Elon Musk's answers in 2017
Q1 How dangerous is AI and how do we make it go well?
We are headed towards either superintelligence or civilization ending. Intelligence will keep advancing unless civilization is put into stasis or destroyed. We have to figure out a world we would like with digital superintelligence in it. link
It comes down to two things: solving the machine-brain bandwidth constraint and democratization of AI. If we have those two things, the future will be good. link
Narrow AI such as vehicle autonomy is one thing. Deep AI, or AGI, that is much smarter than the smartest human on Earth is a dangerous situation. link
He said it is difficult to appreciate how far AI has advanced because of a double exponential: hardware capability and software talent going into AI. Predictions are almost always too conservative. link
He said people should be really concerned about AI. It is a fundamental risk to the existence of human civilization in a way that car accidents, airplane crashes, faulty drugs or bad food were not. link
He said it is the biggest risk we face as a civilization. One role of government is to ensure the public good and that dangers to the public are addressed. link
Says the biggest impediment to recognizing AI danger is overconfidence in one’s own intelligence. link
Q3 Where do you stand politically?
He thought it was worth trying to argue for sustainability on the councils. After the Paris Accord withdrawal he could not stay on: the country needs to keep its word, and the agreement was not even binding. link
Q5 Why go to Mars / what is SpaceX's mission?
Humanity should be a multi-planetary species out among the stars. That is important for long-term survival — kind of like life insurance for life collectively — and it creates a sense of adventure. link
The critical breakthrough in spaceflight is rapid, complete reusability of rockets, as with aircraft. A $60 million rocket used a thousand times is $60,000 of capital cost. He is confident you can do a fully reusable rocket. link
Becoming a multiplanet species and space-faring civilization is not inevitable, unlike a sustainable energy future; the space trend after Apollo was down. link
He said it is important to have a future that is inspiring and appealing. link
The key was how to pay for going to Mars. If they downsize the Mars vehicle so it can also do Earth-orbit activity, they might pay for it with Earth-orbit work. He thought this version had a shot at being real. link
SpaceX is about helping make life multiplanetary, continuing the dream of Apollo, and contributing to life becoming multiplanetary. He framed Americans as a nation of explorers who gave up the known for the unknown. link
Says it is high time humanity went beyond Earth: there should already be a Moon base and astronauts sent to Mars. link
Q6 When will AI surpass humans?
If it reaches a threshold where it is as smart as the smartest, most inventive human, it really could be only a matter of days before it is smarter than the sum of humanity. link
Imagine being very confident we will be visited by super-intelligent aliens in 10 years or 20 years at the most. Digital superintelligence will be like an alien. link
Q7 Will people still need to work?
Mass unemployment from automation will be a massive social challenge. Ultimately we will have to have some kind of universal basic income. He does not think we will have a choice. link
There will be fewer and fewer jobs that a robot cannot do better. Goods will be abundant and cheap. The harder challenge is how people then have meaning, because a lot of people derive meaning from employment. link
There will be a lot of job disruption because robots will be able to do everything better than us, including all of us. About 12% of jobs are transport, which will be one of the first things to go fully autonomous. link
Q11 What drives you / what is the meaning of life?
As a kid he wondered what the meaning of life was. What really matters is understanding the right questions to ask, and increasing the scope and scale of human consciousness so we can ask them. link
What drives him is trying to figure out how to make sure things are great. That is the underlying principle behind Tesla and SpaceX: accelerate the transition to sustainable energy, and help make life multiplanetary. link
Q12 What's the biggest remaining Starship problem?
Reusability only matters if it is rapid and complete, like an aircraft or a car; you do not send the plane to Boeing between flights. link
He said they believed they could get to the point, probably by next year, where the Falcon 9 booster can be re-flown within 24 hours, with inspections only and no hardware changed, not even the paint. link
He said reusability is absolutely fundamental to a breakthrough in access to orbit and beyond. Earth's high gravity and thick atmosphere make it hard; it would be easier on Mars. link
Q15 How competitive is China?
He guessed that in 10 years more than half of new vehicle production in the United States would be electric, and China would be ahead of that — by far the most aggressive on Earth on electric vehicles and solar. link
Companies will squawk that regulation will stifle innovation and move to China. He said it won't: Boeing has not moved to China, and the same on cars, unless the regime is really overbearing. link
Says competition for AI superiority at national level is the most likely cause of WW3. link
Q16 How good is Tesla self-driving?
His guess: in probably 10 years it will be very unusual for cars to be built that are not fully autonomous. Tesla cars already have the sensors; it is mostly software. Getting in a car will be like getting in an elevator. link
The demo used only cameras and GPS. Solve vision and autonomy is solved; cameras can be superhuman, maybe ten times better than humans. link
One of the biggest risks for autonomous vehicles is a fleetwide hack. They have to make sure that is basically impossible, and that people in the car have override authority no amount of software can override. link
Explains an Autopilot software delay: Tesla has the most advanced AI neural net of any consumer product, so it gets exhaustive testing. link
Q18 Are we in a simulation?
Video games went from Pong to almost photorealistic. Extrapolate any rate of progress and those games will be indistinguishable from reality. How do we know that did not already happen, and that we are not in one of those games? link
Muses that the singularity for this level of the simulation is coming soon. link
Q20 How hard should founders work?
He said they max out R&D; he spends most of his time on engineering — probably 80% of his week is engineering meetings — and any revenue goes right back into R&D. link
Q22 When will Starship go to Mars?
Any 50-year prediction will be wrong, but he hopes we are on Mars, maybe the moons of Jupiter, traveling the solar system and preparing missions to nearby star systems. He thinks all of that is possible in 50 years. link
Mars is in a useful transfer window roughly 6 months every 2 years. If the ship can get to Mars in under 3 months and back inside that window, you get to reuse it twice as often. link
Plan is to do powered landings on Mars for sure, but with a vastly bigger ship than Dragon. link
The goal was the 2022 Mars rendezvous, then in 2024 to try to fly four ships, two of them crewed (captions: 'for ships' / 'crude'). link
Q23 Will humans stay in control of AI?
If only a few people had ultra-smart AI they would be dictators of Earth. It is extremely important that AI be widespread and that we solve the bandwidth issue, so it is tied to the sum of individual human will — a more even, more egalitarian playing field. link
Q30 How do you hire and manage people?
Gut feel more than anything. He asks people to tell the story of their life and the decisions they made, and about the most difficult problems they worked on and how they solved them. People who really solved it know the little details. link
Q37 How should AI labs manage safety?
When something is a danger to the public there needs to be some government agency, like regulators. He is not a fan of regulators, but aircraft, cars, drugs and food have them. AI is a public risk. The question is whether government reaction speed matches AI's advancement speed. We probably need some kind of regulatory authority. link
First advice: pay close attention to the development of AI. Scientists can get so engrossed they miss the ramifications. For public safety, governments should keep a close eye on AI and make sure it is not a danger to the public. link
Part of the role of government is public safety. The right move is to establish a government regulatory agency that first gains insight, then starts applying rules, as with the FAA, cars, drugs and food, so people do not cut corners on AI safety. link
AI is a rare case where we need to be proactive in regulation instead of reactive, because by the time we are reactive in AI regulation it is too late. link
The first order of business is to gain insight: right now the government does not even have insight. Stand up a regulatory agency whose initial goal is insight into AI activity, then put regulations in place to ensure public safety. link
Argues AI should be regulated like everything else that is a danger to the public. link
Q44 What is Neuralink's near-term goal?
Everyone is already superhuman and a cyborg via phone and computer. The limit is bandwidth, especially output. Solve that with a high-bandwidth neural interface to the cortex so a digital tertiary layer is symbiotic with us. link
A closer merger of biological and machine intelligence. We are already a cyborg via phones and computers. The bottleneck is bandwidth: a computer can communicate at a trillion bits per second, a thumb maybe 10. A high-bandwidth brain interface could achieve symbiosis and maybe solve the control problem. link
Q45 Is there hidden alien or advanced propulsion technology?
It seems probable. One of the great questions in physics and philosophy is: where are the aliens? Maybe they are among us. Some people think he is an alien. Not true. link
Q53 What's your vision for energy abundance?
Eventually almost all houses will have a solar roof, on a 40- or 50-year timescale; in 15 years it would be unusual not to. link
He said eventually almost all houses will have a solar roof. link
He said that 15 years from then it would be unusual to have a roof without solar. link
He called the solar roof something incredibly revolutionary that is beautiful. link
He said the gigafactories will be powered by sustainable energy. link
Tesla's overarching optimization is how many years it can accelerate sustainable production and consumption of energy. If they accelerate it by 10 years he would consider that a great success; even 5 years would still be pretty good. link
He guessed that in 10 years more than half of new-vehicle production in the United States would be electric. link
He said there is no question in his mind that all transport, with the ironic exception of rockets, will go fully electric — planes, trains, automobiles, ships. The unpriced externality is the harm of changing the chemical constituency of the atmosphere and oceans. link
He said just as a big satellite runs purely on solar and a battery, you can power Earth that way. link
He said Tesla insisted on being held to 100 days or it is free. link
Australia is well placed for solar (not too far north or south) and could power the whole country with solar, wind, geothermal and hydro, even export power to Asia. link
He said it matters whether the transition to sustainable energy takes 5, 10 or 20 years. link
He said they worked as hard as possible to show it is not just talk, it is reality. link
Q61 Will you stay CEO of Tesla?
He intends to stay with Tesla as far into the future as he can imagine; Model 3 and the Semi were coming. link
He said he intends to stay with Tesla as far into the future as he can imagine. link
Q107 When will a city on Mars be self-sustaining?
For SpaceX the focus is getting people and cargo there at a cost they can afford. There is a threshold cost per ticket or per ton to the surface of Mars below which a self-sustaining city can develop and above which it cannot. link
Q120 How should government award contracts?
He said NASA's space-station resupply contracting — two competitors, fixed-price, hardware-oriented milestones, recompete if one fails — is a great model that should be adopted throughout government. link
As soon as you make something a cost-plus contract you incent the contractor to maximize costs. Change to competitive commercial bids, at least two competitors, concrete milestones (PowerPoint does not count), and be prepared to fire one and recompete. NASA already did this with commercial cargo. link
Q121 Should governments put a price on carbon?
Says a carbon tax is what is really needed to move the needle. link
Q126 What did Elon promise for the Model 3?
He said Model 3 is a smaller, more affordable version of Model S with less range, power and features. link
He said Tesla was going to be in production hell. link
He said the standard Model 3 would be $35,000 for 220 miles of range. link
Q129 Can the Boring Company solve traffic?
He said he was optimistic tunnelling can be improved by at least fivefold, maybe tenfold. link
He said you have to go 3D, probably down, and use hyperloop-type ideas for longer distances. link
Says public transit has not solved traffic, and multi-level tunnels can, but only with much faster and cheaper tunnelling. link
Q130 What is the Tesla Semi?
He said the Tesla Semi goes 0 to 60 in five seconds. link
He said from day one a Tesla Semi will beat a diesel truck on economics. link
Q131 Why does rocket reusability matter?
He said rapid and complete reusability of rockets is key to opening up space and becoming a multiplanet species. link
He said if not that year, he was confident SpaceX would reach rapid zero-change reflight the next year. link
Q132 How fast is Tesla expanding Superchargers?
He said by the end of next year there would be three times as many Superchargers. link
Says you can now drive anywhere in the US, most of Europe, China and Japan on Superchargers. link