“I’ve NEVER seen anything like this” - Elon Musk drops a bombshell in Interview (Feb 29, 2024)

Summary

在这段视频中，埃隆·马斯克分享了他对人工智能（AI）未来5到10年的看法，特别强调了自动驾驶技术的进步，以及其对世界的深远影响。他提到，自动驾驶技术的成功将大幅提高乘用车的使用效率，并可能使得车辆的使用时间从每周10小时增加到50-60小时。此外，马斯克还探讨了AI在其他领域的应用，包括客户服务中的聊天机器人，以及其对提高生产力的潜在影响。他还预测了AI计算能力的快速增长，以及这可能带来的挑战，如电力和变压器的需求增加。最后，马斯克提到了特斯拉即将推出的低成本自动驾驶车辆，以及新款特斯拉Roadster的性能特点。

Takeaways

• 🚗 埃隆·马斯克讨论了自动驾驶技术的发展，强调了它对于未来汽车自主性的重要性。
• 🌍 马斯克提到，尽管自动驾驶在欧洲的发展较慢，但他们正在致力于首先在美国实现这一技术，随后将其扩展到全球。
• 👁️‍🗨️ 描述了使用数字神经网络和摄像头来实现汽车自动驾驶的方式，这类似于人类如何通过眼睛和生物神经网络驾驶汽车。
• 🤖 讨论了人工智能（AI）在各个领域的应用，包括自动驾驶、在线互动、客户服务和物流管理，强调了AI技术的快速发展。
• 💡 强调了AI计算能力的快速增长，预测这一趋势将继续带来巨大的技术变革。
• 🔋 讨论了电动汽车和AI技术对电力和变压器需求的增加，预示了未来可能出现的能源和基础设施挑战。
• 🚀 揭示了特斯拉正在开发的Optimus人形机器人项目，该项目旨在创建能够执行几乎所有人类任务的机器人。
• 🔌 强调了建立充电基础设施的重要性，以支持电动汽车的广泛使用。
• ⚡ 提到了汽车行业朝着48伏系统迁移的趋势，以提高能效和减少铜材的使用。
• 🚗 讨论了特斯拉下一代低成本、高度自主的电动汽车的开发，预示了电动汽车市场未来的变化。

Q & A

埃隆·马斯克认为未来5到10年人工智能会如何改变世界？

马斯克认为人工智能将极大地改变世界，特别是自动驾驶技术的发展，将使汽车完全自动化成为可能。此外，人工智能的计算能力正在以前所未有的速度增长，这将推动人工智能技术的快速进步，包括在线聊天机器人和图像识别等方面。

为什么在欧洲不太能看到特斯拉的自动驾驶技术？

马斯克提到，特斯拉首先需要在美国使自动驾驶技术工作，之后才会增加尝试在世界其他地方推广的复杂性。

为什么自动驾驶汽车的发展如此困难？

自动驾驶汽车的发展困难，是因为要使汽车完全自动化，需要汽车具备高度的智能，包括能够阅读所有信息并评估其他驾驶者和行人的意图。这基本上是在创建一个能够全面理解周围环境的人工智能。

自动驾驶技术如何改变汽车的使用效率？

自动驾驶技术将大大提高汽车的使用效率，因为目前乘用车每周的平均使用时间仅为10小时，而自动驾驶技术将使汽车像出租车一样工作，可能增加到50到60小时，这意味着乘用车的使用效率可能提高五倍。

人工智能计算能力的增长速度有多快？

人工智能的计算能力似乎每六个月就增长10倍，这是一个前所未有的增长速度。

为什么NVIDIA的市值如此之高？

NVIDIA的市值之所以如此之高，是因为它目前拥有最好的神经网络芯片，这对于人工智能的发展至关重要。

特斯拉在人形机器人领域的计划是什么？

特斯拉正在开发名为Optimus的人形机器人，该机器人能够执行几乎所有人类可以做的任务。

人工智能如何影响生产力？

马斯克认为人工智能将大幅提高生产力，从手工劳动到供应链物流，以及使用先进聊天机器人进行客户服务等方面都将得到改善。

特斯拉如何解决电动汽车长途旅行的充电问题？

特斯拉开发了全球超级充电网络，以确保人们能夫舒适地进行长途旅行，并在任何地方给他们的车充电。

特斯拉的下一代车型将重点放在哪些方面？

特斯拉的下一代车型将是一款成本更低、非常注重自动驾驶的车型。这表明特斯拉正致力于推进自动驾驶技术，同时使电动汽车更加经济实惠。

Outlines

00:00

🚀 对话埃隆·马斯克：AI与自动驾驶的未来

埃隆·马斯克分享了他对人工智能（AI）和自动驾驶技术未来5到10年的展望。他强调了自动驾驶技术的进展，特别是在美国，以及它如何利用数字神经网络和摄像头来实现全自动驾驶。马斯克提到，为了达到这一目标，汽车需要具备相当程度的智能，包括阅读能力和评估其他驾驶者及行人意图的能力。他还讨论了AI在提高乘用车效用方面的潜力，预计未来车辆的使用时间将大幅增加。此外，他谈到了AI在在线服务和计算能力增长方面的快速进展，以及这一进展对市场和技术的影响。

05:01

🤖 优化与AI在生产力提升中的角色

马斯克讨论了AI技术对生产力提升的广泛影响，包括在手动劳动、供应链、物流以及客户服务等方面的应用。他预测了一个通过AI驱动的技术革命，这将是历史上最大的变革之一。马斯克表达了他对AI未来发展的担忧，包括AI计算需求可能带来的电力和基础设施挑战。他还提到了电力需求增长对电压变压器和整体电力供应的影响，预示着未来可能会出现的电力短缺。

10:03

🚗 电动汽车与低压系统的未来

在这一部分中，马斯克探讨了汽车行业中电压系统的变化，特别是从传统的12伏系统向48伏系统的转变。他解释了这种变化如何减少车辆内所需铜材的量，从而提高效率和性能。马斯克还讨论了电动汽车（EV）充电基础设施的重要性，以及如何通过全球超级充电网络解决长途旅行和充电需求。他指出，随着电动汽车和AI技术的同时增长，对电气设备和电力生成的需求将大幅增加。

15:05

🌍 特斯拉未来车型与市场展望

马斯克预告了特斯拉未来的汽车设计方向，特别是关于低成本车型和以自动驾驶为中心的设计。他提到，随着AI技术的进步，车辆驾驶体验将变得更加人性化和直观。此外，马斯克提及了即将推出的特斯拉Roadster，它将提供令人兴奋的性能，例如在1秒内加速到100公里。他强调了特斯拉在创新和伙伴关系方面的成就，并对未来电动车市场的发展表达了乐观态度，尤其是关注欧洲市场中对可负担电动车的需求。

Keywords

💡人工智能

人工智能（AI）是一种模拟人类智能过程的技术，包括学习、推理和自我修正。在视频中，人工智能被认为是未来5到10年内将深刻改变世界的关键技术之一，特别是在自动驾驶技术方面。埃隆·马斯克强调了人工智能在提高车辆自主性方面的潜力，以及它在其他领域，如客户服务和物流中的应用，指出这可能是有史以来最大的技术革命。

💡自动驾驶

自动驾驶技术使车辆能够在没有人类司机干预的情况下导航。在视频中，埃隆·马斯克讨论了其公司在自动驾驶领域的进展，特别是使用数字神经网络和摄像头，而不是雷达。这种技术允许车辆自行导航至目的地，如从家到机场，预示着未来汽车的使用率和效率将显著提高。

💡数字神经网络

数字神经网络是一种模仿人脑结构和功能的计算系统，用于处理复杂的数据输入。视频中提到，使用数字神经网络和摄像头是实现自动驾驶技术的关键。这种技术的应用展示了人工智能如何通过模拟人类的视觉和决策过程来驾驶汽车。

💡优化程序

优化程序是特斯拉开发的人形机器人项目，目标是创建能够执行几乎所有人类可以做的任务的机器人。视频中埃隆·马斯克提到，通过优化程序，他们旨在开发出能够进行复杂任务和提高生产力的机器人，这反映了人工智能技术在自动化领域的广泛应用潜力。

💡计算增长

计算增长指的是人工智能计算能力的快速扩展，视频中提到这一点正在以前所未有的速度增长。埃隆·马斯克指出，人工智能计算的在线增长似乎每六个月增长10倍，这显示了该领域技术进步的速度和规模。

💡NVIDIA

NVIDIA是一家专注于图形处理单元（GPU）和人工智能技术的公司。在视频中，埃隆·马斯克提到NVIDIA因其在神经网络芯片领域的领先地位而市值巨大，强调了高性能计算在推动人工智能发展中的重要性。

💡电力需求

电力需求在视频中被提及为人工智能和电动汽车发展的一个重要挑战。随着这些技术的快速增长，对电力的需求也在增加，包括为高性能计算提供动力和为电动汽车充电。埃隆·马斯克指出，未来可能出现电力供应短缺的问题。

💡变压器

变压器是一种电气设备，用于调整电压以满足不同设备的需求。视频中埃隆·马斯克提到，随着人工智能计算需求的增加，变压器成为了一个瓶颈，因为需要大量变压器来降低电压以供给人工智能服务器。

💡48伏系统

48伏系统是指使用48伏电压的电气系统，与传统的12伏汽车电气系统相比，它能更高效地管理电力。视频中埃隆·马斯克讨论了将汽车电气系统从12伏升级到48伏的好处，包括减少所需的导体（如铜）量，这是向更高效电动汽车过渡的一步。

💡充电基础设施

充电基础设施是电动汽车普及所必需的，包括充电站和其他充电设备。视频中埃隆·马斯克提到了建设全球超级充电网络的重要性，以确保电动汽车用户可以方便地进行长途旅行，这是推动电动汽车接受度的关键因素之一。

Highlights

Cars becoming fully autonomous in the near future using AI and cameras.

Self-driving technology needing to assess intention of drivers and pedestrians.

AI compute increasing exponentially, by 10x every 6 months.

Shortages happening first in AI chips, then voltage transformers, then electricity.

Next generation Tesla vehicle focused on autonomy and lower cost.

Electric cars and AI driving large increases in electricity demand.

Infrastructure needing to keep pace with technology advancement.

Possible to have 0 to 100 km/h acceleration under 1 second.

Automotive moving from 12V systems used for a century to 48V.

48V system allows 75% reduction in copper wiring.

AI progress feels increasingly human-like in how it drives.

Mass adoption of electric vehicles could increase utility 5x.

Lidar not needed for self-driving, cameras and neural nets sufficient.

Demand for AI chips making Nvidia's market cap massive.

Seeing the biggest technology revolution in history.

Transcripts

so

yeah a true changer of the world and

he's continuing to to change all of us

with wonderful Vehicles he has brought

to us a space launch system with as

efficient as nothing before new people

on the on the space uh from us ground

after the space shuttle and I could go

on for half an hour but I don't do that

welcome big hand to Elon

Musk thank you

than oh how are

you uh good

fantastic we have we have had two we

have two days of AI and geni and

whatsoever but you look

real um my head is very big in the

screen and in reality too

suppose no Elon we have had so much talk

about Ai and these two days so we need

your opinion what do you think about the

next 5 to 10 years on AI please give us

a view artificial insemination

artificial intelligence what do you

think about

that but try try audience I say um so um

yes artificial intelligence um is

um is is certainly a profoundly change

the world one of the

most significant ways is

self-driving um

so we've been putting a lot of effort

into

self-driving uh technology for cars um

you you don't see it quite as much in

Europe um because we we first have to

make it work in the US uh before we

increase the complexity of trying to

make it elsewhere in the world um but I

think we're quite close to having the

car be fully autonomous so for example

like right now I could I'm inan Austin

and if I want to drive say to the

airport the car could take me to the

airport with no interventions uh using

and all it's using are digital neuron

Nets other words artificial intelligence

and cameras so there's no lar radar

nothing in fact if you think about how

um humans drive cars humans are

biological neural Nets and we use eyes

so it's the eyes

and bi biological neural Nets the analog

is or

digital the analog for digital uh is

cameras and uh digital neural Nets U

this is working remarkably well the it

has been quite difficult to do this

because turns out that in order to use

in order for this to work the car has

to really be quite fully intelligent

really as as a subset for example has to

learn how to read

everything and and and how to assess

intention among drivers and

pedestrians uh so it's it's it's you end

up

creating sort of a baby artificial

general intelligence to solve this but

this will obviously be extremely

profound um and obv I'm talking about

aut motive here I'll get into the

broader question in a minute um but the

uh the average use of a passenger

vehicle is only about 10 hours per week

out of 168 hours the point at which it

is aut

I think you would see probably a third

of the hours being used because it would

be like a

taxi uh so maybe my guess is like 50 to

60 hours which means that the utility of

a passenger vehicle would increase by uh

a factor of

five um and um this is this is gigantic

um but we'll see so that that's one

motive then there's what people think

have have an experienced online where

you can ask it a question and ask for an

essay or um a picture or to understand a

picture um and that is progressing

rapidly what I'm seeing in terms of AI

compute is I I've never seen any

technology Advance faster than this

the the the artificial intelligence

compute coming online appears to be

increasing by a factor of 10 every 6

months like

now obviously that cannot continue at

such a high rate for forever or or it

will exceed the mass of the universe but

U I've never seen anything like it and

this is why you see nvidia's U market

cap being so gigantic uh because they

currently have the best neural net chips

I mean I think the didn't nidia's market

cap exceed the GDP of Canada or

something recently anyway it was quite

High um so yeah you know it may go

higher who knows

um the chip Rush is um bigger than any

Gold Rush that has ever

existed um so and then there will also

be robots like humanoid humanoid robots

not not just sort of industrial

industrial industrial robots we of

course familiar with you use industrial

robots we use industrial robots everyone

does but they're very they're programmed

explicitly and they they don't walk

around but Tesla with

for example the Optimus program uh is

making a humanoid robot that is capable

of doing almost anything a human can do

um I mean I hope I just hope the the

robots are nice to us you know

um we hope so we hope so too Elon that

the robots will be nice to us and maybe

let's say in this regard because you are

working on optimos um what can we expect

in terms of let's say support which is

going to come maybe also a little bit

with regard to um production um because

in the BOS connected world we were also

discussing application of AI gen into

production uh in order to uh boost

productivity maybe a couple of thoughts

on this one

too

well the side of AI is that I think

productivity will increase

dramatically um so across every field um

whether it be manual labor to supply

chain

Logistics um there's already a lot

of movement to use um chat Bots like

sophisticated chatbots for customer

service for example where they can

answer quite quite complex questions

already but I mean I think

we really are on the edge of probably

the biggest technology

Revolution um that has ever existed you

know there's supposedly sort of a

Chinese curse may you live in

interesting times Well we live in the

most interesting of

times the most

interesting um and for for a while I was

a little bit it was making me a bit a

bit depressed frankly cuz I was like

well will we be will they take over will

we be useless um but then the way that I

reconciled myself to this question was

would I rather be alive to see an AI

apocalypse or not I'm like I guess I'd

like to see

that you know that's a good

thought that's a good thought because

you mentioned it tfold every six months

that's an exp exponential curve right

exponential curves are beautiful but

there's kind of a question where does it

ever end right do you see any end yet or

will it go on for three four five years

like this because then we talk 100 folds

right thousand folds of time yeah that's

why obviously yeah I mean you you can't

you can't grow at that rate without

quickly exceeding the mass of the

universe so clearly it's going to hit

some constraints the constraints that AI

compute and very predictable I've

actually predicted this over a year ago

um so over a year ago the shortage was

chips um CH neural net chips then I it

was very easy to predict that the the

next shortage will be voltage step down

Transformers so because exactly you've

got to feed the power to these things so

if you've got 100 to 300 kilovolts

coming out of your utility and it's got

to step all the way down to you know 6

volts that's a lot of stepping down on

so now we're in step down and this is

why not that funny joke which is that

they need Transformers to run

Transformers um because you know the AI

is like there's this thing called a

transformer in AI um it's a it's a neur

a neural I don't know it's a combination

of sort of neural Nets um and they're

running out of Transformers to run

Transformers then the next shortage will

be

electricity so I think next year you

will see the electricity that they just

can't find enough electricity to run all

the

chips I mean during the chip crisis we

certainly all did have many contacts

fortunately we were able to master it

together of course we all would like uh

to avoid a second chip crisis but with

regard to your com uh to your comment on

on power demand um we all know today

vehicles have uh three uh to four

kilowatt power demand for the

peripherals uh but this is continuously

going up and Tesla um has recently

launched the cyber truck which I believe

is mostly coming with 48 volt uh do you

see this as a let's say further

disruption of automotive industry uh

another technology change within

Automotive yeah I I think it you know

the automotive industry has been living

with 12 volts for a very long time a

century roughly um right I think it it

started off as six bolts and then they

needed a a more powerful um starter

motor and then they doubled it to 12

volts although it's it's actually more

like 13 s

volts you're right absolutely

yeah and and the bus voltage actually

varies a lot for those since there

probably some electrical engineers in

the audience it's like 12 volts is a

very rough approximation um but it's

it's it's really it's really kind of a

arbitrary low voltage um and if you look

at say power ethernet it's running at 50

volts or roughly 48 volts um and we set

it at 4 48 volts because anything really

above that you start to run into

regulatory concerns um and and power

ethernet is like I said roughly at that

level um but the advantage of obviously

quadrupling the voltage is that um you

can substantially reduce how much copper

uh or conductor is used in the car so

you can use very small copper wires for

um you know it's roughly a roughly a

quarter not exactly a quarter but

roughly a quarter as much as as much

copper is needed uh for 48 volts as for2

um so I think it's it's a logical move

um uh that's that's that's really the

next step for the low voltage

architecture for the car I think um

ultimately long-term War cars will be 48

volts mhm it's it's interesting that um

many people love electric vehicles and

want to have them but some are harmed by

a missing infrastructure that would be

the same actually as the AI thing we

will have lots of chips ready but we

will not had the power to to operate

them there's a race between

infrastructure and Technology right now

and you're also certain technology may

win which is a problem because actually

infrastructure should win we should have

all the infrastructure we need is that

correct yeah are you talking about like

sort of charging infrastructure for cars

and that kind of thing um well the the

reason Tesa developed the

sort of global supercharge network was

specifically to address that point the

point you're making which is uh people

need to be comfortable that they can

travel long distances um and charge

their car wherever they go um

unfortunately you you electricity is

almost everywhere um and so it's uh you

know it's not well not all that rare

unless you're really far out in the

countryside or something um

so yeah um you need this super need you

need the charging infrastructure and the

electricity to charge electric cars in

fact the the the simultaneous growth of

electric cars and um AI both of which

need electricity both of which need

voltage

Transformers um actually I think is

creating a a tremendous demand for

electrical equipment and for uh

electrical power

generation maybe one more question Elon

with regard to um new vehicle designs

coming to the market of course you

cannot share with us uh confidential

Secrets uh but but anyways what other

things you believe let's say we will see

over the next years in the market and

maybe just um a small comment in Europe

we see a shortage of affordable battery

electric vehicles um in the offerings uh

because many people in Europe are only

able to spend like 20,000

25,000 uh for their V Vehicles so so

what are the new vehicle types vehicle

designs you you are thinking about at

Tesla well our next Generation vehicle

will be a a low a lower cost vehicle um

but one also that is very focused on

autonomy um like I said

the we really are um getting to the

point where the car can drive itself

well um

when you're in the in the US I would

recommend perhaps giving a you know

trying out our car and see seeing how it

works um it's remarkable you know it it

as the AI gets better and better it

actually feels quite humanlike in the

way that it

drives it does things that you wouldn't

expect a computer to do it's it's not

like a robot dancing you know it's like

it feels smooth and and intuitive um so

our next Generation vehicle um which we

we're quite far along on is is very much

anticipating an

autonomy um and then of course we've got

some fun things as well like on our um

the new Tesla Roadster which will be

able to do 0 to 100 kilometers in under

1

second great thank you very much you

know that's we need that's good news

yeah that's doing some fun stuff too you

know that's that's the that that's the

dessert that's the cherry on the cake

you knowly but it's it's fun you're

going to have the fun stuff too yeah

right well that's very good news because

in the end there's a lot of people here

in the room that still on their list to

buy an electric vehicle and you just

announced something which they probably

want to buy so thank you very much thank

you very much and uh thank you extremely

for 15 years of partnership and work

together we are inspired Always by your

speed attitude or graphicness and things

come out thank you very much to Elon

Musk all right and thank you for your

support very very much appreciated

thanks