Musk forecasts AI abundance, robots and Mars missions at FII8

At a 2024 Riyadh event, Musk discussed AI risk, energy, humanoid robots and ambitious schedules for Mars and autonomous cars.

EMElon MuskInvité
Chaîne
FII Institute
Publié
29 oct. 2024
Vidéo
22:29
Lecture
3 min
Original
anglais
Ouvrir sur YouTube
En faisant défiler, la vidéo se réduit en bas à droite et continue. Les sous-titres suivent votre langue de lecture.
Langue de lectureDisponible en 2 langues
Cet article n’est pas encore disponible en français ; il s’affiche dans l’original en anglais.
RÉSUMÉ

En 30 secondes

At the Future Investment Initiative in October 2024, Elon Musk predicted rapid AI progress and a future in which robots make goods and services abundant, while estimating a nonzero chance of severe AI harm. He discussed energy constraints and proposed truth-seeking AI as a safety approach. His timelines for robots, Mars missions and unsupervised Tesla vehicles were forecasts; FSD (Supervised) requires active driver oversight.

  1. 1Musk forecast rapid AI progress and acknowledged riskHe estimated a 10–20% chance of a bad outcome; neither the rate nor the probability is established by the discussion.
  2. 2AI growth will require power and infrastructureHe identifies transformers, installation and electricity as potential constraints.
  3. 3Robot numbers and prices are long-term guessesMusk predicted more humanoid robots than people by 2040 and a $20,000–$25,000 price point.
  4. 4Mars and robotaxi timelines were 2024 expectationsHe outlined conditional Starship missions and projected Tesla deployment dates; the supervised feature requires a driver.

À retenir

  1. Musk’s 10–20% estimate of a bad AI outcome is a subjective judgment.
  2. He expects computing, transformers and electricity to constrain AI expansion at different stages.
  3. His 2040 robot counts, prices and abundance scenario are speculative forecasts.
  4. Tesla says FSD (Supervised) requires active driver supervision and does not make a vehicle autonomous.
  5. Population change depends on fertility, mortality and migration, not fertility alone.
TEXTE INTÉGRAL

Lire le texte remanié, chapitre par chapitre

Transcription seule
01

AI progress and downside risk

Musk offers forecasts for AI capability and discusses his subjective risk estimate and safety approach.

Peter Diamandis welcomes Elon Musk to a Future Investment Initiative conversation in Riyadh and recaps recent developments at Musk’s companies, including xAI’s computing cluster, Starship’s fifth flight and the Cybercab reveal.

Asked about AI progress, Musk says capability is difficult to quantify but estimates it could improve tenfold per year. In October 2024 he predicts systems may match individual human abilities in a year or two and exceed the combined abilities of humans several years later. These are his forecasts, not measured results.

Musk estimates a 10–20% chance of a bad AI outcome while saying a good outcome is more likely. He calls advanced AI a significant near-term existential risk; the percentages are his subjective estimates.

He also raises low fertility as a long-term concern and projects sharp population declines in some countries. The [United Nations](https://www.un.org/sustainabledevelopment/blog/2024/07/press-release-wpp2024/) notes that population trajectories depend on fertility, mortality and migration; a current fertility rate alone does not determine a country’s future population.

Diamandis asks what could reduce AI’s downside. Musk advocates a truth-seeking system that values humanity, criticizing what he sees as political bias in other models. He presents this as his approach to safety, not a proven guarantee.

02

Energy and national AI infrastructure

He identifies infrastructure constraints and predicts countries will develop computing capacity.

Discussing xAI’s Colossus cluster, Musk says electrical transformers, installation and eventually energy supply may constrain AI growth. He expects demand from computing and transport electrification to increase and predicts solar will dominate energy over the long term.

Asked what countries should do, Musk says training frontier models requires substantial computing resources and specialized expertise. He predicts that national AI computing infrastructure will become common over time.

03

Humanoid robots and abundance

Musk forecasts robot numbers and prices and describes a possible high-abundance economy.

Diamandis asks about humanoid robots by 2040. Musk predicts there could be more robots than people and that a useful robot might cost $20,000–$25,000. He acknowledges his timing can be optimistic; these figures are speculative.

Musk and Diamandis discuss a possible future of abundance in which AI and robots make goods and services much cheaper. Musk’s description depends on a favorable AI outcome and is a projection rather than a demonstrated economic result.

04

Starship, autonomous cars and Optimus

He outlines conditional Mars plans, vehicle timelines and speculative business valuations.

On Starship, Musk predicts uncrewed launches toward Mars at the next transfer opportunity and possible crewed missions two years after successful uncrewed flights. He argues that regulatory approvals slow launch progress and expresses a preference for a Trump administration; these were 2024 expectations and political views.

Musk forecasts unsupervised driving for some Tesla models in Texas and California in 2025 and Cybercab volume production in 2026. [Tesla’s FSD (Supervised) documentation](https://www.tesla.com/support/fsd) says the supervised feature requires active driver oversight and does not make a vehicle autonomous.

He predicts that future autonomous vehicles will be much safer than human-driven cars and that Optimus will become a very large product. He cites multitrillion-dollar valuation scenarios and a future of “universal high income”; these are speculative projections, not established performance or returns.