Latest / Bad Decisions Podcast / "Superintelligence by 2028", Sam Altman's Warning at the India AI Summit
Transcript
- 0:00What is our beautiful people? Welcome back.
- 0:02To the Bad decisions, show where we come to you with everything
- 0:06in the world of technology, 3D, AI, business, art, you name it,
- 0:11we got it. Every Monday.
- 0:13Every Wednesday. Every Friday.
- 0:14And today's Friday, we are back with another episode.
- 0:17It's funny because when we start researching for this show, we
- 0:21realize there are too many topics are released in between
- 0:2448 hours of the show. We have so many things to cover
- 0:28today, so many exciting news. Pharaohs, are we ready?
- 0:30I swear to God, we sleep for six hours, not 8, not 10-6 hours.
- 0:35And we wake up, we're like, how did so many things happen when I
- 0:39was asleep? This wasn't the case two years
- 0:42ago. This is the case in 2026.
- 0:44So we are trying to take all of the information out there and
- 0:49reduce the noise, get the signal and give it to you.
- 0:52And we've got interesting things today.
- 0:54We've got McConaughey, Matthew McConaughey, the actor who
- 0:58worked in Interstellar. I'm sure you guys have seen that
- 1:00movie. If you haven't checked it out,
- 1:02him and Timothy Xiaomi did an interview or a talk with an
- 1:07audience of actors and they talk about AI.
- 1:09So we're going to start with that, but then we're going to go
- 1:12to. The AI summit in India.
- 1:14With all of the world leaders and tech CEOs after that, we've
- 1:19got fouls report. The media generative report,
- 1:22we're going to cover that and we're going to talk about 11
- 1:25labs, AI agents, insurance, which is quite interesting
- 1:30topic. Also, we're going to talk about
- 1:32Google Gemini 3.1 Pro. Release.
- 1:35That's right. Shall we get started?
- 1:37Let's go. OK, first up, Pratt, let's pull
- 1:39up Matthew Mcconaughey's take on AI.
- 1:42This has been going around on the Internet.
- 1:44Cnni think it's a CNN. Event.
- 1:46I don't think it's CNN. It's a CNN event, I think.
- 1:49CNNN is broadcasting it. If you think so, why?
- 1:52CNN may be for you, but it's not going to be enough to see.
- 1:54It before we start so just pull it back.
- 1:57I believe, I haven't seen the full clip, but I believe one of
- 2:00the actors stood up and asked a question about AI.
- 2:04And this is Matthew Mcconaughey's response to that
- 2:07question. So we do not have the question
- 2:09here, but this is the response to the person in the audience.
- 2:12So let's hear it again from the beginning.
- 2:16It's coming, it's already here. Don't deny it.
- 2:21It's not enough. It may be for you, but it's not
- 2:24going to be enough to sit on the sidelines and make the moral
- 2:27plea, the moral plea that no, this is wrong.
- 2:31It's not going to last. There's too much money to be
- 2:33made and there's it's too productive.
- 2:35It's it's tear. All right, So I say get, get,
- 2:41get your own, your own yourself, voice, likeness, etcetera,
- 2:46trademark, whatever you got to do.
- 2:47So you did that. Yeah.
- 2:48Get own yourself. So when and if when it comes,
- 2:52Not if it comes, no one can steal you, but they're going to
- 2:56have to come to you to go can I or they're going to be in breach
- 2:59and you'll have the chance to be your own agent and go yeah, for
- 3:02this amount or no? OK, it's coming.
- 3:06Is it going to be another category or is it going to
- 3:09infiltrate our categories? Damn sure.
- 3:10Going to infiltrate, infiltrate our category.
- 3:14I think it'll end up. Does it become another category?
- 3:16Will we be in five years having films, the best AI film, the
- 3:20best AI actor? Maybe.
- 3:23I think it might be. That might be.
- 3:24The thing is that becomes another category.
- 3:26I'm not sure it's going to be in front of us in ways that we
- 3:30don't even see it. It's going to get so good.
- 3:32We're not going to know the difference.
- 3:35That's one of the big questions. What we're doing right now is
- 3:38the question of reality. That's it's more hazy than ever,
- 3:43in a very exciting way, I think, but also a scary way.
- 3:49Could you do me a favor and Scroll down, wait, Scroll down a
- 3:52little bit, Scroll down. Let's see this go go down a
- 3:55little bit more. Yeah.
- 3:56So someone was like, understand that he's an investor in the 11
- 4:01Labs AI voice generation company.
- 4:04So I do want to start it this way, Farah, before we begin
- 4:09talking about our opinions and what we think about what Matthew
- 4:12McConaughey just said, it is important to consider that he
- 4:16does have an invested interest in AI adoption because he's an
- 4:20investor in 11 Labs, which we know of, and potentially other
- 4:25AI companies. So that means he is for AI.
- 4:28He does believe that there's a future in which AI will be
- 4:32adopted in a lot of different industries, specifically his
- 4:35industry, which is the entertainment industry.
- 4:38So that's one thing we have to consider before we continue.
- 4:41But Farah, what did you think about it?
- 4:43Do you own yourself now? Do you have trademarks?
- 4:46I I kind of disagree with some of the parts of what he was
- 4:51saying, but I think generally he was accurate when he was saying
- 4:55it is coming. It is too productive.
- 4:57People are making a lot of money with it, But the fact that
- 5:01owning yourself. How do you own yourself?
- 5:03How do you trademark yourself? You can do it in US.
- 5:05Maybe if you are Matthew, maybe you can, but I'm sure if you go
- 5:08to seed Dance now, put Matthew McConaughey, you can get a video
- 5:11out of it. Yeah, this is such AUS or a
- 5:14Western thing like if you're in Iran or India or any of the
- 5:19countries that is not as harsh as the US when it comes to these
- 5:23things like copyrights, then you good luck to you trying to
- 5:27trademark yourself. And also trade marking works
- 5:30when you are a big person like Matthew McConaughey.
- 5:32If you are not anybody in 2026 trademarks marking yourself is
- 5:36kind of useless. It's it doesn't really matter.
- 5:39So his advice of trademarking yourself works for very specific
- 5:44few people like Timothy Chalamet sitting in front of.
- 5:48Him, yeah, I'm sure you can now go to any AI model, put these
- 5:51guys names and then generate image or videos.
- 5:53Yeah, and. Then use it in your UGC
- 5:55commercial. You can go ahead and do that,
- 5:57and by the time they try to do something about it, it's
- 5:59probably too late. Someone has already made enough
- 6:02money and this would have been a problem.
- 6:04This would have been fine. Five years, we are going to get
- 6:07to it now what are you doing? What are you doing, man?
- 6:10He said he owns Simpson. He trademarked it.
- 6:13You can even say bad. Decisions branded onto this
- 6:16show. You can even do bad decisions
- 6:18branded. It's not even our account
- 6:19practice generating with his own.
- 6:20Account No, no. But what I'm saying he can do,
- 6:22he can do bad decisions, branded coffee, and we can use this as a
- 6:26UGC ad for our coffee. Matthew McConaughey approves of
- 6:30our coffee, but even get people to buy it.
- 6:32And by the time he realizes it, it's too late.
- 6:34We've already made our money and the thing?
- 6:36Is but they can sue you for that money.
- 6:38They. Can sue me, but I'm here in the
- 6:40UAE, he's in the US like you know, it, it has to it's the
- 6:43mobile system. And again, now the problem is
- 6:46it's not just one person. Thousands of people can do it at
- 6:50the same time because the barrier to entry to the ability
- 6:54to barrier to scam the barrier to scam is super low.
- 6:58But let's let's go over the whole statement, right.
- 7:01So the statement was that it's too productive.
- 7:04There's too. Much money.
- 7:06There's too much money to be made and that the moral plea is
- 7:11not going to work for a long time.
- 7:13I agree with that. I agree.
- 7:14With all of those statements, let's go back to the
- 7:17agricultural and industrial revolution, right when we went
- 7:21from hunter gatherers to farmers, when we went from
- 7:25farmers to industrialists, when we went from industrialists to
- 7:29technologists. All of these revolutions, when
- 7:32they happen, lots of jobs were displaced, lots of people were
- 7:36unhappy because the change brought about a lot of
- 7:39uncertainty. But ultimately, all of us
- 7:42listening right now and watching whenever you guys, wherever you
- 7:46guys are, we are here because of the sacrifice everybody made
- 7:51before us. Like the fact that we have the
- 7:52ability to even listen to each other on the Internet is because
- 7:56we had all of those revolutions to be able to build, you know,
- 7:59laptops and computers and phones and the Internet.
- 8:02So with every single revolution and evolution, we do have a lot
- 8:07of change that impacts people negatively and impacts others
- 8:12positively. And what he's trying to say is
- 8:15that whenever there is lots of productivity boom in a specific
- 8:19area and lots of money to be made, humans will naturally
- 8:23gravitate towards it. Whether it comes at a sacrifice
- 8:26of other human beings, it doesn't matter.
- 8:28It's always been the case in history.
- 8:30If you just look at history and how people have evolved, how
- 8:33different countries or different kingdoms or different states
- 8:37have won over the other states, It's always been about becoming
- 8:40more productive, becoming faster, more efficient and more
- 8:44effective. And artificial intelligence is
- 8:46just way too big to ignore. So yes, there is a moral plea to
- 8:51be made and he's announcing that, which is good.
- 8:54But what he's saying is, I'm assuming, based on the question,
- 8:57again, I wasn't at the talk. I'm not sure what people were
- 9:00talking about prior to him stating this claim, but I assume
- 9:04a lot of people are just saying, you know what?
- 9:05I'm not going to even learn AI or I'm not even going to bother
- 9:08with it. And he's like, you know what?
- 9:09That's only going to work so long.
- 9:12You need to understand how this works.
- 9:13You need to do something about it because these big companies
- 9:18are naturally going to adopt AI. No, he also mentioned that it is
- 9:23coming to their industry. Yes.
- 9:25So he said in five years we can have a separate category for the
- 9:29best AI actor to for the best AI film or the categories can be
- 9:34mixed. We might have the best
- 9:36supporting role to be an AI role or to be a live action role.
- 9:40But that's very interesting when he mentioned it.
- 9:43Are we going to have a best AI actor or best AI film in the
- 9:48Academy Award? Probably yes.
- 9:50All of these companies are spending money.
- 9:52All of these studios are making money about it.
- 9:55So maybe. Let me let me send something to
- 9:57Pratt right now. Actually, this, this is about
- 10:00this exact topic. So one of our friends, he showed
- 10:05us a video. I was completely unaware of
- 10:07this. Pratt, I'm going to send you
- 10:08this on WhatsApp if you can do me a favor and pull this up.
- 10:12This is a video of a concert, an Iranian concert.
- 10:16You know which one I'm talking about with the AI singer.
- 10:19Oh, it was just here. Yeah.
- 10:21I'm sure a lot of people do not know about this.
- 10:23I want just I want people to see this.
- 10:25So you talked about the category of best AI actor, right?
- 10:29So this singer, well, that's a. Hologram.
- 10:32I would say that's a pre recorded video.
- 10:34Yeah, yeah. But do you understand?
- 10:35OK, So just a context for you guys, that girl that you see
- 10:39doesn't actually exist in real life.
- 10:41This is an LED the screen hologram of her singing a song
- 10:45and dancing to a live audience with another real singer.
- 10:50So there's a real singer's concert and she's Co singing
- 10:53with him in this concert and she has her own fan base on social
- 10:57media. So you can play the audio just a
- 10:59little bit for people to understand, like the cloth
- 11:08simulation and realism is up there.
- 11:15OK, you can stop it. So this is already a thing.
- 11:20You know what I'm saying? In 20252026, when you're saying
- 11:24about the category of AI actors, I'm pretty certain in 10 years
- 11:28time you're going to see something like that.
- 11:30But then how would you? So imagine you will have an AI
- 11:33actor that will consistently play in different movies.
- 11:36Or this AI actor is just made for this movie, and then that's
- 11:40it. It will die.
- 11:42That's a very. That's a very.
- 11:44So can I own an AI actor? So imagine I own Farhad, the AI
- 11:49character. It would be a young billionaire
- 11:52in one movie. It can be a poor athlete in
- 11:54another movie. Yeah.
- 11:55Or do I have to create different Leonardo DiCaprio?
- 11:57Exactly. Like different roles, different
- 11:59matches, but it's still the same character.
- 12:01If Farhad's movie has been the most popular, then you can use
- 12:08him in another film because people become attached.
- 12:11To the actor. Yeah, exactly.
- 12:13Like, what's his name? The guy who's playing the
- 12:16Mandalorian, what's his name? Pedro Pascal.
- 12:19Pedro Pascal. Like he's in every movie now,
- 12:21right? The first time he played in as
- 12:24Oberyn in the Game of Thrones, people got attached to his
- 12:27character. Then they started using him in
- 12:28other movies. Even the movies that you cannot
- 12:30even see his face. Yeah, because people and people
- 12:32watch it because of him. So you can see that what I'm
- 12:35saying is what Matthew McConaughey is talking about is
- 12:38not too far fetched. It's based in realism.
- 12:41And I do agree that at some point you have to stop the moral
- 12:45police if you're concerned about your future.
- 12:49What he's saying makes sense, although he does have an
- 12:52interest, as we talked about, definitely these AI companies.
- 12:56It's important to understand where things are going.
- 12:59So it was interesting because this whole post has been going
- 13:02viral on the Internet, people taking sides and just wanted to
- 13:05show you guys what it's all about.
- 13:07What do we have coming up next? We have the AI summit, which
- 13:10happened for a couple of days in India.
- 13:13We had all the tech leaders, tech CEO Sam Altman was there,
- 13:16Sundar was there, France President Emmanuel Macron was
- 13:20there. We had a lot of celebrities and
- 13:24important figures in this summit so.
- 13:26This was a five day event held in Delhi that was focused on
- 13:30ethical AI development, governance and innovation for
- 13:34societal benefits. As Fallon mentioned, it had
- 13:37world leaders. We're talking prime ministers,
- 13:39presidents and tech CEOs, namely CEOs of Open AI, Anthropic,
- 13:45which is the owner of Claude, Google, NVIDIA, Jensen Huang.
- 13:49Was there No. No, no, he was sick.
- 13:50He didn't go. I saw there was a.
- 13:53Jensen Jensen didn't go because he was sick.
- 13:56Bill Gates didn't go because of Epson files.
- 13:58No, no, that's not true. That's true.
- 14:00Who did you check with? I checked with grok.
- 14:02No, he didn't go, he's supposed. To go, I checked with grok.
- 14:05Search Jensen. OK, Pratt, look.
- 14:06It up. Pratt, Jensen.
- 14:08This is fun. OK, Jensen.
- 14:10Jensen Did Jensen Huang and Bill Gates attend?
- 14:14India's they supposed to. They pulled out last minute.
- 14:16No, no, no. Just say no photos.
- 14:17You will. See.
- 14:18No, no, no. Let me just check here.
- 14:20Fact Check. Fact Check, Fact Check.
- 14:22Realtor then you will see how hard source is reliable or not.
- 14:25OK, so while Pratt looks that up, I'm just gonna quickly prad.
- 14:32You're not typing. Are you typing?
- 14:35OK. All right.
- 14:37Nope, he didn't go. Wait, no, we needed Jensen.
- 14:42Go to the Indian AI Summit. Let's see.
- 14:45You see, no, he was expected to participate, but he canceled his
- 14:50trip. Farhad sources are.
- 14:54Better than AI's Farhad TV network.
- 14:56Wait, don't trust me with all of your news.
- 14:59I think you should pass over the fact that figures here because I
- 15:02have a very credible. Source.
- 15:03I check with, I don't. Use AI, I check.
- 15:06AI hater? No, I am an AI hater.
- 15:10I check what is it. OK, I'm gonna have to talk to
- 15:14Elon Musk after the show. I did Bill Gates.
- 15:17Bill Gates I again. Didn't Epstein file?
- 15:19He didn't go. He was scared.
- 15:21See, I'm so good. I'm so good.
- 15:25So guys, you want to listen to me for the rest of this podcast
- 15:29far as you can do some? Basic stuff.
- 15:32Bill Gates wasn't in the earlier that week for separate meetings.
- 15:37He did. He did afterwards.
- 15:39OK, this is this is crazy. It's so funny because I actually
- 15:42was. Far as any other false news you
- 15:44want to report, no. No, no, no.
- 15:46Fake news, fake news, fake news. So OK, so yeah, Bill Gates and
- 15:50Jensen weren't there, but let's talk about why this whole India
- 15:54AI Summit is. Important If Bill Gates and
- 15:56Jensen were there, they would be in the hand holding meeting the
- 15:59one that they are putting their hands up.
- 16:01I don't know, maybe they didn't want to be part of the hand
- 16:03holding. The hand holding was kind of
- 16:05wholesome. There was a well part of the
- 16:08hand. Holding.
- 16:09We're going to cover that. We're going to cover.
- 16:10That Yeah. So why is this India AI Summit
- 16:13even important? Why should we care?
- 16:16Traditionally, world leaders, world leaders, don't do that.
- 16:23World leaders only held conferences with figures that
- 16:27were shaping the world, like energy industries.
- 16:32Oil, oil and gas only. Oil and gas.
- 16:34But now they understand the importance of technology
- 16:39companies. They understand that so much of
- 16:41the future of the world relies in the hands of the researchers
- 16:46and experts and CEOs of these technology companies because of
- 16:50AI. And therefore we're seeing
- 16:52something for the first time here.
- 16:53I don't think 20 years ago you would have seen all of these
- 16:56world leaders want to hang out with the tech CEOs.
- 16:59They don't even match, like personality wise, they don't
- 17:01even match. And right now we're having this
- 17:03massive AI summit in India with all of these guys talking about
- 17:08how AI is going to change the world, some positive, some
- 17:12negative. We have both sides again, but
- 17:14let's talk about the important things that happen at.
- 17:16The I think one of the most important statement was Sam
- 17:20Altman's talk, which he referred to SGI super intelligence, not
- 17:28AGI anymore. Can we play that on our current?
- 17:31Trajectory. We believe we may be only a
- 17:34couple of years away from early versions of true super
- 17:37intelligence. If we are right, by the end of
- 17:412028, more of the world's intellectual capacity could
- 17:45reside inside of data centers than outside of them.
- 17:49This is an extraordinary statement to make, and of course
- 17:52we could be wrong, but I think it really bears serious
- 17:56consideration. A super intelligence at some
- 18:00point on its development curve would be capable of doing a
- 18:03better job being the CEO of a major company than any
- 18:05executive, certainly me or doing better research.
- 18:08Than our best scientists on our current trajectory.
- 18:12We believe we may be only. Break it down, if super
- 18:16intelligence in a couple of years, we are not talking about
- 18:18decades, we are not talking about five years, we are talking
- 18:20about two or three years from now can perform a better job
- 18:24than any CEO and especially he mentioned himself means any CEO
- 18:28of a big company or any researchers, we are going to
- 18:32have a huge displacement in the job market we are going to have.
- 18:37A huge shift. In how people operate and how
- 18:41people earn income. Because now you will replace the
- 18:45researcher, you will replace the executive team, definitely the
- 18:48consultants. Because who are you consulting
- 18:51if you have a super intelligence running everything for you?
- 18:55I disagree to a certain extent with that statement and I'll
- 18:59tell you why. Yes, you are right.
- 19:03If the things that we talk about go wrong, you will just have
- 19:07displacement if it goes wrong. And that's why I'll explain.
- 19:11So what Sam Altman is saying here is that I'm not denying
- 19:15that it could go wrong. And he's right because we do not
- 19:19actually know what will happen when we have super intelligence.
- 19:22Yes, I'll give an example. Humans, before the development
- 19:27of any tools and technologies, we were just at the midsection
- 19:32of the food pyramid in the jungle, in the Sahara, with all
- 19:35of the other animals around us. We've got these fingers and
- 19:38legs, we've got a good vision, we can see far.
- 19:42But fighting against a lion? How much of a chance do you
- 19:45think you will have right now, barehanded, with no tools, with
- 19:48no? Be.
- 19:49Serious, for this conversation, for you to fight with a lion
- 19:53without any tools, can you win? Absolutely not 20%.
- 19:57Absolutely not 20%. No chance. 5%.
- 20:00Brad, can you call in a lion right now?
- 20:02Let's see. Let's see if you're gonna sit
- 20:03down here and say 20%. Should I ask Gronk about it
- 20:08without the tools that we have today?
- 20:11Spears, guns, rockets, all of these things.
- 20:14It would be impossible for a human to fight against a lion.
- 20:17The only way we could is because we had large numbers, large
- 20:20groups cooperating together. And eventually we built tools so
- 20:23we could fight with other animals.
- 20:25And we built so many tools and so much technology that it
- 20:28allowed us to climb up the ladder of the food pyramid and
- 20:32become the predator, the alpha predator.
- 20:35So all other animals can no longer beat us.
- 20:39We can beat all other animals. We weren't, we were not like
- 20:42that in history. And the way we did that also is
- 20:45through intelligence, right? Yeah.
- 20:48So right now, why everything? See, no, a human cannot
- 20:52realistically win against a line without any tools.
- 20:54See. OK, now we fact checked it.
- 20:57OK with grog. What I'm saying is this.
- 21:00The reason why we are what who we are right now is because we
- 21:04are arguably the more intelligent species because we
- 21:11have the ability to create these tools that we have created.
- 21:15Now we are creating potentially a more super intelligent
- 21:19species. Yes, it is dangerous if we do
- 21:23not have control over it. It is dangerous if it is
- 21:27separate from us. You see, the lion was separate
- 21:30from us. We could never domesticate the
- 21:32lion. The lion was strong and it was
- 21:36separate. But imagine if like Avatar you
- 21:39could interface with the lion. I could go take my hair, connect
- 21:42it to the lion and now I controlled the lion.
- 21:45Now I have the power of the lion and my intelligence I'm now
- 21:48stronger. My IQ added with the lion's IQ.
- 21:51Why Elon is making neural link is because he wants us to
- 21:54interface with the Super intelligence.
- 21:57If Farrad could without the use of laptops, talk in every
- 22:01language. If Farrad could memorize every
- 22:04book he just looked at, then Farrah shouldn't be afraid of
- 22:08super intelligence because you become super intelligent.
- 22:11But I wasn't talking about being afraid or going wrong.
- 22:14I was thinking about the job displacement.
- 22:16No, but if if okay if everyone is super intelligent.
- 22:21But not everyone. Job displacement.
- 22:23The current jobs will no longer will.
- 22:25Change. That's what I said.
- 22:27But if you can speak in every language, if you can look at
- 22:30books and memorize everything, the way we communicate and the
- 22:34things we talk about are on such a different level that you want
- 22:38me and you won't even comprehend it right now, we'll be able to
- 22:41immediately explore other planets.
- 22:43I'm saying when we go to the level that everyone can do that
- 22:46everyone has. Well, we are gonna know we are
- 22:49gonna reach this much faster than what you are referring.
- 22:52To. But I'll tell you something, it
- 22:53only happens if we can interface with the Super intelligence.
- 22:57It won't happen if the Super intelligence is separate from us
- 23:00like the lion. I'm talking about lion as the
- 23:02Super intelligence. As an example.
- 23:03If the lion is separate from us and the lion keeps getting
- 23:07stronger, let's say the ChatGPT models and the growth, they keep
- 23:09getting smarter and we cannot control them and we cannot
- 23:13interface with them, then yeah, that kind of seems like a
- 23:16problem. It's right.
- 23:17No, but that's the whole point of neural link, no?
- 23:20Bro neural link. We are too far from neural link.
- 23:23How long it takes for every person on the planet to have a
- 23:26neural link their head? What Elon is saying is in a
- 23:29couple of years, would you not every person on that, no,
- 23:32because no, it depends on the adoption.
- 23:34Like so many people are not going to take it because they
- 23:36are afraid of the government being able to use your previous
- 23:41episode. Like to read your thoughts?
- 23:42In the previous episode for you guys, you who you have watched,
- 23:46you were pro the variables, the pendant, the eyeglasses.
- 23:49Now you are pro Neurolink. I was I was pro technology in
- 23:54the. Episode 200 Pharaohs will come
- 23:56here with an AR glass, a pendant and neurolink in his brain.
- 24:00I will come the same way. Let.
- 24:01Me ask you another question. OK, I want everyone to answer
- 24:04this question in the comments. Pratt, you are going to answer
- 24:06it here. You're going to answer it here,
- 24:08you know tomorrow. I listen.
- 24:10I completely understand the whole discussion and plea about
- 24:15neural link, being able to read your thoughts and being able to
- 24:19even manipulate you, that will be freaking dangerous.
- 24:22I don't want that. But on the other hand, you're
- 24:26competing with other people, right?
- 24:27When you're uploading content, for instance, you have other
- 24:29creators who are also uploading content.
- 24:31When you go into business, you're trying to win deals that
- 24:33the same businesses are trying to win other businesses.
- 24:37Neural link. If it gives you the power to
- 24:40think faster, to think more logically, to memorize anything,
- 24:44to speak in every language, to communicate more effectively,
- 24:48you will be unbeatable in business.
- 24:50You will destroy your competition.
- 24:52And trust me, by the way, that is the promise.
- 24:55They're not just course they will be.
- 24:57Fixing disabilities is the first thing they're solving.
- 25:01Do you know why I read the report?
- 25:02So Chamath wrote a report, a deep dive into Neurolink.
- 25:09I think it's over 50 pages. Unsub stack.
- 25:11I haven't read the whole thing. I read the summary.
- 25:14He said the reason why they even went for disabilities first is
- 25:16because they did a survey. The rate of adoption was going
- 25:19to be so low because people were against, Yeah, of course.
- 25:22And so they said, if we go ahead and fix disabilities first, no
- 25:26one's against that. Like the difference was 80% said
- 25:29no, 20% said yes. The moment they said what about
- 25:31fixing disabled people, it reversed.
- 25:33So they go with disabled people. But the promise, Elon wrote, is
- 25:37that we're going to increase your cognitive performance.
- 25:40Dude, I know you're taking like cognitive boosting kills, right?
- 25:44That help with memory. I take that too sometimes,
- 25:46right? The alpha.
- 25:47Right. That's very different that that
- 25:49doesn't change my thought patterns and it doesn't have the
- 25:53no, no, no. It's I would say it's very
- 25:55different. My vote is.
- 25:56No, I'm not. Saying it's guys vote.
- 25:58Let's see. OK, Brad, let me.
- 26:00Ask Brad, would you take a Neurolink?
- 26:02Of course, who wants to be average?
- 26:07Listen, you're 2 idiots. Here, Brad.
- 26:09Thank you. Let me be clear, for the sake of
- 26:13the audience, I understand for. The show one person with neural
- 26:17link, one person without neural link.
- 26:19We will see how. I'm going to talk so quickly, so
- 26:23clearly, and then you're going to be lagging behind me.
- 26:26Like, but it's everything. Imperfection is the new
- 26:29authenticity. So, people.
- 26:31I'm not going to be. I'm not.
- 26:32Going to be AI? No, you're going to be a.
- 26:34Robot, you know I'm not. Yeah, I'm what?
- 26:38Your imagination of what I'm going to if you like it's like
- 26:41I'm going to generate pixels in real time and lat No, no, no.
- 26:44I'm telling you the reason why it's going to become
- 26:48increasingly difficult to avoid it is because it gives you an
- 26:51edge. Let me give you guys another
- 26:53example examples. Let me give you guys another
- 26:56example. Fuck you.
- 27:02Let me give you another example. How many people do you know that
- 27:07would want to give up their phones if they could, for just
- 27:11one weekend to feel like they are free and liberated?
- 27:17But they cannot because their livelihood depends on the phone,
- 27:21because the business depends on that phone, because their
- 27:24friends are on that phone. I know at least 80% of the
- 27:29people around me would want to have that life, but they can't.
- 27:33They always tell me that I wish I could put my phone away and
- 27:35just hang out with my family for a weekend.
- 27:37They can't. Their business depends on it.
- 27:39I mean, they can't, but they don't want to.
- 27:41Because the phone gives you an edge.
- 27:44Yes, the social media platforms give you an edge.
- 27:47The phone gave you the addiction.
- 27:48I mean the edge short but I personally it's not a lot.
- 27:51I personally use my phone quite a lot.
- 27:53It's 50% if I. Take your phone.
- 27:56No. If I take your phone away, you
- 27:58lose bad decisions. I would no, no, you could take
- 28:00it away during the weekend. I wouldn't use bad decision.
- 28:02I wouldn't. So, Neurolink, I'll take it out
- 28:04during the weekend. No, you cannot.
- 28:05That's a problem. That's a USB.
- 28:07Plug I'll take. It out.
- 28:08Where are you plugging it? In your brain, imagine that some
- 28:12business comes up with that. So sure, actually that's like
- 28:14the so you can take it off during the.
- 28:16Weekend. If it's I say I will do it.
- 28:21It is going to be, it's going to be in your brain, but the the
- 28:24whole system, maybe the CPU will be taken off.
- 28:28No, thank you. You take it.
- 28:30We will see how it goes. I'm being, but you're not taking
- 28:33this seriously. I'm very I'll tell you why.
- 28:34No, you're just saying no because of the implications.
- 28:36I understand. Listen, I'm on your side.
- 28:39I don't think it's cool if if your thoughts can be manipulated
- 28:44that that part is going to kill it for me.
- 28:46But what I'm saying is, on the other hand, there's going to be
- 28:48so many advantages which makes it almost impossible to not take
- 28:52it, just like the phone. I know you are not part of the
- 28:56PR of the team, but you really sound like the PR of the team.
- 29:01You don't get it. I'm just.
- 29:02Saying I do understand the. Benefits.
- 29:03Sometimes I wish, I wish I could also put my phone aside, but I
- 29:06can't because bad decisions depends.
- 29:09I need to. I need to be able to create
- 29:11content to be able to upload it. Bad decisions would be an
- 29:14entirely different company without social media.
- 29:16Yes, 100% wouldn't exist. Wouldn't exist like the services
- 29:19would exist, but even the marketing of our services.
- 29:21Is not exist. Social media.
- 29:24We'll see. I guess we'll talk to you guys
- 29:26in 2028. The.
- 29:28Episode 700 and we will see who has the Neurolink chip.
- 29:32But guys please put it in the comment.
- 29:34I'm genuinely curious who will do the operation to take
- 29:39Neurolink. OK, Oh, there was something
- 29:42funny in the AI. Imagine.
- 29:43When you play FIFA, I'll just beat you every time like you
- 29:47can. You cannot.
- 29:48Again, you cannot play hanging. Out with you anymore.
- 29:51Imagine even on 4. Kids.
- 29:52Christopher Dolo doesn't have a phone.
- 29:54Think about kids. Christopher Dolo doesn't.
- 29:56Have a phone? You're thinking about your
- 29:57boomer generation. Christopher Dolo doesn't have a
- 30:00phone. Let me, let me go.
- 30:01I'm talking about the new generation.
- 30:02Imagine on Roblox and Fortnite. You're a kid and your moral
- 30:05values are different than ours, right?
- 30:08You're playing Roblox and this guy's decimating you.
- 30:11And you're like, how do you do it?
- 30:13Like you're, you're the leaderboard number 10.
- 30:15I have a neurolink, bro. All the kids are going to get
- 30:19like ask their parents for a neurolink.
- 30:21I'm just saying it's going to be so different moving forward.
- 30:23And how how how I understand it's going to be bad in certain
- 30:27ways, but it's it feels like it will almost become inevitable
- 30:32that we will be cyborgs like Cyberpunk.
- 30:38You played cyberpunk. You know what I'm talking about.
- 30:41Yeah, I'm just saying Boomer generation.
- 30:46Not not going to take part of it.
- 30:48Actually I might become a Cyborg.
- 30:49I don't mind the biotic arm but brain not really.
- 30:52Bionic arm should be fine. Biotic arm but with like proper.
- 30:56Skin on top? No.
- 30:57What's the? Matter.
- 30:59Arnold like Arnold, Terminator 2 Actually it was fun I think.
- 31:05Bionic arms are I. Mean they're already there, but
- 31:07the cool ones that you're talking about that give you more
- 31:09power, Yeah, not yet. We still have.
- 31:12Some there was there. Was a funny scene.
- 31:13In the AI Summit where the Prime Minister of India invited all
- 31:17the tech CEOs to go up on stage and they start holding hands.
- 31:22Look at this. This is so wholesome, so
- 31:24wholesome. This is the most wholesome thing
- 31:25on the Internet. They hate each other and.
- 31:28Play it, play it, but pay close attention to Sam Altman.
- 31:32And anthropomorphy on the day, yeah.
- 31:34Yeah, no, Play it, Pratt, Play it.
- 31:36They will zoom in. They're the only ones not
- 31:38holding. Hands follow follow here Come
- 31:45so. Let me give you guys a little
- 31:47bit background. So Sam Oldtman and Dario Amade
- 31:51were once close colleagues in open AI.
- 31:55Dario Amade the CEO of Anthropic was working in open AI before he
- 31:59left and Co founded with his sister Anthropic.
- 32:04So they knew each other very well for a long time.
- 32:07Of course he left open AI and recently during the Super Bowl,
- 32:11they started making ads to basically shit on each other.
- 32:15Anthropic made four ads which were quite funny actually
- 32:19mocking open AIS using ads in ChatGPT and then also ChatGPT
- 32:24started running codecs 5.3 ads that were sharing on cloud code.
- 32:29So they are fighting non-stop and that's why up on stage they
- 32:33are not holding hand. But Pratt, if you Scroll down,
- 32:35there are some wholesome videos made with AI which made it even
- 32:38funnier. Yeah.
- 32:40Wait, is this the one? No, no, this is the.
- 32:44Don't worry, AI has solved that going on.
- 32:48This is better. Oh shit.
- 32:50And Sam Oldman and Mode Amode are going at it.
- 32:54Oh my God, this is it's crazy, but I understand it.
- 33:00I mean, these guys are rivals and but they could hold times.
- 33:04No, but listen. They're.
- 33:05Rivals, and we're talking about billions of dollars at stake
- 33:10between each company, every single decision they make moves
- 33:14the entire economy true. And this is this is crazier than
- 33:18anything we've ever seen before. So that's really cool.
- 33:22Actually, India's Prime Minister also called for open AI systems,
- 33:26not open AI systems, open AI systems, and also talked about
- 33:31watermarking AI content, which I agree with by the way.
- 33:34We discussed that together. I was reading the other.
- 33:37Day, I think it was Runway CE OS that mentioned that we need to
- 33:40watermark human generated content rather than watermarking
- 33:44AI generated content because there will be so much inner
- 33:48generated content in the world. That would be very hard are to
- 33:50manage them. You should be able to tell the
- 33:52difference exactly. You should be able to.
- 33:54Tell the difference. There was other talks going on
- 33:58at this same exact summit. There was investments that were
- 34:01announced, deals between India and US for AI chips and a lot
- 34:06more. So if you guys want a full
- 34:08breakdown, I recommend you guys go and look up the AI Impact
- 34:12Summit in India. Of course, Jensen Huang wasn't
- 34:15there, Elon Musk wasn't there, Mark Zuckerberg wasn't there, or
- 34:19there was a Meta representative. Alexander Wang was there from
- 34:22Meta. But I'm saying Meta, Mark
- 34:24Zuckerberg himself and Elon weren't there, which is sad.
- 34:27I would have wanted to see them. Next up, we've got Fowls, the
- 34:33state of generative media, volume one.
- 34:37What is this and why is it important?
- 34:40Fowl is a developer API and platform for generative video
- 34:46and image models. So for people who do not
- 34:48understand what that is, think about Higgs field that creatives
- 34:53and freelancers use, where we have the amalgamation of all the
- 34:57different models in one place. This is the same thing but for
- 35:00enterprises because it's the API that different businesses can
- 35:04use. And for instance, if you're a
- 35:06fashion brand and you want to use generative images in your
- 35:10fashion brand like Zara, you'll just go to foul and get the API
- 35:14of Cling or Nana Banana from these guys and then give them
- 35:18money. And then they will deal with all
- 35:20of the credits They have, all of the models you can think about
- 35:24with video generation, image generation, audio generation.
- 35:28And they created this report to tell the world about the
- 35:32adoption of AI. Now, before we dive deep into
- 35:37the report, me and Farhad and Pratt looked at it.
- 35:40And I do want to tell you, after looking at it, reading it, also
- 35:43talking to Grok about it, we found out that the report is
- 35:46actually quite insightful and data rich.
- 35:49And the claims are generally aligned with the industry
- 35:52trends. But it does appear a little
- 35:55optimistic and sometimes self promotional, which makes sense
- 35:59given that foul themselves have products about AI.
- 36:04So if bad decisions was to make a report, I'd probably talk a
- 36:08little bit nicely about myself, you know what I'm saying?
- 36:10So that kind of makes sense. It's not too much.
- 36:12It's just you know, gonna let you guys know about that in
- 36:14advance. So the thing that really.
- 36:17Caught my eyes was the generative media timeline.
- 36:20Yeah, if we open this and when you look at these dates, it felt
- 36:25for me like 10 years ago when we have all of this and it's just
- 36:29been a year, isn't it nuts? It's chat.
- 36:31GPT image 1 is March 2025, one year ago.
- 36:36And we have Sea dance 2 point. O now it feels like OK, not ten
- 36:40years, it feels like 3 years, 2-3 years at least.
- 36:43This is last year where we had the first image generation on
- 36:47ChatGPT. That means the first time you
- 36:49could go and ask for an image to be made, all of those memes that
- 36:53we saw on Instagram, that was last year.
- 36:55That's true because we made. If you remember the flipbook and
- 36:59we made the birds using ChatGPT for our ultimate Unreal Engine
- 37:02program and that was so that was one year ago and then we had
- 37:05Sora 2 in September back in September.
- 37:07That was a huge wave. They were making a lot of funny
- 37:11videos and memes and now we have sea dance 2.
- 37:13Of course, cling 3 point O. But if we progress this much in
- 37:19one year, then we are going to have this show in 2027 March
- 37:24where video generation is going to be.
- 37:27This is why I agree. With Matthew McConaughey, what
- 37:30he said he said sure I understand the plea but this is
- 37:34all moving towards the direction that we cannot ignore and father
- 37:38this is the perfect example of that.
- 37:40If within one year we go from the first image generated on
- 37:43ChatGPT to C Dance 2 point O where we have perfect animation
- 37:48between Tom Cruise and Brad Pitt fighting in LA, where it's so
- 37:52good that Netflix and Disney have to step in and throw a
- 37:56cease and desist letter at Byte Dance, that's when you know this
- 38:01is just moving way too fast and our concept of time is
- 38:04completely off. If you thought this is like 2-3
- 38:06years ago and I thought the same thing, this means it's going
- 38:09really fast. Now you can go ahead and look at
- 38:12the the report itself, but we'll give you guys a summary and what
- 38:16we think about it. The main highlight of the report
- 38:19is the rapid advancement in all AI models and the enterprise
- 38:24adoption and future trends. I'm going to emphasize that
- 38:29enterprise adoption because they stated and this surprised us,
- 38:3388% of organizations deployed AI in at least one function by
- 38:392025. One thing that they forgot to
- 38:42mention that we research and it's probably true is that they
- 38:45have a scope limitation which is large enterprises.
- 38:49When you think about small and medium sized businesses and
- 38:52individuals, I don't think the number is 88%.
- 38:56And the reason we know that is because sometimes me and Fat
- 38:58freak out were like, we got so much is happening in the AI
- 39:01world. Oh my God, it's moving way too
- 39:03fast. And then we go out on the
- 39:04street, we talked to 10 people and they're like, AI is ChatGPT
- 39:08AI. I'm not using it.
- 39:09Like, what are you talking about?
- 39:10So yeah, OK, maybe everyone's touching ChatGPT, but they're
- 39:14using it for basic things. They only have 800 million.
- 39:17User Yeah, 4 How many? Seven billion. 8 billion.
- 39:20Actually, we don't. Have that.
- 39:21If you look at it globally, we don't have that.
- 39:24That's another thing, right? Global.
- 39:25Diversity is probably a concern when you look at this report.
- 39:28I think it's more Western focus. No, I think it's mostly silicon.
- 39:32Valley focus. Yeah.
- 39:33US focus. Yeah, US focus.
- 39:35Probably Europe as well. But what I'm saying is that
- 39:38we're looking at these numbers. The numbers are true, but just
- 39:42understand that they are mostly for bigger enterprises and
- 39:46probably just Western focus. Now let's talk about the numbers
- 39:49because they are fascinating and they are factual.
- 39:52So this report is not just based on Thal's own data.
- 39:56They also looked at other reports made by the likes of
- 39:59McKinsey and so on and so forth. So then the number that actually
- 40:04fascinated me after the dates was that they added a total
- 40:08number of 985 models into the platform in 2025.
- 40:14And I'm just going to give you the numbers here.
- 40:17Pratt, I'm not sure if you can find it, but I did this with
- 40:19Grok. So they added 450 video models,
- 40:23406 image models, 59 audio models, 353-D models, and 35
- 40:31speech models, bringing them to a total of 985 integrations in
- 40:36just one year. Models by different companies,
- 40:38different researchers, probably a lot of open.
- 40:41Source A lot of Chinese models, but wow, that's so many and I
- 40:45would imagine these numbers at least double or triple by next
- 40:48year, probably because everyone. Has access to create and train
- 40:52their own. In their report, they mentioned
- 40:54their. 2026 prediction forecasts multimodal word models enabling
- 41:00interactive 3D environments. I think one of the biggest thing
- 41:03that we are still missing, of course, multimodal models that
- 41:06can generate words and then we can move this and move different
- 41:12elements and also be able to interact with it.
- 41:15Also I think models that can have audio, video integrated
- 41:19together, which now we have, but the audio is not that great yet.
- 41:23I think one of the biggest thing that gives off the latest video
- 41:27models is the audio generated on top.
- 41:28I agree with you even with. C Dance, the audio gave it away.
- 41:31There's like a bit of a distortion that makes it fake
- 41:34and unreal. Let's talk about the industry
- 41:36adoption rates because I think these numbers again within the
- 41:39enterprise world, I think they are accurate.
- 41:41It does seem like that at least. So with advertising, we've seen
- 41:45the biggest adoption, right? The industry adoption rates are
- 41:50mostly with e-commerce and advertising that are driving the
- 41:53growth. That's 56% of them using AI in
- 41:58cases of campaign visuals and ads.
- 42:01We've got entertainment and media with 43% mostly in
- 42:04storyboarding in VFX, which makes perfect sense because we
- 42:08do the same thing, right? We've been using AI in a lot of
- 42:11our storyboarding just to spin up ideas really quickly, make
- 42:14sure we get an idea of what our shots look like before we get
- 42:17into them, either in 3D or whereabouts.
- 42:20If it's a practical shoot, the next one is creative software
- 42:23with 31%, then education 30%, retail and e-commerce 19%.
- 42:29That's your product Photography, virtual try ONS and one at the
- 42:32bottom way more. Why retail is so low?
- 42:35Retail is true. Because they're more
- 42:37traditional. They are more traditional, Zara
- 42:40added. Zara added AI.
- 42:41How so? I think they.
- 42:43Have what we? Made with Google Nano Banana Pro
- 42:46where you can submit your photos and it will have you try on the
- 42:49photos I believe. So check with Zara.
- 42:51And then the the last one at the bottom is architecture and real
- 42:54estate. 8% adoption rate for the 3D renders.
- 42:58And I can confirm this. They are actually right.
- 43:01It's so funny that real estate have been moving super slow.
- 43:04That's also one of the more traditional economies we've
- 43:06worked with real estate companies.
- 43:07And it feels like we are working in 1990.
- 43:09There is such. The reason why this report is
- 43:13incredibly important for people who want to take advantage and
- 43:16have a competitive edge. This report shows you where you
- 43:19can position yourself if you're trying to win clients.
- 43:22Think about it with real estate, right?
- 43:24What is the biggest bottleneck? The moment you want to sell to
- 43:27clients, you need visualizations.
- 43:29The visualizations have always taken a lot of time because you
- 43:33need to go to the artist. They need to look at the actual
- 43:36measurements of the place. It needs to be accurate.
- 43:38And then they need to look at the kind of furniture they need
- 43:40to add in. That also needs to be sort of
- 43:42accurate whenever, whether it's the first time you're building
- 43:45the, the, the compound or if it's an interior design just
- 43:50doing a revamp in the house, then you need to go ahead and
- 43:53create the models, texture them, render the photos.
- 43:57Once you get the photos, a lot of people have unrealistic
- 44:00fucking renders. I've seen some of the real
- 44:01estate renders. They're garbage.
- 44:03They don't look photo real. Some companies do a damn well
- 44:06job to make them photo real. They're of course charged a
- 44:09bomb. And then the next problem is a
- 44:12lot of them don't even animate them, so they just stop at 3DS
- 44:15Max photo renders. They just stop at photos.
- 44:17There's no. Animation.
- 44:18But imagine how much better it would be if you could sell to
- 44:22your clients with videos of exactly what they're going to
- 44:24get in real estate. That will completely change the
- 44:28game. And so for a lot of people who
- 44:29are in real estate, think about where you can position yourself
- 44:32because there's a lot to do and clearly a lot of people haven't
- 44:36adopted it yet. This is the adoption rate in the
- 44:41industries. Is there anything else from the
- 44:44report that you want to talk about?
- 44:45No, not really. I.
- 44:47Think the part that caught my eye was the forecast about the
- 44:513D models. They also mentioned which 3D
- 44:55models they were integrating me. The meshy was there, triple was
- 44:59there. I think they are actually.
- 45:01Triple Sr. was the best one. Apparently for them, yeah,
- 45:03triple Sr. For three.
- 45:05D was the best one for actually I think I have it here if you
- 45:08want me to look at it. Let me see.
- 45:15Yes, this is, is it official Pratt?
- 45:17Is that Zara? Yeah.
- 45:19I think she gave a review. About it.
- 45:21Zara's new try. On feature creates a digital
- 45:24avatar for you. There you go, the one on the.
- 45:26Right. The one on the right is her
- 45:28actually after buying the dress, and the one on the left is what
- 45:31she got on the website before buying the dress.
- 45:35That's incredible. That's incredible.
- 45:38No, but you see the biggest problem.
- 45:40Is the fit. Of course it shows how it looks.
- 45:43No, but sometimes all you need. Is just to get a feel of what it
- 45:46looks on you and you're going to get it.
- 45:48So OK, just just for context in case you missed it, this is
- 45:51Zara's AI virtual try on that you can use to see how you would
- 45:57look in the clothes before you make a purchase.
- 45:59So you probably just submit an image to them of your front
- 46:03angle and side angles and then choose the clothes that you want
- 46:06to buy and they will generate a video of you turning.
- 46:10And then you can make a decision if that dress looks great on you
- 46:13or if the hoodie looks good on you.
- 46:15And you can make a purchase without having to go to the
- 46:17store, which makes the purchasing decision so much
- 46:20easier. And Zara is one of the first
- 46:21ones to adopt this. Anybody, any brand right now,
- 46:25any fashion designer, any clothing designer could go out
- 46:28and use AI to do the exact same thing for their business and
- 46:31they'll be ahead of the game 100%.
- 46:34I remember. Was it 5-6 years ago when a
- 46:36sunglasses company tried the AR try ONS and it worked for that?
- 46:40It works really well. Now you can do it like 10. 1000
- 46:42times better than AR Try on 100. Percent.
- 46:45So that's that. Thank you, Pratt, for pulling it
- 46:48up. OK, the next do you have?
- 46:49Anything on the yes so for image generation.
- 46:52They're talking about how Nano Banana Pro is the best and Sea
- 46:55Dream 4.5 became the best in video generation.
- 46:59The major ones were Cling and VO which makes sense.
- 47:02Audio synthesis was 11 labs which also makes sense.
- 47:05I mean we use all of these. 3D modelling was Tripo Sr. and
- 47:09World models was Genie by Google Genie 2.
- 47:12And that's it for the report. It's very lengthy.
- 47:15Recommend you guys check it out. It will help you understand
- 47:18where things are going. And this is just the first
- 47:20volume, so they're going to be doing this every year, which is
- 47:23going to be interesting when we come back.
- 47:25But one thing to note again as a conclusion is the leap that we
- 47:30made in realism and expressiveness in video, audio
- 47:35and photo in just one year. Just imagine where this is going
- 47:39to be in 2027 when we come back for volume to the next.
- 47:44Thing we're going to cover is a news that was broadcasted on X
- 47:48and when I saw it, it reminded me of movies that we used to
- 47:53watch back in 2000 and 1990s when they would announce a new
- 47:57tech or a news, I would say a side news from the tech that it
- 48:02would really blow your mind. So 11 Labs announced that they
- 48:06are the first company to seek your insurance for AI voice
- 48:10agent. Let that sink in.
- 48:13The AI voice agents can now have insurance, which means if you
- 48:17are connecting to these AI voice agents, if they are your call
- 48:21center or if they are booking something for you or if they are
- 48:23closing sales or deals, they are going to be under a policy which
- 48:28if they do something wrong, you can claim damages.
- 48:32And this happened usually when you have staff, if you are
- 48:34working in a big company, you have a huge call center.
- 48:37You might call someone, they might give you a wrong guidance,
- 48:40and you might you might lose 100. $1000 because of the wrong
- 48:45advice that was given to you and I got really interested and I.
- 48:47Started looking up how the corporate insurance right, it's
- 48:50part of the. Corporate insurance.
- 48:51Yes. And I started looking how they
- 48:54actually do it. And 11 labs went through 5000
- 48:59plus adversarial simulation across security vulnerabilities,
- 49:04data privacy, hallucination, prompt injection and safeties.
- 49:09So they tested all these AI agents with different methods,
- 49:12whether can they get into their heads, can they ask for the
- 49:14password, can they ask for the credit card numbers using PROM
- 49:18injection, using hallucination, using different techniques?
- 49:21And they created I think a guidebook or a guideline and now
- 49:24they certified them. So if you're a company now you
- 49:27are getting an AI call center, you can have insurance either up
- 49:30to a certain amount. So the company that is doing
- 49:33that insurance is AIUC. Apparently they are quite big in
- 49:37States and they have been doing this with other type of
- 49:41underwritings in within the states.
- 49:44This is going to be massive. Let's.
- 49:45Take a look at the video 1st and I'll tell you what I think about
- 49:48it. Intake systems is changing fast.
- 49:52Businesses want to adopt AI, but only if it's safe, secure, and
- 49:56reliable. That's why at 11 Labs, our voice
- 50:00agents are now insurable. In partnership with AIUC, the
- 50:05risks associated with voice agents can be quantified for the
- 50:08first time with a certification, ensuring that agents adhere to
- 50:12more than 50 safety criteria and our agents meet 75% of platform
- 50:18level requirements by default, giving enterprises a head start
- 50:22towards certification and insurance coverage.
- 50:25Trust, safety and financial security are now deployable with
- 50:3011 labs agents. That is straight out of a movie.
- 50:37It's straight out of a movie. I cannot believe it.
- 50:39This is gonna change the game though.
- 50:40You know that right? This is gonna build so much
- 50:42trust for companies to now rely on AI voice agents because the
- 50:48reason they wouldn't prior to this, as you said, is they can
- 50:52ensure humans and they can train humans and they just didn't
- 50:55trust the AIS. But now with 11 labs getting the
- 50:59backing of AIUC 1 certification and saying, you know what, if
- 51:02anything goes wrong, there is insurance.
- 51:05That's how much we believe in it.
- 51:07That's how much the market believes in it.
- 51:09We'll pay you back. This is when your restaurants,
- 51:11this is when your tech companies are going to slowly switch to
- 51:14using AI agents. There was another news that we
- 51:18wanted to talk about Poly AI. Did we talk about that in the
- 51:21last? No.
- 51:22OK, so. Poly AI.
- 51:24Very similar to 11 Labs Does AI voice agents specifically for
- 51:29enterprises, restaurants, hotels, casinos, casinos?
- 51:33And they? Just did their announcement
- 51:35video with Gordon Ramsay and they're showing why their
- 51:38product is superior to any other traditional method of call
- 51:42centers. And so this would be exactly one
- 51:46of those, right? This is exactly the same thing.
- 51:48Recently raised $200 million from NVIDIA and other big
- 51:52companies and they said they are currently handling 500 million
- 51:57calls. There are 500 million, that's
- 51:59huge. But I think if they are handling
- 52:02hospitals, I don't think restaurants will get 500 million
- 52:05calls, but probably big corporation.
- 52:07Yeah, they are being married. Yeah, let's let's say no, no.
- 52:09Problem. Let's take a look at this one,
- 52:12Chateau Ego. Let's see what all the fuss is
- 52:15about, shall we? Welcome to to make a
- 52:19reservation, say reservation reservation ask for.
- 52:23A menu please. Seriously.
- 52:25Sorry, no reservation to make a reservation, say reservation.
- 52:29Madame Suru play reservation. Sorry, please hold your
- 52:35doughnut. This is true.
- 52:41Gordon, what can I do for you? I would love a table.
- 52:44For tonight at your new restaurant.
- 52:46Yes, how did you know my diary? Has an opening at 7:00.
- 52:51Right after your guided meditation session.
- 52:53That's perfect. That's kind of creative.
- 52:55Guests about the guided meditation.
- 52:58Yes, it'll be breathing when you sit down.
- 53:00You're good to go, Gordon. Finally.
- 53:04Someone who listens. OK, pause.
- 53:07It's right there an. Amazing.
- 53:09Ad Amazing ad. Amazing ad.
- 53:11Can I just say? Something.
- 53:13Isn't it funny that all the futuristic AI companies make
- 53:18their ads with the most traditional methods of film
- 53:21making or items that represent nothing futuristic.
- 53:25If an AI uses film, digital, not digital, sorry, film cameras
- 53:30that make it look like those really dreamy old styles of
- 53:35capturing film. Because right now when you watch
- 53:37movies, a lot of them are shot with digital cameras.
- 53:39But some directors love to shoot on film, which is much more
- 53:43complicated, but it's the older way of doing it and it creates a
- 53:46very specific look. Open AI's ads are not AI
- 53:50generated, even though they do have AI generated imagery.
- 53:54They use traditional methods of film making to promote their
- 53:57material. And it's so funny that he even
- 54:00here with Paul the AI, when he dropped the flip phone, he
- 54:03didn't pick up a new iPhone with Paul the AI.
- 54:06He picked up an old, old telephone for the AI
- 54:09conversation. Why do you think they did that?
- 54:11No. Why not an iPhone?
- 54:13Why not a futuristic device to create that?
- 54:16Feeling in you. There's a nostalgic feeling in
- 54:19the phone. In the telephone, Yeah, but it
- 54:22is. But.
- 54:22Why would they use that instead of just the more futuristic
- 54:25thing? Can I explain why that makes you
- 54:27talk about it now? Yes, that's the thing.
- 54:30Because if you. Go back to ads in the 2000s,
- 54:33right? They everyone wanted things to
- 54:35be futuristic. If you remember the Sony ads and
- 54:39the vibes of everything we saw in commercials, it was all
- 54:42futuristic tech. But now that we are in the
- 54:45future, they're trying to go back to the traditional stuff
- 54:48and use the old phone to promote Poly AI, which was quite cool.
- 54:52Why do you guys think they did it?
- 54:54Let us know in the comments. I do want to talk about the
- 54:56actual product itself. I think people, yeah, then we
- 55:01can read people's comment. Actually that's we can.
- 55:04Let us know if you guys. Want us to the comments
- 55:07comments. There's a lot of comments to be
- 55:09made why Poly AI is interesting or AI voice agents.
- 55:14I haven't had that much trouble here in the UAE with calling
- 55:17anyone. I think this place is not
- 55:19understaffed. It is sometimes even
- 55:21overstaffed. Like they make sure you get
- 55:23responses quickly. They have WhatsApp chats.
- 55:26When you go to Abu Dhabi, you can get on WhatsApp and talk to
- 55:29Yas Island about any questions you have.
- 55:31This is not the case when I was in the West, in Canada or even
- 55:35in Malaysia. You do have problems when you
- 55:37want to call a hospital, when you want to call the government
- 55:40company, when you. Want to call the immigration?
- 55:42Call the hospital. You want to call any specific
- 55:45place, even a restaurant. Sometimes you have to wait for
- 55:4730 minutes and you're behind this tune that is driving you
- 55:51insane for 30 minutes and then they hang up the phone on you
- 55:55and they have to call again. I've seen that instance so many
- 55:57times with Farhat that it annoyed us.
- 56:00Poly AI and 11 Labs are trying to solve that specific problem
- 56:05because I know a lot of people are annoyed with call centers
- 56:08and the fact that you have to wait and never get an answer.
- 56:10This could potentially. Solve trained.
- 56:13Yes, sometimes you are connected to someone in your bank.
- 56:16They have absolutely no idea what product you're holding,
- 56:19what is your account, what problem you're having.
- 56:21Imagine every single call is routed to a single server which
- 56:25knows all the history of the call, which knows what was your
- 56:29problem? How many times it that you
- 56:30called your bank said oh have to connect it to a different
- 56:33department, they connect it to a different department, what was
- 56:36the problem? Can you explain again?
- 56:37Then you have to explain everything again.
- 56:39It happened to me dozens of times.
- 56:41I explained the scenario four times.
- 56:43Nobody could help but imagine. Now all of it are consolidated
- 56:47in one place and they can solve your problem.
- 56:50The only problem I have. With this is.
- 56:52Now you have a. Problem.
- 56:54You don't have problem with neural link.
- 56:56You have a problem with voice AI.
- 56:57I don't have problems with neural link.
- 56:59By the way, I never said I. Didn't, but the only problem I
- 57:02have with this is that you lose the one thing that made humans
- 57:06humans. You lose the emotional touch.
- 57:09I don't think you can ask the AI touch.
- 57:12You had the call center. I'll tell you, you cannot.
- 57:15Persuade them you cannot. Persuade them you cannot get the
- 57:19AI to do your favors. You cannot start sweet talking
- 57:22and be like, hey, how are you doing today?
- 57:25I hope you have a beautiful day and then you guys, now you know.
- 57:28Where he gets his restaurant reservation.
- 57:31How he gets his credit cards approved.
- 57:34How he gets his loan. Sweet talking the call center.
- 57:37How to win and influence friends by.
- 57:39Dale Carnegie Like doesn't apply it.
- 57:42Doesn't apply again. You cannot call an AI voice
- 57:45agent and be like, hey, I hope you have a great day.
- 57:47I'll call you back tomorrow if we remember the chat, but it's
- 57:50it's hard coded to follow rules. Yes, I will call you tomorrow.
- 57:54Yes. Oh, actually, you remember I've.
- 57:56Done that before bank. I remember the I've done and.
- 57:58It works I. Remember, I was in a call
- 58:00center. I had to call someone multiple
- 58:02times throughout the year, and I even found out they had their
- 58:05leg broken. And I would call them after six
- 58:07months asking them how their leg was.
- 58:08And the fact that I remembered made them more accommodating to
- 58:13my requests. It wasn't something dodgy, it
- 58:15was just me trying to be nice and getting favors in return.
- 58:18This happens on a daily basis. Yeah.
- 58:20What's the problem? I just AI is hard coded.
- 58:23Not hard coded, but it's. Hard on the rules?
- 58:26No, no, it is hard. There is a.
- 58:27Guardrail, a global prompt for the AI things that you cannot
- 58:31do. If the Table 1 is booked, do not
- 58:34give it to anyone. Sure, with humans.
- 58:35I've seen it in Dude, I've seen this in person in France.
- 58:39We went to a French restaurant fully booked.
- 58:42We spoke in English right the first time the guy was in saying
- 58:45to us that this is impossible, you have to wait two hours.
- 58:48Me, if I went back, we said what if we try to speak French with
- 58:50them? We went back, spoke in French
- 58:52and they became immediately more accommodating and they get off
- 58:56the table after 15 minutes. You cannot do that with AI.
- 58:59There's like what I understand pros and cons pros.
- 59:03And cons. So again, it's.
- 59:05I agree with that. You get.
- 59:06What I'm saying right, that's the one thing that you.
- 59:08Mean it's like there's no way I can pull off my fucking sweet
- 59:11talking voice agents anymore. Poly AI.
- 59:14That's the part that bothers me. That's not what you guys think
- 59:18about right? Expect like fucking s s in the
- 59:22comments. But next thing we got Google
- 59:26Gemini 3.1 Pro released yesterday.
- 59:31It's the same core intelligence that powers Gemini 3, but now
- 59:36scaled for practical applications.
- 59:39There are a few examples in their tweet that they shared.
- 59:42One is creating animation, the other one is creating 3D.
- 59:47It wasn't a 3D environment that they created using this model.
- 59:50You can code that. Yeah, you can code that.
- 59:51So yeah, this is the one that now.
- 59:53They're comparing 3.1 pro with three, you can see the quality
- 59:57has shifted drastically so on the left you're looking at.
- 1:00:02Three pro yes, and on the right it's 3.1 pro and the prompts you
- 1:00:07can read at the very bottom is the same exact.
- 1:00:09Prompt for both sides. Both sides.
- 1:00:12But it's much. Better on 3.1 pro.
- 1:00:14It understands how to. Create a bigger environment.
- 1:00:17It has more detail and of course it's more usable.
- 1:00:19The one on the left seems like a kids drawing.
- 1:00:21The one on the right is something you would like a
- 1:00:23stylized you would kind. Of see on a website, yes.
- 1:00:26And can we look at the? Modeling.
- 1:00:29Yes, this one was very interesting.
- 1:00:32So basically this is a 3D interactive simulation that you
- 1:00:37would have on a website. Yes, right, with particles.
- 1:00:40Yes, pretty cool. But one of the Can I just say
- 1:00:43something? Go ahead.
- 1:00:44Before we went on live for this show, I did one test, 1 test.
- 1:00:49OK, one test doesn't count. And I actually gave it Open AI's
- 1:00:52logo and I asked it to animate it in a minimalistic tech way,
- 1:00:58generate an SVG animation just like the prompts we saw above.
- 1:01:01And I saw the thinking and the reasoning and it blew my mind.
- 1:01:06First of all, it understood that this was opening eyes logo.
- 1:01:10Of course it did. I was like, OK, now I'm going to
- 1:01:12design this. First of all, I need to
- 1:01:13understand what it's made of. And it was like, oh, it's like
- 1:01:16petals of flowers. And I was reading it understand
- 1:01:18all of this and it said, OK, I can use the X&Y coordinates of
- 1:01:21the beginning of each of the petals and try to animate the
- 1:01:24strokes. And then it failed.
- 1:01:25And it said, OK, I failed. I'm going to use another
- 1:01:27mathematical formula to be able to create the petals and animate
- 1:01:30them. And it did that.
- 1:01:32But at the end of it, when I got the SVG animation, even though
- 1:01:35there was a background, everything was animated.
- 1:01:37It fucked up the logo. There was like extra sides to
- 1:01:39the logo. So my first Test failed, and I
- 1:01:42imagine it's the same thing. When you see these demo videos,
- 1:01:45understand that there's always failures that are hidden because
- 1:01:49no one's going to upload failures, right?
- 1:01:51It's a compilation of what works and one of the biggest
- 1:01:54challenges. The UI.
- 1:01:55Oh yeah, the UI was ugly. I really love.
- 1:01:58Google products, especially Gemini's of AI products.
- 1:02:02But the interface. That's why Higgs field is good.
- 1:02:05Higgs field's UI, taking the models and then putting.
- 1:02:08In a good user friendly UI and I say that someone who uses tech
- 1:02:14every single day. If someone is not tech savvy and
- 1:02:18get behind these softwares, they're gonna bang their head
- 1:02:22twice to the wall right next to themselves.
- 1:02:24So we need better. It's built by coders, so yes.
- 1:02:27It's built by coders for coders, yes.
- 1:02:29That's why you need a platform like Hicksfield to come and
- 1:02:32create a wrapper to be able to cater to creatives.
- 1:02:35Yes, but you can see from their post they've created some
- 1:02:38incredible work. I urge every single one of you
- 1:02:41guys, if you're interested, to go ahead and try it out for
- 1:02:43yourself. By the time the next episode
- 1:02:46goes live on Monday, we'll have a lot more time to play around
- 1:02:49with 3.1 Pro and give you guys a more detailed analysis.
- 1:02:54But so far I'm amazed by the reasoning and the fact that it
- 1:02:57does actually create SVG animated files that you can use
- 1:03:01on your website or applications. It's just I haven't done enough
- 1:03:04testing to be able to prove to you how good it is.
- 1:03:09We're going to do that by the next time.
- 1:03:10But this is not the only thing that Google announced.
- 1:03:12They also announced Google Pomeli by Google Labs.
- 1:03:16Pomeli, Google. Pomeli and this.
- 1:03:18Is a project aimed for small and medium sized businesses to
- 1:03:23quickly generate scalable social media campaigns and content.
- 1:03:28What do you actually do? First, you build your business
- 1:03:31DNA by uploading and capturing, having the platform capture your
- 1:03:37color palette, your logo, your visual identity, and then it
- 1:03:41creates editable creatives that you can use on your application,
- 1:03:46on your website, for ads on Meta and on Google, which is quite
- 1:03:52interesting. It is an experimental feature.
- 1:03:54It is not a product that you can confidently say, I'm going to
- 1:03:59use this, I'm going to fire all of my guys and my team.
- 1:04:02We tested it with bad decisions branding.
- 1:04:04Pratt, what was the result? What do you think about it?
- 1:04:07It's pretty decent, but it's not something that we would use, to
- 1:04:11be honest. So the quality isn't there yet.
- 1:04:14This is OK. This is let me tell you why this
- 1:04:17show exists. Pratt, I sent you WhatsApp
- 1:04:19images of people talking about this product.
- 1:04:22Could you pull them up for me really quickly?
- 1:04:24The reason why I want you guys to tune in every Monday,
- 1:04:28Wednesday, Friday is because we are trying to be as unbiased as
- 1:04:32possible. We're going to give you our
- 1:04:33analysis and our analysis could at times be wrong because we
- 1:04:37only have so much time. My analysis Farad's.
- 1:04:39Analysis could at times. Be wrong, but we tried to give
- 1:04:42you guys our full analysis and have you also make your own
- 1:04:46judgments. The problem I have with a lot of
- 1:04:49the people on the Internet is the things they say once they
- 1:04:52immediately see something without any analysis.
- 1:04:55So zoom in here. Holy shit, Google just decided
- 1:04:59to kill thousands of startups. Who?
- 1:05:02Died. Did you try the product?
- 1:05:07Like it? Do you think it's going to
- 1:05:08happen that fast? I don't think so, but it's
- 1:05:12important to understand that things like this are happening.
- 1:05:15So this is what the platform looks like.
- 1:05:16Thank you for pulling this up, Brad.
- 1:05:18You give it your fonts, you give it your logo, your color
- 1:05:20palette, you upload your content in there and you ask it then to
- 1:05:24create the creatives for you. For a lot of the e-commerce
- 1:05:28brands that have no creative lead in their team and all they
- 1:05:32care about is just uploading a lipstick on Meta ads, this would
- 1:05:37work brilliantly. You've worked in ecommerce
- 1:05:39before. Like how much of their content
- 1:05:41is actually moderated? There's only a few I would say
- 1:05:43that really care about the creatives.
- 1:05:46If you go to Alibaba, the photos look ugly, so whatever you get
- 1:05:51out of this would be much better than that.
- 1:05:52Even some of the Amazon shops are still very ugly.
- 1:05:55There you go. Which is why if.
- 1:05:56You look at foul. Also the biggest adoption rate
- 1:05:59was in e-commerce and I believe this is targeted specifically
- 1:06:02there. If you're someone who are who is
- 1:06:06working with clients to create product brand photos, photo
- 1:06:11shoots, you can now use this to your advantage.
- 1:06:14I mean, test it out for yourself.
- 1:06:15You might get completely different results.
- 1:06:17This is going to give you an advantage.
- 1:06:20If you're working with clients, let's say restaurants,
- 1:06:23lipsticks, makeup, all of these brands need photo shoots.
- 1:06:26If you can go ahead and use this and say, hey, if someone's doing
- 1:06:29it for a week, I can do it for three days.
- 1:06:32I can do it much faster because you have tools like this.
- 1:06:34So go ahead and try it out. Could be helpful to you guys.
- 1:06:37That is Google, Pomeli, Farhad. Do we have anything?
- 1:06:41Else no. This will conclude our Friday
- 1:06:43session and we had a lot of topics to cover.
- 1:06:48If you are watching this on YouTube, we ask you to please
- 1:06:52subscribe and like this video. If you are watching it on
- 1:06:55Spotify, it would be amazing if you guys can follow us on
- 1:06:58Spotify and leave us a five star review and if you are watching
- 1:07:01on X, you can follow us on XI. Do want to say one last?
- 1:07:05Thing before we leave, we talked a lot about the adoption and how
- 1:07:09the industries, not industry industries are changing.
- 1:07:14Every single piece of news we looked at today is just a
- 1:07:17testament to that. So please go ahead, do your own
- 1:07:20research, look into these. We have not yet created any sort
- 1:07:25of educational content around these topics.
- 1:07:29Not yet yet. Not yet, as you guys know in
- 1:07:33bad. Decision Studio, we've always
- 1:07:35been creating educational content, trying to teach you
- 1:07:37guys how to use tools, whether that was Blender, whether it was
- 1:07:40Gaussian splats, whether it was AI at times in the beginning
- 1:07:44like 2 years ago, whether it was Unreal Engine.
- 1:07:47If you guys do want to see educational content, I'm talking
- 1:07:51about practical educational. So here we're doing analysis.
- 1:07:54I'm talking about educational content that practically teaches
- 1:07:57you these tools. Let us know in the comments.
- 1:07:59Hopefully we're going to read the comments.
- 1:08:03We're going to read the comments.
- 1:08:04Don't worry. Let us know in the comments and
- 1:08:07we can talk about it there. We'll see you guys again on
- 1:08:10Monday. Thank you so much for watching
- 1:08:13and listening. If there's anybody that you
- 1:08:14think would be appreciating this podcast right here and would use
- 1:08:18the information, please send it over to them.
- 1:08:20And thank you so much again. We'll see you guys later.
- 1:08:23Ciao.