Elon Musk's answers in 2018
Q1 How dangerous is AI and how do we make it go well?
Says confusing narrow AI with general AI, which has vastly more compute and an open-ended utility function, means humanity is in trouble. link
He said the danger of AI is much greater than nuclear warheads and it is insane to have no regulatory oversight. link
Says nothing will affect the future of humanity more than digital super-intelligence. link
Q5 Why go to Mars / what is SpaceX's mission?
He said Falcon Heavy could send people back to the Moon if you did a bunch of launches with orbital refilling, though he would not recommend that: the new architecture (BFR) is the way to go. link
Jokes that you can’t show up at Mars in something the size of a rowboat. link
On a longer time scale: extending life beyond Earth and making life multiplanetary, in the spirit of Asimov's Foundation. link
Says life cannot just be about solving sad problems; there need to be things that inspire people. link
Says BFR will take over from Falcon, with the goal of enabling anyone to move to the Moon, Mars and the outer planets. link
The window of opportunity may be open a long time or a short time; they should not assume it stays open, and should become a multi-planet civilization as soon as possible. link
Q7 Will people still need to work?
Says universal income will be necessary if AI takes over most human jobs. link
Q8 Why are your timelines so often late?
He said anyone who closed the design has a mental checklist scrolling of all the ways it can fail — thousands of things can go wrong and everything has to go right. link
He said they tried to cancel the Falcon Heavy program three times at SpaceX because it was way harder than they thought. The initial idea was just sticking two first stages on as side boosters. link
He felt optimistic about the timeline but said people have told him his timelines historically have been optimistic, so he was trying to recalibrate. link
Model X was an exercise in hubris — they got carried away with the art and it probably delayed Model 3 a little. link
Q11 What drives you / what is the meaning of life?
He rejects the idea that SpaceX is about money: a rocket company is one of the dumbest and hardest ways to make money. link
He said technology is the closest thing to magic we have in the real world. link
Q12 What's the biggest remaining Starship problem?
They wanted a highly reusable heat shield capable of absorbing heat from interplanetary entry velocities. The ship is by far the hardest because it comes in from Mars/Moon transfer velocities, way harder than low Earth orbit. link
The ship part is by far the hardest. He thought they understood reusable boosters; reusable spaceships that can land propulsively are harder, so they were starting with the hard part first. link
The ship would do about 150 metric tons to orbit fully reusable. Rapid reusability is the fundamental breakthrough needed in rocketry. link
From September 2018: hop flights next year (2019), high-altitude high-velocity ship flights in 2020, and first orbital flights in about two to three years if things go well (~2020–2021). link
Q14 What do you think of OpenAI?
He mentioned OpenAI among SpaceX, Tesla, Boring Company and Neuralink, and said OpenAI is a non-profit and he was not on the board. link
Q15 How competitive is China?
He says the pace of innovation is the fundamental determinant of competitiveness. link
Q16 How good is Tesla self-driving?
He says the hardest thing to predict about autonomy timing is regulatory approval. link
He says the statistics are unequivocal that Autopilot improves safety. link
He says Tesla thinks even the current autonomy cuts the probability of a death by 50%. link
He says the issue with serious Autopilot accidents tends to be complacency. link
Says perfect is the enemy of good: a system that on balance saves lives should be released. link
Says Autopilot work has focused on safety and that Version 9 in August will begin enabling full self-driving features. link
Says it will be easy to bully a self-driving car because it always yields, so there may be a manual override. link
Q18 Are we in a simulation?
He said the argument for the simulation is quite strong given any rate of improvement over long time frames. link
He said if we are in a simulation, base reality outside it is most likely really boring. link
Jokes that if there is no listener, a simulation probably wouldn't spend computing power on sound. link
Q20 How hard should founders work?
Most people don't realize he is in the factory, design or engineering meetings or production about 89% of the time; Twitter is almost nothing. link
He said whenever you try to do something and avoid hard work, good things do not come from that. link
Says nobody ever changed the world on 40 hours a week. link
Q21 What is the purpose of X?
A friend suggested Twitter would be a good way to talk directly to the public and bypass the media, so what he wanted to say would not always be filtered through the lens of a journalist. link
Q22 When will Starship go to Mars?
He said they might be able to do short-hop flights with the spaceship part of BFR maybe next year (2019). link
He said it is conceivable they do a first test flight in 3 or 4 years of a full-up overall test including the booster, capable of going to the Moon shortly thereafter. link
With orbital refueling/retanking, BFR (captions: VFR) was designed to take 100 tons all the way to the surface of Mars. link
They would do a bunch of test launches without people on board before putting anyone on board. link
Q23 Will humans stay in control of AI?
He said the tiny straw of information between biological and digital self must become a giant river. link
Q26 What do you think of tariffs?
Points out that a US car going to China pays 25% duty versus 2.5% for a Chinese car coming to the US. link
Says he just wants a fair outcome where tariffs and rules are equally moderate. link
Q29 How do you think about doing good?
Put money and attention into making a product people love so sales grow by word of mouth; Tesla spends nothing on ads, and he pays full retail for his own cars. link
Q30 How do you hire and manage people?
He says machines are bad at picking up pieces of fluff, so Tesla removed an over-complicated automated step. link
He says idea generation far exceeds the ability to execute, so Tesla must stay focused. link
He says the number of third-party contracting companies Tesla uses has gotten out of control. link
He said a company is just a group of people gathered together to make products. link
Q39 What about unions at Tesla?
Says workers can form a union whenever they want and he cannot stop them; they just do not want to. link
Says nothing stops the Fremont team voting union, but asks why they would pay dues and give up stock options. link
Says the factory injury rate under UAW with Toyota/GM was twice today’s non-union rate. link
He says the UAW chose to leave the Fremont plant before Tesla arrived, and Tesla is the reason people there have jobs. link
He accuses the UAW of driving GM and Chrysler into bankruptcy and abandoning NUMMI workers, the plant Tesla took over. link
Says he has never stopped a union vote or removed a union; UAW abandoned the factory. link
Endorses freedom to form a union, and freedom not to. link
Says Tesla has always provided healthcare, hourly wages above union and Tesla stock. link
Says he is not against all unions, but the UAW has a record of destroying productivity. link
Says the UAW lacks the 30% of signatures needed for a union vote at Fremont. link
Says workers in pro-union California can form a union anytime, but the UAW can’t get enough signatures. link
Q44 What is Neuralink's near-term goal?
Neuralink's goal is to help solve brain injuries — strokes, dementia, Parkinson's, epilepsy — using computer chips, with a longer-term view to achieving symbiosis with artificial intelligence, which he said could be important on an existential level for civilization. link
He said the purpose of Neuralink is a high-bandwidth brain interface so humans can be symbiotic with AI. link
Q46 Do regulations hold you back?
Defends Tesla's policy of issuing recalls before anyone is injured. link
He says regulators respond to public pressure and the press. link
Proposes direct democracy, short laws and automatic expiry of rules to prevent death by bureaucracy. link
Q47 Has Tesla come close to failing?
2008 was brutal: third consecutive Falcon failure, Tesla almost went bankrupt, financing closed 6 p.m. Christmas Eve, last hour of the last day. link
In April 2018 he said Tesla would be profitable and cash-flow positive in Q3 and Q4, so it would not need to raise money. link
He accepts the criticism that Tesla was not yet profitable, saying you are not a real company until you are. link
He says the use of third-party contractors had got out of control and promised a big clean-up. link
He credits hard work and sacrifice by talented people for the production outcome. link
Says complexity is the devil and the team is working incredibly hard on the ramp and profitability. link
Says Ford and Tesla are the only two American car companies to avoid bankruptcy, and working less is not an option. link
Q50 Why keep Tesla and SpaceX as separate companies?
He said it is similar at both companies: he is CEO and chief engineer / chief designer of the rockets, and he drives the technical direction of the product. link
Q53 What's your vision for energy abundance?
In April 2018 he said Tesla batteries were delivering power at 662 locations in Puerto Rico, with several hundred more coming. link
He says South Australia took a chance on the world's biggest battery and it worked out really well. link
Says Singapore has enough land to make most of its electricity from solar, at about 250MW per square km. link
Says early Teslas could act as a battery outputting power, and that this may be worth revisiting. link
Q59 Will Tesla own the robotaxi fleet?
He describes owners adding their cars to a shared autonomous fleet, able to recall the car at will and restrict who uses it. link
He describes owners offering their cars as a kind of Robo Lyft or Robo Uber. link
Q79 When does Tesla stock have its NVIDIA moment? (asked on X)
He says Tesla has no interest in satisfying the desires of day traders. link
Q99 When will Starship go to the Moon? (asked on X)
He was not sure they would test a flight around the Moon, but probably they would try that without people before sending people. link
Q107 When will a city on Mars be self-sustaining?
About half his money was intended to help problems on Earth and half to help establish a self-sustaining city on Mars. link
Q117 What's the long-term vision for Starlink?
Says if the test satellites succeed, Starlink will serve the least served. link
Q118 What's your algorithm for designing products?
Admits excessive automation at Tesla was a mistake, his mistake, and that humans are underrated. link
He describes a robot built to place fiberglass fluff on battery packs; tests then showed the part was not needed. link
He says the path to manufacturing efficiency is velocity and density, not just spending more capex. link
He says a great production system is primarily a software problem. link
He says moats are lame and the rate of innovation is what matters. link
He says Tesla's understanding of production is improving dramatically, exponentially. link
He says some things suit manual operation and others suit automated operation. link
He recalls the ironically foolish fiberglass mats on top of the battery pack. link
He says Model Y will be a manufacturing revolution because Tesla does not want to go through the Model 3 pain again. link
He says a really complicated machine needs a ton of maintenance to keep operating. link
He says little money is spent early in a development program, but early decisions have massive implications. link
He says Tesla tried to automate placing fiberglass fluff on battery packs, which was ridiculous. link
Q126 What did Elon promise for the Model 3?
Says Model 3 is the whole point; the expensive cars were made to get to the mass-market car. link
Says Tesla must reach 5k/week of Model 3 before adding complexity. link
In May 2018 he said the focus was getting Model 3 to about 5,000 a week at a good margin and making Tesla profitable. link
In May 2018 he said he was quite confident Tesla would achieve GAAP net income and positive cash flow in Q3 2018. link
He says Tesla could have made the Model 3 much easier to produce. link
Says Model 3 had about half a million reservations, so even at 5,000 a week it would take two years to fill existing demand. link
Announces the dual-motor Model 3 Performance specs: 0-60 mph in 3.5 s, 155 mph top speed and 310 miles of range. link
Says the Model 3 Performance costs $78k with all options (apart from Autopilot), about the same as a BMW M3 but quicker. link
Explains shipping the minimum-cost Model 3 right away would make Tesla lose money and die. link
After the Consumer Reports braking test, he said a firmware update would fix Model 3 braking distance within days. link
He said it was quite likely Tesla would achieve a 5,000-car week by the end of that month. link
Q127 What is the new Tesla Roadster?
He said the Roadster shows an electric vehicle can outperform a gasoline car in every way. link
Announces a SpaceX option package for the new Roadster with about 10 small rocket thrusters. link
Says the new Roadster is meant to beat gas sports cars on every performance metric and move the halo effect to electric cars. link
Says the trade-off is that the Roadster becomes a two-seater, with the pressure vessel in place of the kid seats. link
Promises the next-generation Roadster will be out of this world for people who love to drive. link
Q128 Why did Tesla build the Cybertruck?
Says the Tesla truck will have dual-motor all-wheel drive and load-adjusting suspension as standard. link
Says the pickup will have 240V outlets for heavy-duty power tools, with no generator needed. link
Q129 Can the Boring Company solve traffic?
Says the Boring Co loop would have thousands of parking-space-sized stations instead of a few big subway stations. link
Says Boring Company tunnels and Hyperloop will prioritize pedestrians and cyclists over cars. link
Describes his tunnel transport as a 150 mph underground, autonomous, electric bus. link
He said tunnels are the only way he could think of to address chronic traffic in major cities. link
He said in theory you can have hundreds of levels of tunnel and solve any level of traffic. link
He said the keys are figuring out how to tunnel fast, safely and at a reasonable cost. link
Says the Boring Company will use tunnel dirt to make bricks for low-cost housing. link
In May 2018 he said the DC-to-NY tunnel had started and he hoped to start LA to SF the next year, as a true Hyperloop. link
Says Boring Company Hyperloop will take people from city centre to a spaceport in 10 to 15 minutes. link
He said the thing about tunnels is you can go 3D underground. link
He said we need to build tunnels way faster and for a lot less money. link
He said the probability of something falling on your head is much higher with lots of flying cars. link
Q130 What is the Tesla Semi?
Says he is optimistic about beating the Semi specs announced at the unveiling for the same price. link
He says commercial trucking is driven by cost per mile, not looks. link
He says the Semi will take quite a bit of revenue from railways. link
He says economics are fundamental for heavy-duty commercial trucking. link
He expects truck platooning to affect the rail industry negatively. link
He says a 50% rise in diesel costs destroys trucking companies' business. link
He said production of Model Y, and something similar for Semi and Roadster, was targeted for the first half of 2020. link
Q131 Why does rocket reusability matter?
Urges Russia to focus on reusability: single-use rockets cannot compete, just as single-use aircraft could not. link
Says it will seem bizarre in the future that rockets were crashed into the ocean instead of reused. link
Q132 How fast is Tesla expanding Superchargers?
Plans an old-school drive-in and roller-skate restaurant at a new Tesla Supercharger in LA. link
He says the Supercharger network was never meant as a walled garden and other automakers can use it if they share costs. link
He says long range matters more than a super fast charge time. link
Says Tesla will probably unveil the next-generation Supercharger in late summer. link
Says Tesla will extend Superchargers to South America around the end of the next year. link
Q133 Can the world solve its water problems?
Says desalination costs only 0.1 cents per liter. link