Elon Musk's answers in 2014
Q1 How dangerous is AI and how do we make it go well?
Recommends Superintelligence and warns AI is potentially more dangerous than nukes. link
He said machine intelligence is advancing much faster than almost anyone realizes, even in Silicon Valley, and a superintelligence with a bad utility function could conclude the way to stop spam is to get rid of humans. link
We should be very careful about AI. With artificial intelligence we are summoning the demon — the guy with the pentagram and holy water who is sure he can control it. It did not work out. link
He said with artificial intelligence we are summoning the demon. link
Q5 Why go to Mars / what is SpaceX's mission?
He would love SpaceX to be the company that brings humanity to Mars, and to see it while he is still alive. link
SpaceX was founded to make radical improvements to space-transport technology, particularly reliability, safety and affordability. link
He wanted to see humanity go beyond Earth with people on Mars, plus widespread electric vehicles and renewable energy. link
It is a much more exciting, inspiring future if we are out there exploring the stars rather than forever confined to Earth. He wondered when NASA was going to Mars. link
He went to the NASA website and there was no plan to go to Mars and no plan to really even take the next step in space exploration. link
Becoming multiplanetary is life insurance for life as a whole, so the light of consciousness propagates into the future, using a small share of economic output. link
The wise move is to do it now: the window of technology is open for the first time in the four-and-a-half-billion-year history of Earth, and it may close. link
Primary focus should be problems on Earth, but for 1% of our resources we could buy life insurance for life collectively. link
Q8 Why are your timelines so often late?
He said there must be an allowance for failure when you try something new. link
Q12 What's the biggest remaining Starship problem?
Rockets had not evolved since the 1960s. Lower-cost engines, structure, electronics and launch operations, run with little overhead, led to roughly three-to-four-fold cost improvements over other U.S. rockets. link
Q14 What do you think of OpenAI?
He said he created OpenAI as a nonprofit so there would be no fiduciary duty to profit from the AI technology, intending to democratize AI power. link
Q16 How good is Tesla self-driving?
He likened it to Waze: cars operating collaboratively as a network with crowd-sourced intelligence. link
He said Tesla would probably achieve true autonomous driving in five or six years. link
He said the standard for fully autonomous driving must be much greater than for a person. link
Q18 Are we in a simulation?
He said the odds we are in base reality are one in billions, and we should hope that is true because otherwise we either create indistinguishable simulations or civilization ceases to exist. link
He answered an audience VR question by saying maybe we are in a simulation right now, then sketched video games becoming indistinguishable from reality. link
Q20 How hard should founders work?
He said hours vary; these days about 80–85 a week, and for a while over 100, with the pain of work hours increasing exponentially. link
He framed a 90-hour week as 45 hours for SpaceX and 45 for Tesla, adding it is a bloodbath on a personal basis. link
He said that, particularly when starting a company, you need to work super hard. link
He said if someone else works 50 hours and you work 100, you get twice as much done in a year. link
Q22 When will Starship go to Mars?
He said the first Mars mission would probably be in 2018, sending a Dragon 2 to the surface on a heavy rocket. link
He said SpaceX should hopefully be able to send the first people to Mars in about 12 years. link
He said SpaceX would send a Dragon to Mars in 2018; he would not be on that flight. link
The subcooled methalox system is intended as a fully reusable Mars transportation system, not merely to LEO. He thought they could start to see some test flights in the five or six year time frame. link
Q30 How do you hire and manage people?
He said the most important thing when creating or joining a company is to attract great people. link
Q37 How should AI labs manage safety?
If he had to guess at the biggest existential threat, it is probably AI. He was increasingly inclined to think there should be some regulatory oversight at the national and international level so we do not do something very foolish. link
Q47 Has Tesla come close to failing?
He thought Tesla would most likely fail, but if something is important enough you should try even if the probable outcome is failure. link
He called 2008 a terrible time: three rocket failures in a row, Tesla financing falling apart, SolarCity difficulties and a divorce, everything going wrong at once. link
In 2008 the choice was Tesla going bankrupt unless he put in remaining capital, including borrowing from friends. link
They had maybe a week's worth of cash in the bank or less, and only two or three business days left in the year. He never thought a nervous breakdown was possible; that was about as close as he was going to come. link
Q53 What's your vision for energy abundance?
The message is you can drive for free forever on pure sunlight — even if the grid breaks down in a zombie apocalypse, you can still charge. link
SolarCity is about sustainable energy creation; Tesla is about sustainable energy consumption. link
The biggest impact Tesla will have is not the cars they make themselves, but showing that you can make compelling electric cars that people really want to buy. link
He said he has no fundamental dislike of hydrocarbons, but a non-renewable resource will not work long-term. link
There are 100 million new cars a year and 2 billion gasoline or diesel cars on the road. You do not just need one gigafactory; you need like 200 gigafactories just for new-car production. link
He loves the idea. Everything will go electric — everything will go fully electric, except for rockets. link
Q93 How can we get fast internet on Mars? (asked on X)
Communications is very important. We will need terabit-level communication between Earth and Mars, which means a tight-beam laser system and relays, because sometimes Mars is on the other side of the Sun. link
Q99 When will Starship go to the Moon? (asked on X)
The Moon is not a necessary step, but if you have a rocket capable of going to Mars you might as well go to the Moon along the way — like crossing the English Channel relative to the Atlantic. You probably end up with a moon base because why not. link
Q106 When will Starship demonstrate orbital refilling?
With physics we know today, the building blocks are there: a densified methalox rocket with on-orbit refueling, a Mars colonial fleet built up between Earth-Mars synchronizations every 26 months, then the fleet departs at the optimal transfer. link
Q107 When will a city on Mars be self-sustaining?
The long-term ambition of SpaceX is the technologies for a self-sustaining city on Mars. Wings and a runway do not work on the Moon or Mars, so they have to get good at propulsive landing. link
The goal should be a self-sustaining civilization on Mars, not simply a mission. A flags-and-footprints trip would be awesome pictures and a new altitude record, but it is not the thing that fundamentally changes the future of humanity. link
Q118 What's your algorithm for designing products?
He described first-principles reasoning: boil things down to the most fundamental truths and reason up from there. link
Q120 How should government award contracts?
Congressional records for FY13: the Air Force paid an average of $380 million for each national security launch, while subsidizing ULA's fixed costs more than a billion dollars a year even if they never launch. SpaceX's price was well under $100 million. link
Had SpaceX been awarded the missions ULA received under its recent non-competed 36-core block buy, he said they would have saved taxpayers $11.6 billion. link
When a competitive environment exists, the government should use firm fixed-price FAR Part 12 contracts that properly incent contractors to deliver on time and on budget, and eliminate the billion-dollar annual subsidy to ULA. link
As a country we decided competition and the free market are good and monopolies are not. From the point Boeing and Lockheed's launch businesses stopped being competitors, the costs doubled. link
Fixed-price competition is the better way to go when there is competition. Cost-plus after a competition lets companies raise prices as costs grow. Mission-unique increments could be cost-plus; the vast majority of the contract should be fixed price. link
Q124 When will Dragon carry astronauts?
He said Dragon 2 would be able to land anywhere on Earth with the accuracy of a helicopter. link
He said it would carry seven astronauts for several days. link