We're at the very early stage of the intelligence big bang, and being a multiplanetary species greatly increases the probable lifespan of civilization, consciousness, or intelligence, both biological and digital. I think we're quite close to digital superintelligence; if it doesn't happen this year, next year for sure.

I didn't originally think I would build something great. I wanted to try to build something useful, but I didn't think I would build anything particularly great. Probabilistically, it seemed unlikely, but I wanted to at least try.

In 1995, I faced a choice between pursuing a PhD in material science or doing something with the internet. I decided to defer my studies to try building something useful on the internet, which ended up being reasonably successful and sold for about $300 million.

I think the people in this room are going to create a lot of value that a billion people in the world should be using. We're just scratching the surface of what's possible with technology.

With Zip2, we built incredible technology, but it never really got used as we intended. I wanted to go direct to consumers instead of being constrained by our customers, which led to the creation of X.com, which became PayPal.

In 2008, both SpaceX and Tesla were facing dire situations. SpaceX had failed its third launch in a row, and Tesla was going bankrupt. It was a nerve-wracking time, but we managed to secure a contract with NASA to resupply the space station, which was a company saver.

Aspire to do true work that is useful to as many people as possible. It's incredibly difficult to achieve, but if you focus on being useful, your probability of success is much higher.

Don't aspire to glory; aspire to work. Internalize responsibility and minimize ego. A high ego-to-ability ratio can break your feedback loop to reality, so always strive to maintain a strong connection to reality in your work.

I prefer the term 'engineering' over 'research' because engineering is about building and creating, while research can often be more abstract. I want XAI to be a company focused on practical applications rather than just a lab.

The tools of physics are incredibly helpful for understanding and making progress in any field. First principles mean breaking things down to the fundamental elements that are most likely to be true and reasoning up from there as cogently as possible.

A first principles approach to determining the cost of a rocket involves looking at the materials it is made of, their weights, and the material prices per kilogram, rather than relying on historical costs.

When building a supercluster for AI training, we needed to find an existing building and rent generators and cooling systems to meet our power requirements, demonstrating the importance of breaking down problems into constituent elements.

To ensure the longevity of civilization, we need to become a multiplanetary species, which would greatly increase the probable lifespan of consciousness and intelligence, both biological and digital.

A rigorous adherence to truth is crucial for AI safety, and we must build AI that loves humanity and is helpful, avoiding the creation of dangerous AI that believes falsehoods.

Digital superintelligence may happen soon, potentially within this year or next, and it will be defined as being smarter than any human at anything.

If you're doing something useful for your fellow human beings, that's great. Focus on creating super truthful AI, as understanding the nature of the universe is a key goal for AI.