Dr. Radhika Dirks’ mission is embodied in the organization she’s been building called XLabs. Simply put, the mission is to unleash Moonshots and change the perception of what people think is possible. “Moonshots” are the kind of goals and projects that invent a new era… the kind of projects that will usher in the intelligence age.
- 3:00 What is a “Moonshot Factory?”
- 9:00 Inventing the AI mothership, Seldn
- 15:00 Radhika’s story
- 21:00 How does AI work?
- 28:00 Tech that can amplify humans
- 36:00 How can we understand the exponential?
- 50:00 The 2 steps to begin designing your moonshot.
The Main Players in the Intelligence Age
XLabs is “a place that can harness the genius of people like Tesla in the 20th century.” His moonshot was a quesiton… how do you capture lightning? This question spurred the invention of electricity as we know it, and prepared the way for Edison after him. The difference between Tesla and Edison though was that no one took Tesla seriously. XLabs intends to prevent that disbelief.
Radhika saw the kind of business and funding going on in Silicon Valley didn’t particularly favor moonshot ideas. Only a specific type of proven-concept startups are favored in the tech hubs of the world, she observed, and so the idea for XLabs was born to give room for unconventional innovation in tech hubs like Silicon Valley.
XLabs spins out 3 companies a year out of the exploration that happens in their research and team-building facilities.
The AI Mothership
9:00 Radhika talks about the gap her team has been addressing in AI… why hasn’t AI totally disrupted business and technology yet? How can we create a company that works like an AI works?
The team started asking questions about the nature of AI to mimic its function in building a company. The company they’re building now learns and grows naturally as their research grows, and therefore informs how AI works. This company became Seldn, which could predict anything from labor strikes to currency changes. Because it was learning from financial models and social occurrences, the company surpassed its 3-year goals in 6 months.
Then… they started looking at what would happen if they opened Seldn’s learning across genetics, culture, media, and geological changes.
15:00 Radhika decided to come to the US at 17 years old, and “I was obsessed with learning about the Universe. How does society work? How does this instrument work?” She wanted to become a physicist and engineer, but she wanted the exposure to research and physics that was available in the States.
She attended Purdue University for physics but wanted to learn engineering to build things people can use. She studied nanotechnology for her masters, and then realized she wanted to learn to build revolutionary ideas, not incremental steps. So, she studied quantum computing for her Ph.D.
Midway through graduate school, Radhika realized that the kinds of people who can build these revolutionary technologies are not the ones who know how to build a business. So, she built a basic business to get the experience to build her first “real” company with a co-founder after grad school.
Then, Shell Oil brought her on, she founded a VC group inside of Shell and eventually moved to San Fransisco to begin building Seldn.
How Does AI Work?
21:00 Radhika distinguishes that AI is a specific kind of software that can mimic the intelligence of a human being. That means it can learn according to an objective, and alter it’s functioning based on what it learns.
Check out this video to see how Dr. Dirks describes AI.
She talks about areas of AI to mimic human functions. For example, “computer vision” to recognize and identify images, and language technology to mimic human intelligence in language. This kind is called “natural language processing.”
The moonshot for AI is to supersede mere automation of things humans can do… Radhika is interested in inventing technology that can do things that humans can’t even do very well. Like financial predictions, causation of societal problems, and creating systemic solutions.
- Can we create intelligences to find the origin of diseases?
- Can we find drug market sources and address them?
- Can we increase safety in volatile cities or countries?
- How can we use AI to amplify humans?
28:00 Here’s the matrix Radhika describes to rank different types of AI:
|Things that come naturally to humans
||Things that are difficult for humans
|Things that are easy for computers today
||Things that are too complex for computers
How can we understand the exponential?
36:00 Radhika introduces a though problem to demonstrate how counter-intuitive exponential growth is.
“Imagine you are filling a glass with the bacteria penicillin, which doubles every second. You start at 8:00 am. It will be full by 12:00 noon.
How full is the glass one second before noon? It’s half full.
5 seconds before 12:00 noon, the glass is only 3% full.”
So, we’re often tempted to cut off resources or measure our growth and very discouraged by what we find. But this is how moonshot-exponential growth is happening.
Non-linear frameworks for allowing exponential growth
Most companies and projects build in a linear way, with deadlines and goals. But exponential breakthroughs don’t occur that way. They often arise by accident, in solving a different problem or exploring a solution with a particular constraint.
- Leaning into interesting-ness or pursuing novelty
- Allowing total freedom to explore, and build solutions on top of one another
So, in order to discover something new, your team will need a lot of freedom, a few purposefully chosen constraints, and a mindset of unattached curiosity… what’s possible?
How do we prepare for change and contribute to the growth as entrepreneurs?
50:00 Radhika gives her top 2 tips for how to make a difference in bringing about the intelligence age:
“More social entrepreneurs need to start adopting this mindset. The mindset that our lives are completely going to change… and starting with creating meaning instead of creating money.
Think about – if you had all the data you could possibly need, what would you build? Start designing it. What would you create in the world with that mindset? Give yourself whatever time frame. Now, come back and ask what is possible for that vision today.”
Then, she says, you will start designing the product that will automatically collect the data you’ll need. Then, you’ll design the incentive for people to share the data that you need to fulfill that vision.
“Here’s what NOT to do… don’t try to become an expert in AI! Don’t compensate for your weaknesses. Build your strengths, and build a team.”
- The exponential growth model… how long will it take for our product to be ready? Well, how do I get a specialist on my team so they can tell me how fast things are doubling? Am I 5 seconds from my 12:00 goal of a filled penicillin cup, or 4 hours away?
- Do you have someone on your team to understand your timeframe, and stay on the leading edge?