44: Naval Ravikant (part 1) | End Games

It became very clear to me, probably about a decade ago, that we’re just on a trend line where on a long enough time scale I can 3D print a nuclear weapon in my house. On a long enough time scale, I can take a synthetic biology lab in my room and create a weaponized virus that combines the most virulent features of smallpox and the longevity of AIDS and spreads like the flu.

Drones are essentially guided bullets. It’s a bullet with a tracking system... I think they easily end commercial aviation as we know it

At any given time on the planet, there are a few people – and one is too many – if you gave them a button that would end the planet, they would press that button

On a long enough time scale, I think commercial air travel will come to an end because weaponized drones will be too prevalent

There’s no free lunch here. The act of creating the technology required to save us will first create the technology that will destroy us.

Just look at how we treat every other creature on Earth to get an idea of how AI will treat us

What if it turns out that all the things that we do to make day-to-day life a little more pleasurable, ranging from taking SSRIs to using Facebook, make your average baseline pleasure line slightly higher, but actually increase the blow-up risk.

I don’t think general AI is high up on the list of issues, not because it’s impossible, but because it’s going to take a very long time. It's improbable in the next 50-100 years.

If you look at how much of this AI stuff is being peddled today by people who did not have "AI" as a title on their business card five or ten years ago, it tells you that there's no such thing as a true AI expert who we should all be following and listening to

It’s been getting easier and easier to destroy something thanks to technology, and it’s getting harder and harder to defend against

If you were a Bond villain trying to take over the world, the three technologies that you would look at are synthetic biology, hidden nuclear weapons, and hunter-killer suicide drones that are miniaturized…like nano drones that use pheromone tracking to find their targets.

The basic idea is that if we cross certain technological red lines, there’s some non-zero chance that a super artificial intelligence could arise, seize control, and imperil humanity

Saying you’re an “AI expert” just means your identity is tied up in the technology, making it hard to tolerate any criticism of it

Physical privacy is dead

When suicidal mass murderers really go all in, technology is the force multiplier.... When someone goes nuts with a knife versus a machine gun, the person with the more powerful technology is going to kill more people.

It’s just the nature of the nature of the Faustian bargain with technology that we get so much more power over our natural environment. That power includes the ability to destroy things and the destructive powers arrive long before the protective powers do.

It's going to cost a hundred billion dollars plus to develop an AI

At some level, it’s just hard to image how we tame nature without the ability to end nature

Creating destructive power is actually a lot easier (compared to protective powers) - “From the moment we first split the atom to the first nuclear bomb was a much shorter trip compared to the first nuclear power plant”

Even though they signed the Genova Convention, I will bet you there are multiple working bioweapons labs in the world today