly petting a dog will make it happy, while pulling its ear will make it upset. The child quickly learns how to interact with his or her environment, be put in danger, or cause pain to a sentient being. Critics posit that teaching this sort of intelligence is going to be much harder than Dr. Li and her team want to admit.

 

What to Expect in the Near Future

For Dr. Li, her focus on World Labs goes much further than prestige. She is seeking to shape the world of AI to reflect human values while making humans' lives better in the long run. She wants it to serve society rather than to wreak havoc and cause irreversible harm
1
.
If she succeeds, AI will go far beyond data and smarts, it will understand who we are and the world we live in. In her mind, AI is only as valuable to humanity as its understanding of its challenges
2
.
So, get ready for AI to dive into uncharted territory!
waccglobal.org favicon
youtube.com favicon
2 sources
Related
How does Dr. Li's vision for AI align with current ethical concerns in the AI industry
What are the key human values that Dr. Li emphasizes in her approach to AI
How does Dr. Li plan to ensure that AI serves society rather than causing harm
What specific projects is Dr. Li working on to shape AI with human values
How does Dr. Li's background in AI and ethics influence her leadership at World Labs