In a conversation with XPRIZE Founder Peter Diamandis, Musk laid out what he sees coming from the rising force of AI, while laying out a timeframe for SpaceX’s first unmanned and manned Mars missions, and lots more.
Elon watch: Skip this bit if you’ve been paying attention lately
It’s certainly been a wild couple of months for the world’s richest (known) human. SpaceX, of course, made a spectacular world-first, catching Starship’s massive Super Heavy booster with “Mechazilla” chopstick arms built into the launch tower, clearing the path for the biggest space launch vehicle in history to start coming straight down from orbit, and land pretty much stacked right up on top of its booster, ready to launch again within hours. This will revolutionize the space industry (again), crash per-weight launch prices (again), and it also lays the foundation for Musk’s ultimate lifetime goal of getting a sustainable human colony built on Mars.
Tesla’s Cybercab/Robovan launch was an entertaining dog ‘n’ pony show in comparison, but set out the company’s plans for fully autonomous robotaxis in detail. Neuralink wired up its second brain implant patient, Alex, who has been using the chip to play video games and operate CAD software hands-free.
Meanwhile, X Ai stood up the world’s most powerful AI training supercomputer cluster, Colossus, building all the infrastructure and wiring up 100,000 nVidia H100 GPUs in a ludicrous 122 days start to finish. The company now has more GPUs at its disposal than Google AI, OpenAI, Meta AI, Microsoft or nVidia itself – and Musk says it’ll have double the processing power within a few months thanks to 50,000 additional H200 GPUs.
Jensen Huang says building xAI’s computing supercluster in just 19 days, from concept to training, was a “superhuman” effort and the only person in the world who could have done it is Elon Musk pic.twitter.com/yUGBnxDM0T
— Tsarathustra (@tsarnick) October 13, 2024
So, if his team has the right approach, Musk has put himself in an excellent position to elevate X’s Grok AI language model past OpenAI’s GPT models, to potentially become the global leader in AI technology over the next year or two.
Not that you’d notice, because over the last few weeks he’s seemed much more interested in politics than anything else, literally jumping up and down as Donald Trump’s #1 supporter as the US presidential election bears down on the world. Indeed, he sees Trump as an opportunity to get his hands on the levers of government, and is now promising to head up a Department of Government Efficiency (DOGE) that would slash at least US$2 trillion, or nearly a third, of the US federal budget should Trump return as president.
Yesterday’s discussion between Elon Musk and Peter Diamandis
Musk certainly enjoys keeping busy, but took time to join Peter Diamandis via video link for a chat about the future of AI as part of a Future Investment Initiative Institute event at the King Abdulaziz International Conference Center in Riyadh yesterday. Enjoy the full 22-minute video here, or we’ve pulled out some key quotes below with timestamped links.
Elon Musk and Dr. Peter Diamandis #FII8 Conversation on the Future of #AI
On the pace of AI improvement
(1:38) “It’s a difficult thing to quantify exactly, but I certainly feel comfortable saying that it’s getting 10 times better per year … So four years from now, that would mean 10,000 times better. Maybe 100,000 times.”
On when AI will overtake the output of humanity
(2:07) “I think it’ll be able to do anything that any human can do, possibly within the next year or two. How much longer than that does it take to be able to do what all humans can do combined? I think not long. Probably only, I don’t know, three years from that point. So like 2029, 28, something like that.”
On AI as an existential threat to humanity
(3:05) “It’s most likely to be great. There’s some chance, which could be 10 to 20%, that it goes bad. The chance is not zero that it goes bad, but overall one could say the cup is 80% full, maybe 90% … I think AI is a significant existential threat, and something we should be paying close attention to. It’s probably the most significant near-term threat.”
On global population collapse
(4:08) “The longer term threat is global population collapse. Birth rates have been collapsing pretty much worldwide, and we’re headed for a situation where, for example, based on current birth rates, South Korea would have about a third of its current population, Europe would have about half its current population – much less.
“And I should say, those numbers are if the birth rate were suddenly to return to 2.1 per woman, which is the stability point. If this effect continues, you would see really many countries become 5% of their current size or less within three generations.”
On the future of energy
(9:15) “There will be a tremendous amount of energy that’s needed for digital intelligence, and also for the electrification of transport. Those two things are a big deal … In the long term, almost all the energy that we get is going to be from the Sun … One way to look at civilization is progress on the Kardashev scale … It’s not clear to me we’re above 1% on the Kardashev scale one, which means you’ve harnessed all the power of a planet …
“Kardashev scale two is where you’ve harvested all the power of your sun. The Sun is overwhelmingly the largest source of energy in the solar system, everything else amounts to maybe a trillionth … So almost all energy long-term will be solar. It rounds up to 100%, that’s how much of the energy in the future will be solar.”
On the state of humanoid robotics by 2040
(13:48) “I think by 2040, probably there are more humanoid robots than there are people … I’m often optimistic on timing, although the press will report when I’m late, but not when I’m early – for example, our Shanghai factory, we thought it’d take about a year and a half, and we did it in 11 months. Our Giga Nevada factory, we thought two years, we did it in 18 months. The Colossus cluster … So I’ve been early actually, many times …
“But once you get out to 2040, that’s a long time from now, going on 25 years, I think there’ll be at least 10 billion humanoid robots. The price point will be, I think, quite low, probably $20-25,000 for a robot that could do anything …”
On a future of abundance
(15:25) “Assuming we’re on the good path of AI, I think we’ll be in a future of abundance … Basically anyone will be able to have any goods and services they want. The actual marginal cost of goods and services will be extremely low in the future.”
On the Super Heavy booster catch
(16:00) “Not bad for humans! … No AI was involved in that whatsoever, I’m glad to say we did that entirely with human brains and without AI. I think in the future, the AI might look at that and say ‘not bad for a bunch of monkeys!'”
On when SpaceX will get to Mars
(16:33) “I think we’ll be able to launch some Starships to Mars in two years. So the next Mars window in in about 26, 27 months … We’re at the beginning of a Mars transit window now and they occur every 26 months. So in just over two years, we’ll be sending our first uncrewed Starships to Mars. And then if those work out, and we don’t increment the crater count on Mars, we’ll send humans two years after that.”
“I feel more optimistic about it under a Trump White House than a non-Trump White House, because the biggest impediment to progress that we’re experiencing is over-regulation … It takes longer to get the permit to launch than to build a giant rocket … America and a lot of countries are getting a slow strangulation from over-regulation. Unless something is done to push back on that, it’ll eventually become illegal to do almost any large project, and we won’t be able to get to Mars.”
On Tesla’s Optimus humanoid robot production schedule
(20:59) “Optimus starts limited production next year in 2025, and then should be in volume production in 26, and then will grow to, I think, ultimately be the biggest product of any kind, ever … I kind of agree with Ark Invest and Cathie Wood that autonomy – like, sort of robotic taxis – makes Tesla kind of like about a $5 trillion company. The Optimus robot, I think, makes Tesla a $25 trillion company …
“It’s not even clear what money means in that future … It does become kinda post-capitalist … Looking at the most likely bright side, we’re headed for an age of abundance … It won’t be universal basic income, it’ll be a case of universal high income, is the most likely outcome.”
It’s always hard to know how to treat Elon Musk’s visions of the future, so you’ll have to administer your own preferred grain of salt to the statements above.
Musk is unquestionably the most fascinating figure of our time in the field of technology, controlling a dizzying array of companies, many of which are pushing boundaries and making astonishing progress in cutting-edge fields. Space travel, advanced AI, humanoid robotics, electric and autonomous transport, brain-computer interfaces, social media … He may be the single best-placed human to tell us what the future will bring, because (ironically for the world’s biggest proponent of self-driving vehicles) he might just have his hands on the steering wheel to a greater degree than anyone in history as humanity moves into a hyper-technological era.
He’s also incredibly divisive, and unlike most billionaires, lives his life very much in public. His companies’ achievements are thus near-impossible to separate from his increasingly right-wing “Dark Gothic MAGA” politics, his incendiary X feed, his frequently childish sense of humor, his flirtation with conspiracy theories, his … complicated family and family history, his naked contempt for unions and regulators, his notorious temper, his well-known tendency to demand “hardcore” commitment and crazy hours from his employees, and any one of a thousand other things that have earned him 57th place on Ranker’s current list of most-hated celebrities.
He’s certainly leaving future historians no lack of material to work with.
