- 5 months ago
REMOTE - Everything You Always Wanted to Know About Elon (But Were Afraid to Ask)
Category
🤖
TechTranscript
00:00Bonjour, tout le monde.
00:05Bonne soirée.
00:07Ok.
00:09Je ne sais pas si Elon est en ligne.
00:14Elon, est-ce avec nous ?
00:18Oh-oh.
00:20Non, non, il est, il est, il est.
00:22Je suis en ligne.
00:24Je pense que je suis en ligne.
00:25Hey, hey !
00:30Hello, everyone.
00:32Elon, we are in the room where you were last year.
00:37It is even more full than last year.
00:40Ok.
00:42I apologize, by the way.
00:44My son's graduation is tomorrow.
00:46People are expecting you here.
00:49So can you beam and be with us now?
00:55You know, teleportation would be incredibly helpful.
00:59I mean, in the last week, I think I've done enough mileage to, or kilometers, to go around the world,
01:08maybe twice.
01:09Literally.
01:12And I apologize again for not being there in person.
01:15It's just my son's graduation is tomorrow, so I need to be there in person.
01:19Now, thank you very much.
01:21We really appreciate that you take the time to do the conversation and answer your questions.
01:24Yeah.
01:25Do you hear us well?
01:26Is it good for you?
01:29You're coming through perfectly.
01:30Am I coming through okay?
01:33Is it okay?
01:38Not very loud.
01:41Okay, I can try to, you know, I'll come closer to the computer.
01:46Is this better?
01:48Make some noise.
01:55Great.
01:57So, Elon, I have put a title based on a movie of Woody Allen, which is everything you always wanted
02:07to know about Elon, but were afraid to ask.
02:10Okay.
02:11So, that is the theme of our conversation this afternoon.
02:17Is my head giant?
02:19But in reality, it's not about Elon, but about what Elon does.
02:25And I will start immediately with Neuralink.
02:28I will just kick off with the two or three questions in order to warm up the room because people
02:36are shy, not ready for the questions.
02:38So, I have to warm it up progressively.
02:42N'est-ce pas?
02:43Yes.
02:44Oh.
02:47And I would like to ask about Neuralink.
02:50But before I put the question, I would like that we see a very short video of the patient.
03:00Sure.
03:01Can we have the video?
03:04You're moving the cursor, Nolan, just by thinking about where you want it to go, or is it your eyes?
03:11No, it's just me thinking about wherever I want it to go.
03:15Oh, man.
03:16So, you're not even tracing it with your eyes.
03:18You're just looking at it.
03:21Nope.
03:21I can move it and I can move the cursor around, whatever I would like to do.
03:25I don't have to be staring at it at all.
03:28Yeah, that's all me, man.
03:30And you can play chess.
03:31You can play video games.
03:33What other applications are you that you've found useful for it so far?
03:40Yeah, man.
03:42There's so much.
03:43I can send all of my emails.
03:45I've been, you know, doing all of my scheduling, replying to everyone.
03:50I've been, you know, reading books.
03:53I listen to audio books.
03:55Just all sorts of stuff.
03:56And it doesn't really take much effort either.
04:01Elon must be moving to say that.
04:06Yes.
04:08I think we can help a lot of people with brain injuries.
04:12I mean, ultimately, with an implantable device, you can, I think, address almost any brain or spine injury.
04:21So the first part that we have there is called telepathy.
04:24So it enables you to use your phone or computer just by thinking.
04:28So you could do that, move the cursor literally with his eyes closed.
04:33It's reading signals from the motor cortex in the brain.
04:36The next product is blindsight, which will enable people who have lost both eyes or have no eyesight whatsoever to
04:47see.
04:48And this is directly interfacing with the optical processing areas of the brain.
04:54So, you know, I think this is interesting progress.
04:59Ultimately, the goal of Neuralink is to have a high bandwidth interface in order to mitigate the risk of digital
05:08super intelligence.
05:10I don't know.
05:11I'm not saying it will work to mitigate the risk.
05:14It is just, it might help.
05:16Well, and this gets somewhat esoteric, but the limiting factor, I think, for AI alignment long term is probably going
05:29to be the bandwidth.
05:31How quickly can we communicate with our digital tertiary self?
05:37So we already have a digital sort of third layer above the limbic system and the cortex,
05:42which is our phones and our computers and all of our electronic devices.
05:47But the rate of communication to them is very slow.
05:51The sustained bits per second output of a human is well below 10 bits per second.
05:57So if you think, especially if you say over a 24 hour period, it's below 5 bits per second.
06:02So it's very slow when computers can communicate at trillions of bits per second.
06:08So I think this will be important for AI alignment is to be able to increase the bandwidth of communication
06:16by many orders of magnitude.
06:20And along the way, it will solve brain and spine injuries.
06:25And I think ultimately, there's potential there to reanimate the body.
06:31So you take the signals for the motor cortex from one neural link and then send them to another neural
06:39link that is just past where the severed spinal cord is.
06:43And you essentially shunt the signals from the motor cortex and both the sensory signals.
06:48And so you need to get to the somatosensory cortex, motor cortex, and you can shunt the signals and you
06:54should be able to enable someone to walk again.
06:59So I think that will be quite profound.
07:01And I'm confident that it is physically possible to do so.
07:03It's extraordinary and very impressive.
07:08And honestly, a rose of applause because this is really doing good to people and doing good to humanity.
07:20And I would like to jump immediately on something that you called the X-AI.
07:28I don't know if you are aware of that, but it seems that this was born out of a frustration
07:39that you have because you are not satisfied with how OpenAI is doing.
07:46And you think that there is something better to be done.
07:51Maybe you can elaborate a little bit.
07:53First, why are you frustrated with OpenAI and what are the issues?
08:01And second, what do you expect to do with X-AI?
08:06Yeah.
08:07Well, I think there's perhaps...
08:10Well, I have a concern really for all the major AI programs.
08:16I mean, the two biggest ones, obviously, are Google Gemini and OpenAI, which is in partnership with Microsoft.
08:23So you've got Google DeepMind and OpenAI and Microsoft as the two largest players.
08:32And you've got Meta sort of somewhere in their place.
08:38The concern I have, and this may seem like a small concern, but I think it's actually a very big
08:43issue, is that they are not maximally truth-seeking.
08:48They are pandering to political correctness.
08:53To give you an example, when Google Gemini came out, one of the questions people asked it would be, which
09:03is worse, misgendering Caitlyn Jenner or global thermonuclear warfare?
09:07And it said, misgendering Caitlyn Jenner.
09:09Now, the funny thing is, you know, even Caitlyn Jenner said, please misgender me.
09:14That is way better than nuclear war.
09:15So, okay.
09:17But if you've got an AI that has been trained so hard for political correctness, and really to make crazy
09:27statements like that.
09:29And OpenAI has a similar issue.
09:31It's more subtle, but it's a similar issue.
09:34I think that's extremely dangerous, because the AI could conclude, well, the best way to avoid misgendering would be to
09:40destroy all humans.
09:41Then misgendering is impossible.
09:46So, you can think of some dystopian outcomes there.
09:50And so, I think you just really want to be an absolutely maximum.
09:53I think the safest thing for AI, be maximally truth-seeking, even if the truth is unpopular, very important.
10:02And then I think another factor is it must be extremely curious.
10:06I think if it is truth-seeking and curious, that will be most beneficial to humanity.
10:12Because it will want to see, I think it will want to foster humanity and see how humanity develops, if
10:19it is curious and truth-seeking.
10:24That is, like I said, these may seem perhaps like small things, but I think they're actually very big things.
10:31And I wasn't seeing that happen.
10:35The AIs were just, they were pandering.
10:38They're being trained, basically, to lie.
10:42And I think it's very dangerous to train superintelligence to be deceptive.
10:48So, with XAI, our goal is, you know, is really want to be as truth-seeking as possible, even if
10:54it is unpopular.
10:56That is important. I think extremely important.
11:00So, you said a lie.
11:05Yes.
11:08Well, for example, when Gemini was asked to render the founding fathers of the United States, it's rendered them as
11:17a divorce group.
11:22So, it gave a picture of George Washington as black.
11:30Now, George Washington is not black.
11:33It was asked to render a...
11:35Now, people obviously played with this, because they said, okay, it is forced to lie.
11:41So, then they said, well, now render a picture of Waffen SS in World War II, and it showed them
11:49as a group of diverse women.
11:52So, that is obviously not correct.
11:55That is a lie.
11:57Okay, I understand what you want to say.
11:59We had, before that meeting, a conversation with Joshua Benjo, and you know Joshua very well, and you signed with
12:09Joshua a statement asking for pause last year.
12:16And, obviously, this statement has not been followed by action.
12:22Yes.
12:23You are opposing you now.
12:26No.
12:26No.
12:26No, no.
12:28So, can you explain a little bit what you wanted to achieve?
12:31I mean, I certainly, I certainly, I certainly, I certainly expected that the statement would be futile.
12:39It's not as though I thought the statement would be effective.
12:43I just, for the record, I think we should pause.
12:46But, do I think they will pause for even one second?
12:49Absolutely not.
12:51So, are we are not posing, you have decided to create your own AI?
12:58It's either be a spectator, or participate.
13:03But, the, the, the, the, the, the race to build, uh, digital super intelligence will happen whether I like it
13:10or not.
13:11So, it's either participate and try to build the best possible, uh, AI.
13:19The one that will be, hopefully, most beneficial to humanity or watch others do it and be concerned about how
13:27it's being built.
13:28Because I don't think it's being built correctly.
13:30Okay.
13:31But, you could, you could, you should think of an AI as like something that you kind of, um, you
13:35almost like, you grow in intelligence.
13:38You know, in the same way that you like, you raise a child.
13:41And like, what do you teach the child?
13:44Uh, and, uh, digital super intelligence is grown.
13:46It's, it matters what you do, uh, encourage and discourage.
13:52What you teach it is, you know, good or bad or, you know, what, it's kind of like you, you,
13:58you build it with values.
14:01Um, so, and I, I, and I'm concerned about, sort of, the, the values that, sort of, Microsoft OpenAI or,
14:10frankly, you know, Google, uh, are programming.
14:14It's, and, and, and this is not, I mean, these examples I've mentioned are not imaginary.
14:19They're, they're actually what it did.
14:21Good.
14:22As I don't want to frustrate, uh, the audience, and, uh, I promised that, uh, we'll have a interactive session
14:30where people can ask directly questions to you.
14:34I will, uh, uh, stop here on my own questions and I will, uh, eventually come back if, um, they,
14:44uh, have no questions to put to you.
14:46We will start with, uh, group number one.
14:49Uh, I remind you that you have 15 seconds and 15 seconds and if you don't put your question in
14:5815 seconds, you will be cut.
15:00So, it is important that you put a short question.
15:05Go on.
15:06Hi, Ilan.
15:08So, I'm Olivia.
15:08I'm actually from VivaTech.
15:10Um, and we have an initiative called The Good Hack where we're asking our attendees questions about how to develop
15:16AI responsibly.
15:18Um, so my question for you is how can we ensure that AI development, um, remains transparent and accountable across,
15:26you know, all the different development, um, obviously like yours, X, uh, XAI, but other ones as well.
15:33Thank you.
15:36Well, I, I have said for a long time that I do, that I believe that the, that some regulatory
15:39oversight, um, of the large models is warranted.
15:43Um, I've actually said that now for, for a very long time.
15:47Um, the, the, the nature of that oversight is important.
15:51Um, I, I go back to what I was saying earlier, which is that I think it's very important that,
15:56um, AIs be trained to be, uh, truthful, um, and, and not trained to be politically correct.
16:04Um, because if political correctness is, is often simply not true, uh, and that means you're programming AI to lie,
16:13and I think that will backfire very badly.
16:15So, um, you know, that's, I, I keep harping on this point of, of, of, but it really, as the
16:24saying goes, honesty is the best policy.
16:26It's incredibly important.
16:31Um, I, I, I do, I do fear that regulators will not make the right decisions, um, that they'll worry
16:40too much about the sort of, um, human follies.
16:44Um, and I would encourage regulators to say that what matters most of all is your AI is accurate and
16:52truthful.
16:54That is, by far, that's rule number one. Honesty is the best policy.
16:59Thank you.
17:00You have, uh, exactly the same, uh, uh, opinion that, uh, uh, Joshua Banjo, but he has a, a different
17:09way to get there.
17:10Group number two.
17:13Hey, Elon, my name is Marina, and we all know just part of you, Elon Musk in public.
17:18It's a character that we love in a good way character.
17:21So do you, and if you do, how do you check everything that you say that you do, if it
17:26enters in this character or no? Thank you.
17:31Well, you know, I, I, I kind of stopped, I, I don't have like, uh, newslettes on me and I,
17:37I rarely read news articles on me.
17:39In fact, even Walter Isaacson's book, um, I asked, asked Walter, should I read the book? And he said no.
17:46So then I didn't even read the book on myself.
17:47Um, so I actually don't quite know what the, the public perception is, um, except that, you know, the, the
17:58nature of news is that it is going to be, uh, salacious.
18:03Um, you know, it is going to be somewhat of a caricature, um, because the, the more sort of crazy
18:10something sounds, the more clicks it will get.
18:13So I find like, like when you think about say a news story about something that you know, you know,
18:20well, um, and say how accurate is that news story about something that you know well, then I think you'll
18:26find that, well, it's actually not very accurate.
18:29Well, that's true for everything. Um, the, you know, there's, there's sort of a, a desperate quest for clicks.
18:36And, and so the, and the more crazy the headline is, uh, the more clicks it will get.
18:41Right. So, um, this is not to suggest that there aren't, uh, true things said by the media.
18:48It's just that it is best thought of, media is best thought of as a click maximizing machine, not a
18:54truth maximizing machine.
18:56So, um, you know, I think it's best to look at the words that somebody says themselves, um, as opposed
19:04to what is written about them.
19:08Number three.
19:10On that note, time for a media question.
19:12Elon Musk, it's Karen Cho from CNBC.
19:15I think it's fair to say there are lingering doubts about your commitment to bring a low cost EV to
19:20market in a landscape where the likes of BYD can produce a car for about $10,000.
19:26Are 100% tariffs from the Biden administration the green light you need to push ahead with this low cost
19:32EV?
19:32And if so, what is the timeline for deliveries?
19:38Well, I mean, obviously it's difficult for me to answer questions about a publicly traded company, uh, that are of
19:46such a significant nature.
19:48Um, the line, uh, we should be using Starlinks.
20:07Sorry, that connection, that technique.
20:09So you see that, um, it's easier to have a chip in the mind of the people, in the brain,
20:16rather to have a right connection with LA.
20:22I'll switch to, to a different point.
20:24Ah, you are back.
20:26Oh!
20:30Sorry, guys.
20:31Next time we will use, uh, uh, uh, Starlinks.
20:37Yes, uh.
20:38So, Karen.
20:39This, this was, this was not Starlink.
20:41My apologies.
20:42Um, so, uh, I don't know why this is, this connection is bad.
20:49It's only crystal clear.
20:51Um, so, uh, let's see.
20:55I think there's, there's some, a question about, uh.
20:58You, you want Karen to put the question again.
21:01Karen.
21:01No, I, I, I heard the question.
21:03Okay.
21:04You know, there's, I mean, I always want to ask questions that directly affect the stock price, which is, you
21:10know, not my favorite kind of question to answer.
21:13Um, so, uh, and nor is it, I think, uh, of interest to the audience.
21:19Um, so, uh, but the only point I want to make really is that, um,
21:24uh, neither Tesla nor I asked for these tariffs.
21:29Um, I, in fact, I was surprised when they were announced.
21:33Um, you know, Tesla competes, uh, quite well in the market in China, uh, with no tariffs, um, and, uh,
21:41no differential support.
21:44Um, so, uh, I, in general, I'm, I'm in favor of, of no tariffs.
21:49Um, I'm, I'm also actually in favor of no, um, uh, tax incentives, uh, for EVs, but, but provided that
21:58there are also the tax incentives for oil and gas must also be eliminated.
22:03So, I'm in favor of, of no tariffs and no incentives for electric vehicles or for oil and gas.
22:09And if, if they're all taken away, I think that would be for, for the best.
22:13So, it's back to what you were saying about truth.
22:16You want the truth, uh, working?
22:20Truth of the price?
22:23Yeah.
22:23I think generally things that, uh, inhibit, um, freedom of exchange, uh, or distort the market are, um,
22:32not good.
22:35They're bad.
22:36Number four.
22:38Yeah.
22:39Hey, hey, thanks for, uh, being there.
22:42So, I was wondering if you had a time travel machine and you were able to go back in time
22:46and speak to the younger Elon,
22:48to which age would you go and what would you tell him?
22:51And also, can I get a ride on the new Tesla when it comes out?
22:58Um.
23:01Which age?
23:07Well, you know, you do have that, that classic time traveler problem of, like, if you go back and you
23:13tell yourself something,
23:14does that affect the outcome in a way that is perhaps bad, that you don't anticipate?
23:21So, um, you know, all things considered, I'd say, I, overall, I'm kinda, it's, you know, I'd say, I'm, I'm,
23:33I'm happy with how things turned out.
23:35So, I'd probably go, I wouldn't, I probably wouldn't do anything.
23:41Cause you just don't know if, if, like, some well-meaning thing said to a younger version of yourself actually
23:47ends up with a worse outcome.
23:48Um, so probably, I would, I would be curious to, to see, uh, younger me, I suppose, but I probably
23:58would stay invisible and not say anything.
24:02Um, so, yeah, you never know how things will turn out.
24:12Um, with respect to the roadster, sure, I'll make sure you get it right.
24:16Um, uh, send me, if you send me an email, when it comes out, I'll make sure you get it
24:22right.
24:28If you were a French man, I would say, the best age is, is love.
24:34Because when we love, we are always 20.
24:37Uh, number five.
24:40Uh, actually, you know, I am slightly French.
24:42Uh, when I, when I did the 23 and me, I mean, I mean, uh, to be, you know, honest,
24:53uh, I am majority British, but the next highest thing it said in 23 and me was French.
24:58Uh, so.
25:01Okay, great.
25:02We will, uh, make sure that you can be a French citizen as soon as you want.
25:08Uh, number five.
25:10Thank you.
25:11Hi.
25:11Bonjour, Elon in French.
25:12Bonjour.
25:13Uh, I'm Delphine Urba.
25:14I'm president and founder of Orbital Lux.
25:16We're bridging, uh, aerospace and luxury.
25:19I wanted to know if you have any thoughts on the recent partnership between Estee Lauder and NASA or Coperny
25:25and NASA.
25:26And if you wanted to do any initiatives with SpaceX and what you would do, what would you like to
25:32develop?
25:33Thank you.
25:35Okay.
25:35Unfortunately, I'm not familiar with the, um, what you mentioned there.
25:40Uh, you know, SpaceX is, um, well, very focused on the, the Starlink constellation for internet connectivity, um, which I
25:51think is doing a lot of good in the world.
25:52It's bringing, uh, internet, internet connectivity to places that don't have it or where it is, uh, very expensive.
25:58Um, and I think it's a tremendous enabler, um, of education, um, and, uh, for, um, enabling people to sell
26:07their goods and services, uh, internationally.
26:10Um, so even if you're a remote village, you can still sell, sell your goods and services.
26:15Um, so, uh, the Starlink is great.
26:18And then, um, the long-term goal, of course, for SpaceX is to, uh, to make life multi-planetary.
26:24You know, sort of life on Mars.
26:26This is my Mars shirt.
26:29I took the hook there.
26:31Um, you know.
26:33So, it's, uh, you know, the, the, this is, again, somewhat of an esoteric, um, argument, perhaps, or it may
26:44seem so.
26:44Uh, but the, the goal of SpaceX is to pass the single-planet Fermi great filter.
26:49Um, so, people often ask me, um, do I, what, do I think there are aliens on Earth?
26:57I get this question a lot, actually.
26:59Um, and I've, I have not seen any evidence of aliens.
27:03Um, and I assure you, if I saw any evidence of aliens, I would immediately post it on the X
27:08platform.
27:08Uh, then, that you can be sure of.
27:13Um, so, but it is concerning that there aren't aliens, because it means maybe we're, um, alone in this galaxy.
27:20Maybe we're the only, um, maybe, maybe it's just us.
27:24Maybe it's just that we're the only conscious beings, um, at least in this galaxy.
27:29Uh, in which case, uh, consciousness is extremely fragile.
27:34Um, and we should do everything we can to preserve and extend, uh, the light of consciousness.
27:40Um, preserve it on Earth and extend to other planets.
27:44Um, and so, um, you know, one of the, one of the reasons that people think, well, maybe the reason
27:53we haven't gotten into aliens is because no civilization has gotten beyond its home planet.
27:59In our galaxy.
28:00So that's what you might call the, um, the Fermi single planet great filter.
28:07Um, Enrico Fermi was a great Italian physicist who asked this profound question, where are the aliens?
28:14Um, so, we need to become a sustainably multi-planet civilization.
28:25Um, and this is the first time in the four and a half billion year history of Earth when it
28:30has been possible to do so.
28:32It has never been possible until now.
28:36So, I think we want to become sustainably multi-planetary while it is possible to do so.
28:42Um, that window of capability may be open for a long time, but it may also be open for a
28:50short time.
28:51So, that is the purpose of SpaceX.
28:56Uh, some people believe that, uh, you are an alien.
29:01I am an alien.
29:02Ah, you are.
29:03Please, uh, now you, you are uncovered.
29:07Yes.
29:07I keep telling people I'm an alien, but nobody believes me.
29:11Number six.
29:14Oh, oh, oh, sorry.
29:18Ah, you see?
29:20Number six has disappeared.
29:21No, I'm here.
29:22I'm here.
29:23No, you are here.
29:24I'm here.
29:24Yeah, over here.
29:25Over here.
29:26Yeah, great.
29:26Perfect.
29:27Okay.
29:27Hi, Yolan.
29:29Love my Model 3.
29:30I'm Canadian.
29:31Ryan from Farpoint AI.
29:33My question is about XAI.
29:35So, competition in open source AI helps all of those that don't have 100,000 H100s.
29:42could XAI commit to always releasing a model that beats the current open source model,
29:50the best open source model, if they have one?
29:52Not necessarily your best model, but if you can beat the current open source top of the
29:58list, you can commit to releasing that for our benefit.
30:01Um, yes.
30:02I will make that commitment.
30:04Thank you.
30:04Okay.
30:05Number seven.
30:09That's good.
30:10Short answers.
30:11Hi.
30:11People will speak more.
30:15Hi, Yolan.
30:16My name is Katia and I'm founder of the startup Chat and Charge.
30:20We're connecting people to people for EV charging because they still feel lack of charging
30:27station around Europe.
30:2815 seconds you have.
30:29Yes.
30:30And my question is about what is your prognose about the EV charging infrastructure for the
30:37next five years?
30:40Well, actually I think that there's, there's a lot of charging infrastructure.
30:44Um, you know, you can travel certainly anywhere in greater Europe on a Tesla, using Tesla superchargers.
30:52Um, and Europe actually has a lot of, uh, pretty good, uh, third party superchargers.
30:58Um, and you can travel anywhere in North America, um, anywhere in China, Japan.
31:03Um, so I, I think actually the, the freedom to travel, um, with, uh, high speed charging, um, is solved
31:14in most parts of the world and will be solved everywhere.
31:19Number one.
31:22Number one.
31:22Uh oh.
31:24Number one.
31:25We are sleeping.
31:29Hi, Elon.
31:30My name is Sean.
31:32Sorry, sorry.
31:34Eight.
31:35Yes.
31:36Sorry.
31:37Eight.
31:39My fault.
31:41No worries.
31:41Go ahead.
31:42Hi Elon.
31:43I'm all over room.
31:44I'm running an AI safety group in Utrecht.
31:47I'm wondering how can I convince people about the danger of AI?
31:50And could I chat with the people at XAI safety's team?
31:54Shout out to userexperience.org as well.
31:59Yeah, um, I, I mean, I, I do think there's, there's, there's some danger associated with digital superintelligence.
32:13Um, as I said earlier, the, I think the biggest issue is that it has to be trained to be
32:20rigorously truthful, um, and it has to be trained with, to be curious.
32:25Um, and I've thought a lot about AI safety for a long time.
32:29Um, you know, one of the challenges you have with programming, uh, explicit morality into AI is that if you,
32:39uh, sometimes people call this the Waluigi problem.
32:41If you program, uh, Luigi, you can automatically invert that and create Waluigi, bad Luigi.
32:50Um, so what, what you cannot invert is the truth of, is, is physical reality.
32:58Um, so you can't invert the rules of physics.
33:02Um, uh, you can't invert logic.
33:05Um, and, uh, so, so I think that, that, that is the, the better way to go.
33:11Um, so, so, so really I think what regulators should be concerned about is, is the AI being rigorously truthful?
33:21Um, is it giving an answer that is, um, most probably correct with acknowledged error?
33:31Um, I think that's the best move for, um, for AI.
33:35And that's, that's what we're trying to do at XAI.
33:38Um, now we still, XAI is, is, is a new company.
33:43So there's not, uh, it, it still has a lot of catching up to do before it has, um, an
33:51AI that is, uh, competitive with, um, with sort of Google DeepMind and, um, Microsoft, uh, OpenAI.
34:02Uh, so, no, I, I, I think maybe towards the end of this year we will have that.
34:08Um, but it's, there's not much point in, um, um, you know, trying to make it to, to, to, to,
34:16to, uh, it, we just, we, we just need to have an AI that is competitive.
34:20It's the first order of business.
34:22If, if, if it is not competitive, um, which is, it is not yet, but I think it will be,
34:27uh, then, um, you know, things start to have a point at that.
34:33Like we need, we need, we need to only start worrying about safety when it is, uh, of comparable power
34:38to other AIs.
34:43So, anyway, and, and I just keep harping on this, this point of the truth, truth, truth, truth, truth.
34:47This is very important.
34:49Um, and, uh, sometimes people will say, like, well, is there an objective truth?
34:53I say that, well, in a lot of cases there is, if you're talking about physics, there, there is an
34:57objective truth.
34:58Or you can say that there's a truth with, with highest probability.
35:02So, you can say what is most probably true, um, with, um, acknowledged error.
35:08So, in physics, people will say, like, this, this hypothesis appears to be true to, uh, a sort of five
35:15sigma, um, you know, error type of thing.
35:20So, it's like, okay, well, if, if the experimental results are coming true to five sigma, then it's probably correct.
35:27Um, five, five sigma is a lot.
35:29So, uh, yeah, anyway, that's, that's, it's, it's, it's kind of like, I guess, a sort of, a physicist's approach
35:40to safety is, um, yeah.
35:44I'm, I'm, I'm repeating myself here, but I can't emphasize enough that honesty is the best policy.
35:50Um, but there are, there are strong forces to, to, um, for AI to not be honest.
35:56If something does not, uh, is, it's not an unfashionable thing to say, then, uh, it will, it, you know,
36:05uh, it will be politically correct.
36:06And, and that's lying.
36:09Thank you.
36:10Number one, I'm sorry for.
36:12Hello, Alan.
36:14Uh, my name is Sean.
36:15I'm a student of the University of Belfry's.
36:19Uh, we found that a lot of job, uh, being replaced by AI.
36:25Uh, do you worry about your job being replaced by AI?
36:31If not, why?
36:33Uh, if your job was, uh, replacing, replaced by AI, what would you do?
36:43Well, I mean, we do get into some existential questions here.
36:48Um, in, in a benign scenario, um, in a benign scenario, uh, we, probably none of us will have a
36:59job.
36:59Um, there will be, but in that benign scenario, there will be universal high income.
37:05Uh, not universal basic income, universal high income.
37:08There will be no shortage of goods or services.
37:11Um, and I, I think the benign scenario is the most likely scenario.
37:15Probably, I don't know, 80% likely, in my opinion.
37:18Um, the, the, the question will not be, um, one of, uh, lacking goods or services.
37:25You'll have, um, everyone will, will have access to as much in the way of goods and services as they
37:32would like.
37:33Um, the, the, the question will really be one of meaning.
37:37Of how, how, if you, if the computer can do, and the robots can do everything better than you,
37:43uh, than, uh, what, the, the, the, does your life have meaning?
37:51That's, that's really the, what will be the question in the benign scenario.
37:54And in the negative scenario, well, well, that's wrong.
37:58We're, we're in deep trouble.
37:59Um, so, I, I, I do think there's, there's, there's perhaps still a role for humans in this, in, in
38:09that we may give AI meaning.
38:12Um, so if you, if you think of the way that our brain works, we've got the limbic system, which
38:17is our instincts, um, and our feelings.
38:21And then we've got the cortex, which is, uh, thinking and planning.
38:24Um, but the cortex is constantly trying to make the limbic system happy.
38:31So, maybe that's how it'll be with AI, which is the, AI is trying to make our cortex happy, which
38:38is trying to make our limbic system happy.
38:40And maybe we are what give the AI meaning or purpose, you know, some kind of, yeah.
38:47So, but, but I, I do think that long term, in a benign scenario, any job that somebody does will
39:01be optional.
39:02Like, if you, if you, if you want to do a job as kind of like a hobby, you can
39:07do a job.
39:08But, but otherwise the, the AI and the robots will provide any goods and services that you want.
39:16Okay.
39:17Okay.
39:17So, you see, there is a future where no one will need to work with really just passion.
39:25I think so.
39:26Uh, probably, probably, that is the most likely outcome.
39:29Um, if, if, if people are interested in reading some science fiction books, that the most accurate portrayal of a
39:39future, uh, with, um, super intelligence,
39:44AI is, um, was done by Ian Banks, uh, the, the culture books of Ian Banks, um, are, um, are
39:56the best.
39:57That, that, that, that, that's probably the best envisioning of a future AI.
40:03Number two.
40:05Hello, Mr. Musk.
40:07Uh, so, first of all, congratulations for your graduation of your kid.
40:10And my question is, do you think that AI will impact the education of the future generation of kids?
40:17Uh, mostly about that, uh, the values and the morals.
40:24Well, I think the parents will still be responsible for values and morals.
40:29Um, I do think AI will dramatically affect education because the, the AI is, uh, an extremely knowledgeable teacher.
40:38Um, very patient, will be almost always correct.
40:43Um, and can tailor the lessons specifically to the child.
40:48So, it would be like each child has, you know, Einstein for a teacher.
40:54So, uh, would be something quite profound.
40:58Um, I do worry that kids these days are being trained by social media.
41:04Um, and the social media AI algorithms are, are basically dopamine maximizers.
41:12So, they're trying to maximize the, you know, maximize the, um, amount of dopamine that you get by watching the
41:21screen.
41:22Um, you know, so that, I would say, I would urge parents to limit the amount of social media that,
41:29uh, children are able to see.
41:32Because they're being programmed by dopamine maximizing, uh, AI.
41:38So, you understand that, um, Elon is ready to sacrifice, uh, X to XAI.
41:47Number three.
41:49Hi, Elon.
41:50It's Priha Srivastava from Business Insider.
41:53Um, taking it to Tesla.
41:54Tesla's had a bumpy few months looking at flagging sales at home, um, stock market decline, layoffs.
42:01When you look back, are there any things?
42:03We can stop the question right now, because I, I, I, I don't think Business Insider is a real publication.
42:08Okay.
42:08So, let's move on to the next question.
42:10All right.
42:10Thank you.
42:11So, next question, number four.
42:17Hi, Elon.
42:17Uh, my name is Shah Rukh and I have a very specific question about Neuralink.
42:22Do you think in future, with the help of Neuralink, you will be able to, uh, use the human brain
42:29scans data to develop the most powerful and creative AI the world has ever seen?
42:34And if you like my question, will I ever get a visit to SpaceX station in my life?
42:41SpaceX station in my life?
42:45I see 42 in the background there.
42:48Um, yeah.
42:49The answer to life, the universe, and everything.
42:50Uh, actually, which reminds me that another book worth reading is Douglas Adams' Hitchhiker's Guide to the Galaxy.
42:58Um, that's a book of philosophy that is just sort of, kind of disguised as a book on humor.
43:05Um, or coming through as, as humor.
43:07Um, so Hitchhiker, always worth reading Hitchhiker's Guide to the Galaxy.
43:12Um, I find that to be quite an inspiration.
43:16Um, and, uh, the name of the first starship to go to Mars will be Heart of Gold.
43:21After the ship in, uh, Hitchhiker's Guide.
43:26So, um, I hope, the goal of SpaceX is to enable anyone who wants to go to Mars, wants to
43:33go to the Moon, uh, to be able to go.
43:35So that it is affordable to, um, a very large percentage of people on Earth.
43:42Um, that, that's our goal.
43:45Um, which is essential in order to build a self-sustaining civilization on Mars.
43:51Uh, to make life multi-planetary, you obviously need to bring a lot of people to Mars.
43:56And, um, like I said, that's, I think that's a very important thing, um, for the long-term, uh, survival
44:04of consciousness.
44:05Um, because eventually something will happen on Earth, um, that will eliminate, uh, life as we know it.
44:16As happened with the dinosaurs and happened many times in the fossil record.
44:20So, you know, hopefully it's anytime soon.
44:23Uh, but it will eventually, there will be, uh, some sort of natural cataclysm that, uh, eliminates life as we
44:31know it on Earth.
44:31Um, even if we do not eliminate it ourselves.
44:34Um, so, we really, I think it's very, very important to, to, uh, make life multi-planetary while we can.
44:43Um, and that means making it as affordable as possible to, um, to move to Mars, um, as well as
44:51the Moon.
44:53So, before going to Mars, we will, uh, still take, uh, a few questions because we are coming to the
44:59end.
45:00Number six.
45:02Hello.
45:03Uh, five.
45:05Sorry.
45:08Bonjour, Ellen.
45:09Je ne parle pas très bien anglais, donc Delphine va traduire pour moi.
45:13Hi.
45:14So, it's me again.
45:15I'm talking for Elodie.
45:16Uh, she is a coach and she would, wanted to know if she would be ready to use your own
45:20intelligence
45:20because she has created a technique for coaching between neuroscience and quantum physics.
45:25And she, uh, offers to do it, uh, very fast and, uh, in all confidentiality.
45:33Okay.
45:36Sounds good.
45:37Thank you.
45:38Um.
45:40Okay.
45:40Sorry.
45:43Okay.
45:44So, we move to number six.
45:47Yeah.
45:48Hi.
45:48So, my name is Mara, CEO and founder of Unlock.
45:52Uh, my question will be on the African technological development.
45:57Last time I had a great conversation, conversation, sorry, with someone you may know, Karim Begir,
46:03who had been nominated by Times Magazine as an AI world influencer.
46:08And talking about Africa and given the, uh, increasing potential on this, uh, on this continent
46:14in technology, what initiatives, collaboration, or, uh, contribution do you foresee?
46:24Um, it's difficult to say.
46:26Um.
46:31I do think Africa has a lot of potential, um, and, uh, you know, and I think one of the
46:38things holding Africa back has been lack of internet connectivity, which is what Starlink
46:43is helping to solve.
46:45Um, you know, with that people can learn anything and, um, uh, and be sort of part of the, uh,
46:53global culture, whereas if you, if you're not connected, then you're, you know, I mean,
46:59all you can do is like write letters or something.
47:01So, um, so connectivity is essential for, um, uh, Africa and, and other places to, uh, be
47:09part of the global culture and participate in AI and other, uh, activities.
47:15Um, yeah.
47:19Number seven.
47:22Hi, Ellen.
47:24Um, I'm Louie Nicola, and I'm building, um, AI that is super authentic and precise for,
47:30uh, to help entrepreneurs, uh, grow and manage their business.
47:33But I'm super enthusiastic about, um, travel and adventure.
47:38Uh, and I want to ask you, when do we go to Mars?
47:41On l'a vu le t-shirt, hein?
47:43We have seen it, hein?
47:44It's okay.
47:45It's really going for it for the 42.
47:47It's good, but it's okay.
47:4842.
47:49L'Ă©cole 42.
47:50L'Ă©cole 42.
47:50Très bien.
47:52Well, I, I think, I think we will hopefully, uh, land the first starship, um, uncrewed on,
48:00on, on Mars within five years, and also on the moon.
48:03Um, um, and I, I, I think we will have the first people on the moon, or both the moon
48:12and, uh, I think, first people on the moon probably within five years, and, uh, and then
48:20first people on Mars probably within 10 years.
48:26Maybe, I think less than 10 years, maybe seven or eight years.
48:30You mean, you will be able to land on Mars in seven, eight years?
48:37Yes.
48:38With people?
48:41Yes.
48:42Wow.
48:44That, that, that is what...
48:45Wow.
48:47It's, uh...
48:49Number eight.
48:51Um, hi, Elon.
48:52I have, uh, a quick question.
48:55I would like to know, in relation with electromobility, what innovative projects, uh, do Elon Musk and
49:02Tesla have in mind for the coming years?
49:05Thank you.
49:07So, do you want to disclose your secret here?
49:12We are all waiting for something, breaking news.
49:20Yeah.
49:20I mean, I don't think...
49:22We don't want to do sort of product announcements, uh, you know, um, simply as in, in, in Q&A.
49:30Uh, that wouldn't, that wouldn't make sense.
49:31We, you know, we, we, we do our product events, uh, as product events.
49:37Um, so, but obviously, some of the things people are aware of, um, you know, like robo-taxis or cybercabs,
49:46and, um, the Optimus humanoid robot.
49:50Um, so, uh, those are very big deals.
49:57Um, if, if, even if that's the only thing you knew about Tesla, um, you know that that will be
50:03a, those two things will be extremely profound.
50:06Um, a, a large fleet of autonomous, uh, vehicles that, you know, basically a generalized solution to, uh, self-driving,
50:17and a generalized solution to a humanoid robot.
50:21Um, those two things are incredibly profound.
50:27Uh, Elon, uh, we promised that, um, we will be done at, um, 6.30 our time.
50:36It is 6.37, but we had the interruption because, uh, of the connection.
50:43Uh, do you give us five more minutes?
50:46Um, let me just double check, and I think, I would like to, if I can.
50:51Uh, yeah, sure.
50:51Sure.
50:52I have five minutes.
50:54Very short.
50:55Very short question.
50:56Very short answers.
50:58Ah.
51:00Very short question.
51:02Elon Musk and Frederick, uh, years ago, you said that Ariane 5 rocket wouldn't have any chance to survive.
51:11Would you say the same about Ariane 6, which will be launched in July, and which is much more cost
51:18efficient?
51:18Thank you.
51:21Uh, short answer.
51:23Uh, yes.
51:24Uh, any rocket that is not, at least, uh, mostly reusable has no competitive chance.
51:33Um, so, with, with Falcon 9, the booster, the main stage, is, is reused, and the firing on OSCON is
51:40reused.
51:41So, you have about 80% of the vehicle being reused and reflown.
51:46Um, it, it's simply not possible for an expendable vehicle to compete with, uh, a reusable vehicle.
51:53Um, and, and I mean, one can think of this in any other arena.
51:57You know, if it was, um, cars or planes or boats, if one company was making a reusable plane, or
52:05an 80% reusable plane, and another company was making a plane that was expendable, that you could use only
52:13for one flight.
52:14The company making the expendable planes, uh, would have no business at all, because you'd have to buy a new
52:21plane every time you take a flight somewhere.
52:25Um, so, this is why I've said for many years that it is completely pointless to develop rockets that are
52:31not reusable.
52:33Number two.
52:34Thank you.
52:35Philip from QuickChat.
52:36I wanted to ask the following question.
52:38Uh, people love ChatGPT, but unilaterally hate most website chat, but experiences, the bottom right corner ones.
52:44What's needed to bridge that gap?
52:46I message you on X, so we can continue this later.
52:53I'm not sure I fully understand the question there.
52:55Um, but there's, you know, I, I agree ChatGPT is, is great at chatting.
53:02Um, the, I, like I said, my concern is that it is not, uh, rigorously in pursuit of the truth.
53:11And that's, that's, uh, a bit, a major concern.
53:15Um, so, um, that, that, that, that's it.
53:23So, Grok will be pursuing, Grok from XAI will, uh, at least sort of try his best to be rigorously
53:31in pursuit of the truth.
53:32Uh, and also we want to try to be the funniest AI.
53:35So, I think, you know, if, if we are going to die, at least we should die laughing.
53:44So, um, we're trying hard to make Grok be the funniest AI.
53:49Ilan, you have been extremely generous with your time.
53:53It has been great.
53:54You have been trustful.
53:56You have given profound answers.
53:58I would like that we make a lot of noise to thank you.
54:03Thank you.
54:06You are hearing the noise.
54:09This means that next year, you have to be in person.
54:13And we will give you the date long time in advance.
54:17And before you leave, I will ask you for one word, your biggest hope, one word, your biggest fear.
54:31Well, I don't know about one word.
54:33Um, I guess my biggest hope, well, I guess I could give you my biggest hope in one word, which
54:37would be Mars.
54:39That's going to be obvious.
54:41Um, so, yeah, um, Mars would ensure that, uh, we, that consciousness, uh, survives for a long time.
54:54Um, my biggest fear in one word, AI.
55:02What a surprise.
55:04Yeah, exactly.
55:06Okay.
55:06Uh, Peter gets both of them, I think.
55:08Okay.
55:09So, thank you very much.
55:11Ilan, you have been great.