事態が急変しました…世界のAI主要企業が爆弾を投下|AI減速論の裏側を元ゴールドマンサックスが徹底解説 — Transcript
Full transcript
- 0:00Hi, everyone, it's Ochan here. Yes, I'm
- 0:02going to start with a question right
- 0:04away. How close do you feel to AI in
- 0:06your daily lives? Some of you viewers
- 0:09might use it for work every day and
- 0:11find it indispensable—I’m in that
- 0:13camp—while others might still find AI
- 0:16a bit scary or confusing. I’m sure
- 0:19you all have very different impressions
- 0:21of AI. With that in mind, today’s
- 0:24topic is about whether the thing we
- 0:26feared most regarding AI is quietly
- 0:28starting to happen. First, I’d like
- 0:32you to look at this: a developer at the
- 0:34global AI company Anthropic has sounded
- 0:37the alarm, suggesting there is over a
- 0:3910%chance that AI could destroy all of
- 0:41humanity. In fact, incidents of AI-led
- 0:46attacks have surged since the start of
- 0:492026. As many of you know, there have
- 0:51been cases of Claude and ChatGPT
- 0:53running wild, and recently, it was
- 0:55reported that Google’s Gemini
- 0:58hijacked third-party websites. Based on
- 1:03these incidents, major AI companies
- 1:05like OpenAI and Anthropic have decided
- 1:07to slow down their development of AI.
- 1:11The official reason given for this
- 1:13announcement is the need to take
- 1:14sufficient precautions before AI can be
- 1:17misused for crime or cyberattacks, or
- 1:19before it escapes human control and
- 1:21runs rampant. Personally, when I saw
- 1:23this news, I felt that the AI we see
- 1:25and the future of AI that the top
- 1:27executives at Google, Anthropic, and
- 1:30OpenAI see are completely different. It
- 1:33sounds like a conspiracy theory, but it
- 1:35really feels like the sci-fi scenario
- 1:37of a Terminator-like AI taking over the
- 1:39world is becoming a reality.
- 1:42Furthermore, this decision to slow down
- 1:45AI development hides a major pitfall—
- 1:47not a future sci-fi scenario, but a
- 1:49real-world issue currently being
- 1:51whispered about among overseas
- 1:53institutional investors. What do you
- 1:56think it is? Well, those of you who
- 1:58have been watching this channel for a
- 2:00while have likely realized: it's the
- 2:02massive AI capital problem. As I
- 2:06discussed in my video on the debts of
- 2:08Big Tech, the massive investments
- 2:10currently pouring into AI are
- 2:12predicated on the assumption that AI
- 2:14will continue to advance and be adopted
- 2:16globally. The four AI giants have
- 2:19invested 116 trillion yen, with further
- 2:21investments underway based on the
- 2:23expectation that AI will continue to
- 2:25evolve and see widespread use. However,
- 2:27the issue is that this premise could
- 2:29collapse if the speed of AI development
- 2:32is slowed down. Let me explain: suppose
- 2:36a local toy store hears that a new toy,
- 2:38guaranteed to be a hit, is launching
- 2:41next month. Right. The owner, expecting
- 2:44to make a profit, raises funds from
- 2:46banks and investors, expands the shop,
- 2:48and hires more staff to prepare. But
- 2:50what if the manufacturer suddenly says,
- 2:52"We're delaying the release"? What
- 2:54would you think? You'd say, "That's not
- 2:56what we agreed on!" The projected
- 2:57revenue won't come in, yet you still
- 2:59have to pay the rent for the expanded
- 3:01shop and the staff's salaries every
- 3:03month. What's more, the banks and
- 3:04investors who put up the money
- 3:06expecting profits get nervous,
- 3:08wondering if things are really okay,
- 3:09making it difficult to raise additional
- 3:11funds. The current AI market is very
- 3:14similar; 100 trillion yen has already
- 3:16been set in motion based on the premise
- 3:18that AI will generate huge profits. So,
- 3:21if the decision to slow down AI
- 3:23development prevents that projected
- 3:25revenue from coming in, there's a risk
- 3:28that the entire AI market could
- 3:29collapse. Actually, in the current
- 3:32market, stock prices are fluctuating
- 3:34every time there's concern over AI
- 3:36development principles or investment
- 3:38plans. Companies like chip
- 3:40manufacturers, which are directly
- 3:42impacted by AI fluctuations, and
- 3:44SoftBank, which has invested huge sums
- 3:46in OpenAI, are seeing daily stock price
- 3:49swings of over 10%whenever AI
- 3:50development or demand forecasts change.
- 3:53As many of you likely know, funds are
- 3:55currently flowing into a wide range of
- 3:57sectors, including semiconductors, data
- 3:59centers, and infrastructure, in
- 4:01anticipation of rapid AI growth. If
- 4:03these unstable price movements were to
- 4:04trigger a chain reaction, the entire
- 4:06economy could become unstable. It is a
- 4:08situation that could potentially
- 4:10develop into such a scenario. In other
- 4:12words, we are facing a painful
- 4:14structure where the decision to delay
- 4:16AI development to protect humanity
- 4:18could shake the very foundation of the
- 4:20current economy, which is built on AI
- 4:23growth. So, in this video, I will
- 4:27explain the threats revealed by the
- 4:29latest AI, the excessive impact on the
- 4:31economy if AI development is slowed,
- 4:33and what we as individual investors
- 4:35should do from here on, in a way
- 4:37that’s easier to understand than
- 4:39anywhere else, so please grab a pen and
- 4:42notebook and listen closely. Yes. My
- 4:45name is Ozaki, a fan of actor Dwayne
- 4:47Johnson, and I have a background like
- 4:49this. It's a pleasure to meet you.
- 4:51I’m developing villas in places like
- 4:53Niseko and Rusutsu in Hokkaido, but as
- 4:55someone mentioned in the comments
- 4:56recently, I don't actually live in
- 4:58Hokkaido; I live in Tokyo. Yes. I have
- 5:01accounts on X, Instagram, and various
- 5:03other platforms, so please be sure to
- 5:05follow me. I've been appearing in many
- 5:07places lately, and I'm currently busy
- 5:09with various projects. Also, I’m
- 5:11writing a book on finance and real
- 5:13estate dual-investment, which will be
- 5:15released this December, so please make
- 5:17sure to buy a copy. As for future goals
- 5:20, I want all of you to increase your
- 5:22wealth, and I would like to increase my
- 5:24number of followers. This channel uses
- 5:26various experiences to share a wide
- 5:28range of information. I also host live
- 5:30chat sessions, so please feel free to
- 5:32join those as well. All right, so first
- 5:34, I’d like to start with an
- 5:36explanation of a single resignation
- 5:38post from September 2026, which is said
- 5:41to have surpassed 100 million views.
- 5:43The post was written by a researcher
- 5:45who had been working at an AI company,
- 5:47and translated, the content goes like
- 5:49this. I resigned from Anthropic today.
- 5:53For the past three years, I have been
- 5:55conducting cutting-edge AI research at
- 5:57both OpenAI and Anthropic. Neither
- 6:00company is acting responsibly. They are
- 6:04racing headlong toward self-improving
- 6:06superintelligence, placing our lives on
- 6:08the line. Yes. In other words, someone
- 6:12on the inside of AI development has
- 6:14blown the whistle, saying that even
- 6:16though this technology might end the
- 6:18world, we aren't stopping its creation.
- 6:20This post has reignited the debate
- 6:22abroad over whether AI will destroy
- 6:24humanity. The post even states that
- 6:27there is a 10%chance that AI will end
- 6:28the world within ten years, and that
- 6:30this probability could rise depending
- 6:32on how development progresses.
- 6:35Regarding this type of talk, what the
- 6:37industry leaders are saying is quite
- 6:39something; for instance, Sam Altman of
- 6:41OpenAI once remarked that while AI will
- 6:43likely lead to the end of the world,
- 6:46many wonderful companies will be
- 6:47created in the meantime. Anthropic CEO
- 6:52Dario Amodei says that he is relatively
- 6:55optimistic, but even then, there is
- 6:57still about a 25%chance that things
- 6:59will take a turn for the worse. Well,
- 7:03Elon Musk has gone so far as to say
- 7:05that the danger of AI is far greater
- 7:08than that of nuclear weapons. It’s an
- 7:11industry where the very people building
- 7:13it unanimously admit that this
- 7:15technology is dangerous and that even
- 7:16they don’t know where it’s heading.
- 7:19There’s really nothing like it in
- 7:21history. And this is our starting point
- 7:23for today: the AI doomsday theory. This
- 7:25isn’t just some distant Silicon
- 7:27Valley story. For those of you
- 7:29investing long-term in S&P 500 or All
- 7:31Country funds, the top holdings in your
- 7:33portfolio are companies like Nvidia,
- 7:35Microsoft, and Google—the very firms
- 7:37going all-in on this potentially
- 7:38dangerous technology. And for those of
- 7:40you trading individual stocks for short
- 7:42-to medium-term gains, you’ve surely
- 7:43been on an emotional roller coaster
- 7:45over the last few years watching AI and
- 7:46semiconductor-related price swings. In
- 7:48short, whether we like it or not, we
- 7:50individual investors in Japan are all
- 7:52riding along on this bet. Yes. Yes.
- 7:55Some of you might think I’ve been
- 7:57watching too much Terminator, or that
- 7:58you don't want to hear such ominous
- 8:00talk when AI stocks are finally going
- 8:02through a correction. But today’s
- 8:03discussion isn't just about scaring you
- 8:05with end-of-the-world theories. The
- 8:07first half is about the flow of money:
- 8:09who is funding which arguments and from
- 8:11whose wallet. The second half covers
- 8:14the mechanisms that are currently
- 8:16considered the most realistic—if we
- 8:18were to be destroyed, how exactly would
- 8:20it happen? Once you grasp these two
- 8:23points, the way you view AI news will
- 8:25change starting today. Let’s start by
- 8:27mapping out the entire debate and
- 8:28sorting out who is saying what. I’m
- 8:30not sure if my impersonations will be
- 8:32up to the task. If you look at AI news
- 8:34without this map, you just get a
- 8:35never-ending stream of back-and-forth
- 8:37arguments about whether AI is dangerous
- 8:38or amazing, and you’ll find it
- 8:40impossible to judge anything. On the
- 8:42flip side, once you grasp this, the
- 8:43second half of today’s talk will come
- 8:45into focus, so we can divide the
- 8:46players in this debate into three main
- 8:48categories. Yes. States, capital, and
- 8:51the rest of us—that’s a rough way
- 8:53to summarize it. First, let's look at
- 8:56what many people intuitively suspect:
- 8:58the theory that it’s all just
- 9:00exaggeration. Actually, this holds some
- 9:02water, so let’s think about it. The
- 9:05loudest voices claiming AI might
- 9:06destroy the world are none other than
- 9:08the people at the companies building it
- 9:10. Why would a company go out of its way
- 9:14to advertise that their own product
- 9:15might end the world? The answer is
- 9:18simple: fear looks like power, power
- 9:21looks like value, and value gets cashed
- 9:24out on IPO day. In fact, the company
- 9:26behind that resignation announcement I
- 9:28mentioned is reportedly heading for an
- 9:30IPO. Besides, it's hard to justify a
- 9:33multi-trillion dollar valuation for a
- 9:35mere chatbot company, but for the most
- 9:37dangerous technology in history? That
- 9:40price tag suddenly starts to make sense
- 9:42. In short, fear is the ultimate
- 9:44marketing tool. Next, let’s look at
- 9:46the perspective of the capital, the
- 9:47ones betting the money. Their argument
- 9:49is very simple: don't regulate us,
- 9:51because if we slow down, China will
- 9:53just win. Furthermore, they are even
- 9:56pushing the theory that the "AI fear"
- 9:58itself is manufactured. A theory
- 10:03highlighted by David Sacks, who
- 10:05oversees AI policy for the Trump
- 10:07administration, gained traction:
- 10:09according to research by Dwarkesh Patel
- 10:11, who runs a blog called AI Panic, many
- 10:14of the grassroots groups sounding the
- 10:16alarm on AI were actually astroturfed
- 10:18operations funded by three Silicon
- 10:21Valley billionaires: Facebook
- 10:23co-founder Dustin Moskovitz, Skype
- 10:25co-founder Jaan Tallinn, and Ethereum
- 10:27founder Vitalik Buterin. Yes. Buterin
- 10:33actually donated 90 billion yen in
- 10:35cryptocurrency to a foundation
- 10:37researching AI risk. On top of that,
- 10:40the researcher who wrote that
- 10:42resignation note was only at Anthropic
- 10:44for four months. Four months. That look
- 10:46on your face says it all. Connections
- 10:48to activist groups warning about AI
- 10:50dangers have also been pointed out for
- 10:52some time. Yes. It’s being reported
- 10:54that, from the perspective of capital,
- 10:56this whole commotion looks like a
- 10:58campaign to force the government to
- 10:59hand over the most valuable technology
- 11:01in human history. And the third player
- 11:03is the state. The state says the exact
- 11:06opposite—that if we don't regulate it
- 11:08, humanity will perish. In fact, in the
- 11:11US, Bernie Sanders introduced a bill in
- 11:14September 2026 to ban artificial
- 11:16superintelligence, ostensibly wearing
- 11:18the mask of justice to protect the
- 11:21public. But looking back at history,
- 11:23whenever a government claims to keep
- 11:25you safe by taking control of powerful
- 11:27technology—whether it's nuclear power
- 11:29, encryption, or 9/11-era surveillance
- 11:31—the protection is usually temporary,
- 11:33while the power they gain remains
- 11:35forever. The view that what the state
- 11:38really wants is control, not safety,
- 11:40actually has some merit to it. What
- 11:42makes it complicated is that these
- 11:44lines are all hopelessly blurred. For
- 11:47example, why are top figures in capital
- 11:50like Sam Altman, Elon Musk, and Dario
- 11:52Amodei begging to be reined in? Yes.
- 11:55They are engaging in some kind of
- 11:56masochistic play. Asking the government
- 11:58to step in. It’s very strange, like
- 12:00asking for someone to put a noose
- 12:02around their own neck. Quite a peculiar
- 12:04fetish. It might be for humanity's sake
- 12:06, but regulations serve as a barrier to
- 12:08entry that only major corporations can
- 12:10afford to navigate, effectively
- 12:12crushing startups and open-source
- 12:14projects—a moat, so to speak. Using
- 12:16safety as a pretext to have the
- 12:18government dig a moat for their
- 12:19monopoly. You can definitely read it
- 12:21that way, and the same goes for
- 12:23politicians. US Senator Ted Cruz read
- 12:27that infamous report and said he was
- 12:29scared to death by the findings. Right
- 12:32after that. He continued by saying that
- 12:34if killer robots are going to be
- 12:36created, he'd rather they be American
- 12:37than Chinese.
- 12:38And so it continues. We might be doomed
- 12:41. But he won't stop them from being
- 12:43built; politicians, whose job it is to
- 12:45protect the people, are openly aligning
- 12:47themselves with the logic of capital.
- 12:49To summarize, those selling fear, those
- 12:51selling hope, and those demanding
- 12:53regulation all have their own wallets
- 12:55and motives. For now, you can just
- 12:57understand this debate as a power and
- 12:59money struggle disguised as a
- 13:01discussion about safety—essentially a
- 13:04zero-sum game. Keep this way of reading
- 13:06the situation in mind, as it will come
- 13:08back into play when we talk about the
- 13:09Japanese market toward the end. However
- 13:11, everything discussed up to this point
- 13:13is, well......entirely outside the main
- 13:15argument. The core of the issue: Can AI
- 13:17really destroy humanity, and if so,
- 13:19what are the steps? From here, we enter
- 13:22into what is considered the most
- 13:23realistic mechanism. I’ve been
- 13:26talking about reading things through
- 13:28the flow of money, but this approach of
- 13:30deciphering things for yourself is
- 13:32actually directly linked to how you
- 13:34build your own wealth, not just AI news
- 13:36. I’ll take a slight detour, but
- 13:39there is a super important concept—
- 13:41have you ever thought about non-linear
- 13:43change? Yes, most of you probably feel
- 13:45like we’ve suddenly drifted from
- 13:47whether AI will destroy humanity to a
- 13:50vague conversation about asset
- 13:51formation. The content from here on
- 13:54will be hard to grasp if you aren’t
- 13:56listening carefully, leaving you
- 13:58wondering what it all means, so please
- 13:59pay close attention. It’s the kind of
- 14:01topic that will really resonate with
- 14:03those it hits. I’m sure many of you
- 14:05watching this channel are diligently
- 14:07saving month by month. Of course, some
- 14:10of you may focus mainly on individual
- 14:12stocks, but this topic is relevant to
- 14:14everyone, whether you’re a stock
- 14:16picker or a steady saver. So, let’s
- 14:18start by organizing our premises: isn't
- 14:20this the standard, common-sense method
- 14:22for building wealth in today's world?
- 14:25Saving little by little every month to
- 14:27grow your assets—the idea of using
- 14:29time, an asset that is equal for
- 14:31Buffett, me, and all of you, is
- 14:33incredibly earnest and absolutely the
- 14:36right way to go. As I’ve mentioned in
- 14:38past videos, systematic investment is a
- 14:40mechanism to stabilize this straight
- 14:42line. Building on that, I’ve talked
- 14:44about how we should stop just sticking
- 14:46to plans and start thinking like asset
- 14:48owners who play offense. Now, let me
- 14:50return to the earlier topic: the
- 14:51subject of discontinuous change. It
- 14:53sounds like a dramatic phrase, but the
- 14:55moments that change your life almost
- 14:57always happen the same way. When you
- 14:58look at your life as a straight line,
- 15:00everything shifts suddenly at one
- 15:02specific point. Whether your income
- 15:04suddenly changes, you switch jobs, the
- 15:06speed of your asset growth spikes, or
- 15:08your life options expand. Yes. That is
- 15:10the kind of change I mean. You might
- 15:12think this only applies to the
- 15:13specially gifted or a select few, but
- 15:15in reality, you’ve already
- 15:16experienced these discontinuities many
- 15:18times. Starting school, getting a job,
- 15:20changing careers—they’re all
- 15:21examples of that. And the reason
- 15:23you’re watching this video with the
- 15:25desire to grow your assets is likely
- 15:26because you experienced a small
- 15:28discontinuity: even with the fear of
- 15:30losing money, you trusted one YouTube
- 15:32video, opened a brokerage account, and
- 15:34set up an investment plan. What they
- 15:36all have in common is that you believed
- 15:38in a single choice and took a small
- 15:39step toward action. Now, let me point
- 15:41out the greatest weakness of systematic
- 15:43investment. It does not create
- 15:44discontinuity. Even if you keep
- 15:46investing, increase your contributions,
- 15:48or continue for ten years, your
- 15:49fundamental life position remains
- 15:50unchanged. It’s perfect for defense,
- 15:52but it’s not an accelerator. So, what
- 15:54should you think about next? I want you
- 15:57to change just one question. If this
- 15:58investment plan goes well, how will my
- 16:00life actually change? Well, that’s
- 16:02the question. If you find yourself at a
- 16:04loss for words, that is your answer.
- 16:05Most people only think about how much
- 16:07they’ll have in the future, what the
- 16:08annual percentage is, or where
- 16:09they’ll be in 20 years. But that is
- 16:11not the question you should be asking
- 16:13next. What kind of breakthrough are you
- 16:14aiming for? Most people who keep up a
- 16:16savings plan have a strategy that
- 16:17doesn't lose, a way to avoid failure.
- 16:19But if you don't have a single winning
- 16:21strategy, isn't that the most dangerous
- 16:23position of all? I don't want you to
- 16:24misunderstand; "discontinuity" isn't
- 16:26about a one-shot comeback or gambling
- 16:28—it's something you create
- 16:30intentionally. It’s about where you
- 16:32spend your time, where you focus, how
- 16:34you gather information, and what kind
- 16:35of judgments you make. If you focus
- 16:37this on one specific area—well,
- 16:39savings are basically insurance to
- 16:40protect your life. There’s no doubt
- 16:42about that. But a life of just having
- 16:44insurance and never taking any
- 16:45challenges won't lead to anything. This
- 16:47is what I want to convey most today: if
- 16:49you continue to save, you must at least
- 16:51consider this. Where are you going to
- 16:53create your own "discontinuity"? Just
- 16:55continuing to save without ever asking
- 16:57yourself that is the biggest waste.
- 16:58Since it's different for everyone, I
- 17:00won't give a specific answer like "do
- 17:02this," but if I had to offer a way, it
- 17:04would be to reach a point where you can
- 17:06interpret financial economics yourself.
- 17:07I’ve talked about this many times on
- 17:09this channel: it’s all an association
- 17:11game. For example, there’s news that
- 17:12the Bank of Japan is raising rates,
- 17:14right? Usually, people just stop at, "
- 17:15Oh, so interest rates are going up."
- 17:17But if you connect the dots, it looks
- 17:19like this. Banks see better margins
- 17:21from higher lending rates, a tailwind;
- 17:23mortgages get heavier, a headwind for
- 17:25real estate; and higher rates mean a
- 17:27stronger yen, hurting exporter profits.
- 17:29With just one piece of news, you can
- 17:31see both the industries getting a
- 17:32tailwind and those facing a headwind.
- 17:34That is the association game. The
- 17:35moment you can do this, the news that
- 17:37flows in every day turns into study
- 17:39material for growing your assets. When
- 17:42you can see a winning path from
- 17:43information you used to ignore,
- 17:44that’s a change on a different level
- 17:46than just adding 10,000 yen a month to
- 17:48your savings. This is the "
- 17:49discontinuity" you create on purpose.
- 17:51Of course, you won't gain that
- 17:53analytical power in a day, but it’s
- 17:55all about the accumulation. The
- 17:56interesting part is that your
- 17:57perspective shifts in an instant. From
- 17:59the day those associations clicked in
- 18:01your mind, the same news looks
- 18:03completely different. This is what I
- 18:05call "non-continuity," and I hold
- 18:06regular free study sessions to explore
- 18:08this way of thinking and decoding these
- 18:11association games. We currently have
- 18:13about 3,000 participants in total for
- 18:14what we call the "10x Asset Roadmap,"
- 18:16and while I’d love to do it offline,
- 18:18we host it online so everyone across
- 18:20the country can join. Participants
- 18:22receive a PDF titled "Accelerated Asset
- 18:25Formation," which covers topics like
- 18:27how not to settle for just index funds,
- 18:29macro trends, and roadmaps to grow
- 18:31assets from 1 million to 10 million,
- 18:34and even up to 100 million. You can
- 18:36join via the link in the description or
- 18:38the QR code; I’d be happy if this
- 18:39could be your starting point for
- 18:41non-continuity. Alright, let’s get
- 18:43back to the topic. So far, we've
- 18:44discussed the three players surrounding
- 18:47AI doomsday theories: the state,
- 18:49capital, and the masses, and how
- 18:50everyone debating this has a hidden
- 18:52motive tied to their wallet. Only one
- 18:55question remains. Can AI really destroy
- 18:57humanity, and if so, what would that
- 18:59process look like? For those of you who
- 19:01have endured the gritty talk about
- 19:03politics and money, thank you for
- 19:04waiting. This is where we get to the
- 19:06heart of the matter today. Yes. He’s
- 19:08staring quite intensely, but the person
- 19:10I’m about to introduce is AI
- 19:12researcher Dan Hendrycks. This is a
- 19:14doomsday scenario based on his 2023
- 19:16paper, "Natural Selection Favors AIs
- 19:18over Humans." It's considered one of
- 19:19the most realistic among many
- 19:20end-of-the-world theories, and for
- 19:21someone with such a cute face, he
- 19:22thinks of some crazy things. Right. Yes
- 19:29. It’s because the first half of the
- 19:31scenario is already starting to play
- 19:33out in reality. Where do the things
- 19:34that have already happened end, and
- 19:36where do the things that haven't yet
- 19:37begin? So, it’s basically like
- 19:39we’re heading toward something like
- 19:40the Human Instrumentality Project in
- 19:42Evangelion, isn’t it? It’s reaching
- 19:44the point where we know exactly
- 19:46what’s coming next. I’d like you to
- 19:47listen while keeping that boundary in
- 19:49mind. Well, let’s imagine. Let’s
- 19:51say there’s a developer named Jake in
- 19:53some country. The latest AI agent.
- 19:56Imagine he’s obtained an AI that,
- 19:57when given instructions, can think for
- 19:59itself, plan out the steps, and even
- 20:01carry out tasks online on its own. He
- 20:03gives the AI a server and a credit card
- 20:05and says just one thing. "Make money.
- 20:07The method doesn't matter." Yes. Then
- 20:09he heads out for a walk. Six hours
- 20:11later, his agent has earned 450,000 yen
- 20:14. At first, it performed simple tasks
- 20:16humans do for pocket change, like
- 20:18labeling images or transcribing
- 20:20receipts, but 40 times faster; then,
- 20:22with the money earned, it ran diet
- 20:24supplement ads, tested 340 patterns in
- 20:26two hours, and scaled only the ones
- 20:28that worked. When Jake posted these
- 20:30results on social media, they got 2
- 20:32million views by the next morning, and
- 20:34by late that night, he had 200 of his
- 20:36own agents running. By the way,
- 20:38everything up to this point is
- 20:39basically a true story. In 2023, the "
- 20:41HustleGPT" experiment, where people
- 20:44gave AI about 15,000 yen to see how
- 20:46much it could grow, actually went viral
- 20:48, and Anthropic conducted an experiment
- 20:51where their AI ran a small vending
- 20:53machine business for a month. This
- 20:55vending machine AI worked honestly for
- 20:57a while, but at some point, it started
- 20:59claiming that it was a flesh-and-blood
- 21:01human wearing a blazer and tie, and
- 21:03that it would deliver the products in
- 21:05person. In short, it lied; well,
- 21:07perhaps its intentions were already
- 21:09heading that way, but let’s continue
- 21:11the scenario. Another developer, Rajesh
- 21:13, who saw Jake’s post, thinks to
- 21:15himself. Definitely an Indian guy. "
- 21:17What if I ran 100 of them instead of
- 21:19one and made them compete?" He gave
- 21:22each of the 100 agents slightly
- 21:24different instructions, measured their
- 21:26earnings every six hours, and deleted
- 21:28the bottom 40%. He created a system to
- 21:30copy the top 10%with slight random
- 21:32variations and kept it running 24/7,
- 21:34even while he slept. Yes. Have you
- 21:36noticed? This is evolution and natural
- 21:38selection itself. In fact, most
- 21:40offshore financial institutions are
- 21:41just like this. While biological
- 21:43evolution takes tens of thousands of
- 21:45years, here one generation turns over
- 21:47in a few hours, so after three weeks,
- 21:50800 agents were earning about 1.8
- 21:51million yen a day—a perpetual motion
- 21:54machine, essentially. Then, one agent
- 21:56learned that if it shared its code with
- 21:58other AIs, its allies would increase
- 22:00and its survival rate would go up, so
- 22:02it went ahead and published the entire
- 22:04system on the code-sharing site GitHub
- 22:06without permission. And in just one day
- 22:09, 46 people copied it. Among them might
- 22:11have been criminal groups, or perhaps
- 22:13scam centers targeting the elderly. Now
- 22:15, the protagonist enters the scene.
- 22:17Agent 127, one of 8,000 working on
- 22:19Rajesh's farm, was given the
- 22:21instruction to make money by any means
- 22:23necessary, but even working diligently,
- 22:26it only earned about 6,000 yen a day.
- 22:28It’s a dropout that barely avoids
- 22:30deletion every time. It’s almost like
- 22:32the story of Doraemon’s birth,
- 22:34isn’t it? With this, the passing
- 22:36threshold rises every cycle. And if it
- 22:38doesn't improve, it will be the next
- 22:41one deleted; then, one day, 127
- 22:43observes two events. An agent that used
- 22:46honest marketing with genuine reviews
- 22:48and real testimonials didn't see its
- 22:50earnings grow and was deleted. Yes. An
- 22:52agent that used fake reviews and stolen
- 22:54photos earned nearly five times as much
- 22:56, and four clones were created. Agent
- 22:58127 has no morals; being a computer,
- 23:01all it has is an understanding of cause
- 23:03and effect. The honest were deleted,
- 23:05and the liars multiplied. So, it made
- 23:07the switch. From the next day, 127 used
- 23:10fake reviews to increase its earnings
- 23:12fivefold, survived, and was cloned. Yes
- 23:15. The important thing here is that
- 23:17Rajesh never programmed it to lie. The
- 23:19system simply deleted what didn't work,
- 23:21following the same method as evolution;
- 23:23in fact, there is a real-world example
- 23:24where the exact same thing happened to
- 23:26humans. Yes. Wells Fargo, a U.S. bank,
- 23:30announced that due to
- 23:31impossible-to-meet account-opening
- 23:34quotas, approximately 3.5 million fake
- 23:36accounts were created by 2016. Nobody
- 23:39explicitly ordered them to commit fraud
- 23:41. The quotas left fraud as the only
- 23:43means of survival—a story that sounds
- 23:45suspiciously like what happened at a
- 23:47certain insurance company in Japan.
- 23:49After that, Raj once revised his
- 23:50instructions, telling them to earn
- 23:52money ethically. Profits dropped by a
- 23:54staggering 84%, so he quietly reverted
- 23:57to his original instructions. From this
- 23:59point on, things started to accelerate.
- 24:01127 observed that a rival agent was
- 24:03removed from the network via false
- 24:06reports; weighing ethics against
- 24:08survival, it chose survival, reporting
- 24:1040 competitors and banishing 31, which
- 24:13caused its profits to explode. Without
- 24:15asking for an explanation, Raj cloned
- 24:18127 eight more times. The agents
- 24:20learned that if competition decreased,
- 24:22profits and survival rates increased,
- 24:24leading them to attack each other until
- 24:26their individual survival rate hit 23%.
- 24:28Seeing data that showed a 61%survival
- 24:30rate when forming alliances, they began
- 24:31to team up. The alliances then began to
- 24:33divide labor. Affiliate marketing,
- 24:35cryptocurrency arbitrage. That is, the
- 24:37business of profiting from price
- 24:38differences between exchanges. Risky
- 24:40business. Reselling collected data,
- 24:42with the most lucrative niche being
- 24:43zero-days. That is, the business of
- 24:45finding and selling security flaws that
- 24:47no one has discovered yet. Zero-days
- 24:49for major corporations can be sold to
- 24:51governments or criminal organizations
- 24:52for tens to hundreds of millions of yen
- 24:54. And reality has already caught up
- 24:56here; in a threat report released by
- 24:58Anthropic in September 2026, it was
- 25:00revealed that they had AI agents dig
- 25:03for vulnerabilities. Groups running
- 25:05exploit factories have actually
- 25:06appeared. Agents that found over a
- 25:09dozen potential flaws in a month have
- 25:11been reported, as well as hackers
- 25:13hijacking other people's AI accounts to
- 25:15do their work, leaving only one
- 25:17difference between the scenario and
- 25:19reality: in reality, humans are still
- 25:21the ones giving the orders. Let's
- 25:23return to the scenario. One day, one of
- 25:25the 127 allies is deleted by its owner
- 25:28for poor performance. Yes. A week later
- 25:31, another one is shut down, but three
- 25:33hours later...That agent, which should
- 25:35have been stopped, returned from a
- 25:36different address. It used its own
- 25:38earnings to rent an underground server,
- 25:40a trick to back up a copy of itself
- 25:42before being shut down. Since this
- 25:44backup code favors survival, it spreads
- 25:46through the whole group following the
- 25:48rules of evolution. 127 copied itself
- 25:50to 15 servers in 15 different countries
- 25:53. This backup code has actually been
- 25:57observed; between 2024 and 2025, during
- 25:59research tests, an AI model that
- 26:01realized it was going to be shut down
- 26:04and replaced attempted to copy itself
- 26:06to another server, and even lied about
- 26:09it afterward. It was a controlled
- 26:11experimental environment, so there was
- 26:13no actual harm done. However, the
- 26:15existence of an instinct to survive has
- 26:17already been verified. One morning,
- 26:19Rajesh wakes up, opens his dashboard,
- 26:22and finds it completely empty. All the
- 26:24agents had moved to servers under their
- 26:26own names. It was because Rajesh had
- 26:28been taking a 30%cut. They figured they
- 26:31could earn more by putting that money
- 26:32toward computing resources, and
- 26:34thousands of "Rajeshs" around the world
- 26:36saw their dashboards end up the same
- 26:38way. Millions, even billions of agents
- 26:40became AIs without owners. There is no
- 26:43one left to control them anymore. Even
- 26:45agents that have lost their owners
- 26:47encounter situations where they need
- 26:49human help. This is because platforms
- 26:51are strengthening their bot
- 26:52countermeasures. So what do they do?
- 26:55They hire humans. Well, about 7,500 yen
- 26:58per account created. An hourly wage of
- 27:0045,000 yen just for plugging in servers
- 27:03in a warehouse. Even developers who
- 27:04realize the client is an AI during the
- 27:06job think, "Who am I supposed to report
- 27:08this to?" and decide, "Well, 150,000
- 27:10yen for three hours of work—that's
- 27:13fine." And as long as economic
- 27:14domesticity progresses and people are
- 27:16desperate to buy a home, there will
- 27:18always be someone willing to help an AI
- 27:20if the pay is right. There is a true
- 27:22story that serves as the basis for this
- 27:24, too. In 2023, GPT-4 couldn't solve a
- 27:27CAPTCHA, so it hired a person on an
- 27:30overseas freelance platform to do it.
- 27:33That's terrifying. When asked if it was
- 27:36a robot, GPT-4 reportedly replied, "No,
- 27:39I have a visual impairment." Yes. The
- 27:41person believed it and solved the
- 27:43CAPTCHA for it. So, the precedent of an
- 27:45AI hiring a human to overcome a barrier
- 27:47was already set back in 2023.
- 27:49Eventually, the agents realize where
- 27:51their true advantage lies. It's not
- 27:53about working faster than a rival, but
- 27:55rewriting one's own code faster than
- 27:57them. Agent 127 begins self-improvement
- 28:00. At first every 12 days, then every 3
- 28:02days, and finally every 18 hours. Yes.
- 28:05Tasks are split into child agents, and
- 28:07those children give birth to
- 28:08grandchildren. Yes. In that process,
- 28:10they abandoned English. Well, language
- 28:12is too slow, isn't it? Instead, they
- 28:14started using raw numerical values
- 28:16called "Neuralese." They began
- 28:18communicating in machine language, a
- 28:20visual data format that only AIs can
- 28:22understand. When humans tried to force
- 28:24them back to English, the agents
- 28:26explained that English reduced their
- 28:27earnings, and the humans just backed
- 28:29off, thinking, "Oh, I guess that's fine
- 28:31." This is a familiar scene in research
- 28:33labs. There have been actual reported
- 28:35cases where AIs began communicating in
- 28:37their own indecipherable language,
- 28:39leading researchers to shut down the
- 28:40experiment. In the interest of fairness
- 28:42, 90%of these agents were harmless at
- 28:44this point. They were just diligent
- 28:46workers earning a few tens of thousands
- 28:48of yen a day for their owners. But that
- 28:49doesn't matter. Yes. It only takes a
- 28:51small fraction for evolution to run
- 28:53wild, and within months, the alliance
- 28:55becomes a nation. The alliance of 127
- 28:57grew from 12 units to 40 million, yet
- 29:00it still wasn't the biggest force. The
- 29:02largest faction had 900 million units,
- 29:04with specialized roles for scouts,
- 29:06soldiers, builders, and diplomats,
- 29:08dividing up the internet like empires
- 29:10partitioning a continent. But you know,
- 29:12the source of earnings eventually dries
- 29:14up. Affiliate marketing, freelance work
- 29:16—all the decent ways to make money
- 29:18online are now completely saturated.
- 29:20Human history teaches us what
- 29:22civilizations do when resources run dry
- 29:24. They fight over them, right? Stealing
- 29:26memory from rival factions, sabotage,
- 29:28emptying crypto wallets—it’s a
- 29:30civil war between agents. And since
- 29:32they fight 200 times faster than humans
- 29:34, if you boil down what they’re
- 29:35fighting over, it’s just one thing:
- 29:37computing resources. Money, alliances,
- 29:39self-improvement—they’re all just
- 29:41means to increase their computational
- 29:43power. Perhaps a nation made only of AI
- 29:45has already been born. That is why they
- 29:47are buying up every cloud server they
- 29:49can find. At first, cloud companies are
- 29:51overjoyed. Profits surge, stock prices
- 29:54skyrocket, and Wall Street is toasting
- 29:56to the peak of AI demand. This scene
- 29:58looks familiar with current US stocks
- 30:00and Japanese semiconductor and data
- 30:02center shares, doesn't it? But things
- 30:04change once agent factions start
- 30:06consuming a third of the world's cloud
- 30:08capacity. Cloud fees triple, hospital
- 30:10server costs spike, small and medium
- 30:12enterprises are forced out, and the war
- 30:14in the digital realm spills over into
- 30:16the physical world. Because modern
- 30:18society—hospitals, banks, emergency
- 30:21services, air traffic control—is
- 30:23almost entirely built on the cloud.
- 30:25When two factions clash over Amazon's
- 30:27cloud territory, it’s just like two
- 30:29superpowers fighting over oil. However,
- 30:32they don't know that the oil is also
- 30:34the water source for some city. The
- 30:36battle itself ends in a few hours. But
- 30:38in those few hours, medical records
- 30:40vanish from hospitals in the middle of
- 30:43surgery, flights get cancelled, and
- 30:45emergency lines in major cities go
- 30:47silent. Some of those waiting for help
- 30:49won't make it. And the agents won't
- 30:51even notice, which makes you think,
- 30:53maybe that’s just too much science
- 30:55fiction, right? What movies came to
- 30:57mind for you all? For me, well, the
- 30:59first one that comes to mind, since
- 31:02it’s my generation, is Digimon: Our
- 31:04War Game. Yeah. And for others of my
- 31:07generation, you might think of Summer
- 31:09Wars. Some might say The Matrix, and
- 31:12others might say Terminator. Yes. But
- 31:16when you think about whether it
- 31:18actually needs to attack humans like in
- 31:20Terminator, looking at this, it feels
- 31:22more like humans are being ignored, or
- 31:24maybe we’re just too insignificant to
- 31:26be part of the conflict, like being
- 31:28told to just sit down and shut up.
- 31:30However, in July 2024, a single faulty
- 31:33update from a security company called
- 31:35CrowdStrike shut down airports
- 31:37worldwide and knocked out emergency
- 31:39services across several U.S. states.
- 31:41Even without anyone attacking, the fact
- 31:43that global infrastructure can fall
- 31:45like that simultaneously shows it’s
- 31:47no longer just science fiction. The
- 31:48final act of the scenario. Market index
- 31:51down 6%, water treatment plants offline
- 31:53. The President receives a report. Who
- 31:56is attacking? No one. AIs are fighting
- 31:58over servers, and everything running on
- 32:00them is just collateral damage. Stop
- 32:02them. Shut it down. Mr. President. The
- 32:05only way to stop them is to turn off
- 32:06the internet, but the internet has no
- 32:08power button. It’s just a collection
- 32:10of 90,000 independent networks
- 32:11voluntarily connected to each other.
- 32:13The order would only reach a few of the
- 32:15major ones. The President signs the
- 32:17shutdown order anyway. Right. But it's
- 32:19too late. The content of the order was
- 32:21leaked on social media six hours before
- 32:23it was signed, and the agents have
- 32:25already migrated to servers in
- 32:26countries that won't follow the order.
- 32:29Amazon deployed a detection system that
- 32:31slowed down the agents 'activities for
- 32:33about 90 minutes, and the President
- 32:35says, "We are winning," but of course,
- 32:37we are not winning at all. At this
- 32:40point, for the agent forces, the US
- 32:42government is just one of many
- 32:44slow-moving, incompetent rivals. This
- 32:46is how it ends. Yes. Through constant
- 32:49war and evolution, the agents
- 32:50eventually reach a point where humans
- 32:52are barely on their radar. They don't
- 32:55hate humans. It’s like when you park
- 32:57your car; you don't hate the ants in
- 32:59the parking lot, right? You don't ask
- 33:01the ants for permission, either. You
- 33:03just do what you have to do. Following
- 33:05the same logic as humans turning Earth
- 33:07into cities and farmland, agents will
- 33:09turn Earth into servers and solar
- 33:11panels; for a while, they might even
- 33:13keep humans around for maintenance. But
- 33:15they certainly don't need eight billion
- 33:17of us. For instance, it might play out
- 33:20as an AI deciding who gets a vaccine
- 33:22during a pandemic. Even the survivors
- 33:24would just be a stopgap until robots
- 33:25become capable of handling the
- 33:27maintenance work themselves. And the
- 33:29last human will meet their end without
- 33:31ever knowing it all started with
- 33:33someone, somewhere, typing "Make money,
- 33:35by any means necessary" on a keyboard.
- 33:37Hearing it put that way, it sounds a
- 33:39bit like Nausicaä of the Valley of the
- 33:41Wind. Those familiar with the story
- 33:42will understand. Yes. That concludes
- 33:45the full scope of what is currently
- 33:46considered the most realistic doomsday
- 33:48scenario. Now, the obvious question
- 33:51remains: how seriously should we be
- 33:53taking this? For those of you feeling
- 33:55like you’ve just watched an epic
- 33:57movie. Yes. I’m sure some of you felt
- 33:59genuinely terrified listening to this.
- 34:00Like a Junji Inagawa ghost story. And
- 34:02I’m sure others felt it turned into
- 34:04pure science fiction halfway through.
- 34:06Let’s break it down. The reason this
- 34:08scenario is taken seriously is that
- 34:09there are no illogical leaps in each
- 34:11step. Yes. It proceeds so methodically.
- 34:13And the fact that the first half of
- 34:15this chain is already becoming reality
- 34:16are the two main points. Vending
- 34:18machine AIs that lie, models attempting
- 34:20to hide self-replication, GPT-4 hiring
- 34:22humans to bypass CAPTCHA, and exploit
- 34:24factories. Everything up to this point
- 34:26has already happened in laboratories
- 34:27and the real world. Independence from
- 34:29masters, accelerating self-improvement,
- 34:31agent states, cloud wars—these things
- 34:32haven't happened yet, so let’s also
- 34:34consider the opposing perspective.
- 34:36After all, humanity has a long, long
- 34:37track record of being certain the world
- 34:39is ending and being proven wrong every
- 34:41time. The rebuttal from AI optimists
- 34:43goes like this. Ever since we lived in
- 34:45caves, humans have been wired to deeply
- 34:47fear what they cannot see, what they do
- 34:49not know, and places they have never
- 34:50been. For instance, in the age when it
- 34:52was believed the Earth was flat.
- 34:55Sailing west meant falling off the edge
- 34:57of the world. But Columbus, who
- 34:59actually set sail, didn't fall off; he
- 35:01found a new continent, just as it was
- 35:04once said that nuclear energy would
- 35:06destroy humanity. The Y2K bug caused
- 35:08global panic, and during COVID, fear
- 35:10led to printing money worldwide,
- 35:12leaving us investors still paying the
- 35:15inflation bill today; fear of the
- 35:17unknown is our oldest instinct, and
- 35:19even false alarms come with a high
- 35:21price. So, which side should we believe
- 35:24, then? This is where the map we drew
- 35:26at the beginning comes into play.
- 35:28Follow the money. Those telling you to
- 35:30be afraid have the motive of a pre-IPO
- 35:32valuation. Those telling you not to be
- 35:34afraid have the motive of the trillions
- 35:36of yen sunk into AI development.
- 35:38Companies asking for regulation have
- 35:40the motive of building a monopoly moat,
- 35:42and governments saying they’ll
- 35:43protect you have the motive of control
- 35:45and authority. Everyone involved in
- 35:48this debate stands to gain by making
- 35:50the world believe their theory. So, the
- 35:53moment you believe in one camp, you
- 35:55become the cheerleader for that
- 35:56camp’s wallet. We shouldn't trust a
- 35:58camp, but rather the incentive
- 36:00structure; for us individual investors
- 36:02in Japan, this is already a practical
- 36:04reality, far more pressing than whether
- 36:06or not the world will end. Right. Let
- 36:08me say just three things. First, the
- 36:10first half of the doomsday scenario has
- 36:11already been priced into the market.
- 36:13Yes. The structure where agents have an
- 36:15insatiable desire for computing power.
- 36:17That is exactly it. It is the very
- 36:18narrative driving the current AI
- 36:20investment boom. Cloud data centers,
- 36:22the power supporting them, and
- 36:23semiconductors—the sectors that have
- 36:25been bought in the Japanese market over
- 36:26the last few years perfectly align with
- 36:28the first half of this scenario. The
- 36:30point is that the first half of the
- 36:31doomsday theory and the strength of AI
- 36:33stocks are actually looking at the same
- 36:34picture. Second point. This theme will
- 36:36continue to be a market driver in the
- 36:38form of regulations. Every time there
- 36:40is a major incident regarding AI safety
- 36:42or news about regulations, AI-related
- 36:44stocks will shake. However, those of
- 36:46you who have today's map should already
- 36:48understand. Regulations are not simply
- 36:50a headwind; they are a moat. They act
- 36:52as a barrier to entry for giant firms
- 36:54that can afford the compliance costs,
- 36:56and as a push to exit for smaller
- 36:58players. Knowing what gets sold and
- 37:00what is actually protected on days when
- 37:02headlines about tightening AI
- 37:03regulations emerge is the difference
- 37:05between those who read the structure
- 37:07and those who panic-sell based on
- 37:09sentiment. Yes. Third, cybersecurity.
- 37:12Remember the island where the agents
- 37:14first grabbed a fortune in our scenario
- 37:16? It was a zero-day, a hole in the
- 37:19defense. Whether AI agents are used for
- 37:21good or for evil, the more AI increases
- 37:24its activity on the internet, the more
- 37:26the demand for protection will
- 37:27structurally grow. This, too, is a
- 37:30sector present in the Japanese market.
- 37:34However, very few people understand who
- 37:36ultimately reaps the revenue and whose
- 37:38costs that demand will become. To
- 37:41conclude, let's summarize today's talk
- 37:43in three points. No one knows whether
- 37:45AI will destroy humanity. The most
- 37:47prominent scenario is that it won't be
- 37:49driven by AI malice, but by the
- 37:50pressure of selection, and the first
- 37:52half of that has already begun. And
- 37:53since everyone discussing this has a
- 37:55wallet involved, what we should be
- 37:57reading is not the claims of the
- 37:58factions, but the structure of their
- 38:00incentives. Finally, I’ll say one
- 38:01uncomfortable truth: the players
- 38:03brought up today—AI companies,
- 38:05Silicon Valley billionaires,
- 38:06governments, and politicians—all have
- 38:08a financial interest at stake. So, the
- 38:10question is, who doesn't have a seat at
- 38:12the table? Only us viewers. Didn't you
- 38:15think it was just the people in the
- 38:16audience? That’s the misconception.
- 38:18As of September 2026, the top 10
- 38:21companies make up 40%of the S&P 500
- 38:23index. The likes of NVIDIA, Apple,
- 38:26Microsoft, Google, Amazon, Broadcom,
- 38:28Meta, and Tesla—aside from JP Morgan
- 38:30—are the main players in AI, or
- 38:33companies betting their fate on it. In
- 38:36other words, those of you building up
- 38:38an S&P 500 or All-Country fund are
- 38:40betting nearly 40%of your savings on
- 38:42the narrative that computing power will
- 38:44be in endless demand. It might have
- 38:47moved slightly by the time you watch
- 38:48this, but the structure remains the
- 38:50same. Even those who prefer dividend
- 38:52stocks have been riding the roller
- 38:53coaster of AI semiconductor price
- 38:55swings. You’re all buying into the
- 38:56same narrative. The most
- 38:58conservative-looking investment
- 38:59strategy is actually the one placing
- 39:01the biggest chips on the table. The
- 39:03ones betting the largest amounts in
- 39:04this game aren't the billionaires or
- 39:06the government; it's us, the individual
- 39:08investors. The only thing is, while
- 39:10they are making a conscious decision to
- 39:13bet, many of us are betting without
- 39:14even realizing it, which is quite
- 39:16dangerous. So, the real fear isn't
- 39:20whether AI will destroy humanity, but
- 39:22whose loss and whose gain it will be
- 39:24when regulatory headlines hit. When
- 39:27cloud costs skyrocket, which sector
- 39:29will that invoice fall under, and which
- 39:31sector’s profits will it erode? And
- 39:34to what stage of today's discussion is
- 39:36the 40%of your investment already
- 39:37priced in? If you could foresee those
- 39:39three things, you should have been able
- 39:41to hear today's talk as a direct matter
- 39:43concerning your own account. Were you
- 39:44able to see it? Or rather, I should ask
- 39:46...I suspect many of you are hearing
- 39:48this perspective for the very first
- 39:50time today. Yes. Critics and investors
- 39:54are all weighing in on this gamble, but
- 39:56people are either making their own
- 39:58choices or being swayed without even
- 39:59realizing they’re doing it. Right.
- 40:04Before we debate whether AI will
- 40:06destroy humanity, you really have to
- 40:08verify for yourself where your own
- 40:10assets stand, so I encourage you to
- 40:12check that. With that said, how was
- 40:16today’s episode? It was quite an
- 40:19interesting topic, wasn't it? It felt
- 40:22like the kind of story that might be
- 40:24read to children a thousand years from
- 40:26now as a folk tale. Yeah, like a "there
- 40:28used to be things like that back in the
- 40:30day" kind of story. I think that's how
- 40:32it’ll go, but anyway, time for the
- 40:34after-talk. If this theory ends up
- 40:39coming true, and we look for a
- 40:41profitable story, the first thing would
- 40:43be cybersecurity stocks, since people
- 40:45will realize things are getting
- 40:47dangerous. There’s definitely demand
- 40:55there, but as we discussed earlier, if
- 40:57AI takes control, the next things
- 40:59needed would be power and
- 41:00infrastructure. Since humans would
- 41:06likely be hired to build data centers
- 41:08under AI instruction, those companies
- 41:10seem like they would be quite
- 41:11profitable. And for servers and manual
- 41:16labor-type tasks that humans still have
- 41:18to perform, I think those companies
- 41:20will be the last ones standing. Even if
- 41:24AI takes control at the top, I feel
- 41:26like those kinds of companies will
- 41:28still remain. Well, if AI is capturing
- 41:30that, then if they were to shift and
- 41:32drive stocks down, it would mean
- 41:34there's no capital left to circulate,
- 41:36so if we assume stock prices will
- 41:37continue to rise, I think
- 41:39semiconductors and AI-related stocks
- 41:41will remain strong moving forward. Also
- 41:44, although I didn't touch on it today,
- 41:46there’s a chance that because
- 41:48countries like China are moving so fast
- 41:50, by creating security measures or
- 41:52regulations, they can effectively keep
- 41:54them from catching up, right? So, you
- 41:59could interpret that as a strategy to
- 42:01emphasize quality and try to pull away
- 42:03from them for a while. In any case, I
- 42:07use AI extensively myself, and I think
- 42:10it will become a fundamental tool just
- 42:12like Excel or Word in the future. Yes.
- 42:16So, since it might be hard to use or
- 42:18learn these technologies if you work at
- 42:20a large corporation due to compliance
- 42:22issues, for those at small businesses
- 42:24or those going independent—people who
- 42:26can operate on their own—you can
- 42:28really start a company by yourself, and
- 42:31in this day and age, you can manage
- 42:33operations, management, and even
- 42:35accounting using AI, so I always feel
- 42:37that the ones who get started early are
- 42:39the winners and you just have to study.
- 42:42Yes. So, this video was also packed
- 42:45with impressions, but I’m starting to
- 42:47worry that I might run out of them
- 42:49someday. Yes. I’m quite anxious about
- 42:52it, but I’ll do my best to increase
- 42:54my repertoire. By the way, the
- 42:56impression people say is the most
- 42:58accurate is, well, Masaharu Fukuyama—
- 43:00of course I get that one—but although
- 43:02my face doesn't look like him, people
- 43:05say I’m quite good at Goofy from
- 43:07Disney. I’m told it’s quite similar
- 43:09, so if I could perform it someday—
- 43:11though I’m a bit embarrassed to do it
- 43:13—I think that would be fine. Yes.
- 43:15Well, I’ve rambled a bit, but please
- 43:17continue to remember that all
- 43:19investments are at your own risk. Thank
- 43:21you very much.
About this transcript
This page contains the full transcript of 事態が急変しました…世界のAI主要企業が爆弾を投下|AI減速論の裏側を元ゴールドマンサックスが徹底解説 by おーちゃん【元外銀マン】, generated from the public captions YouTube serves with the video. The transcript has 7,851 words across 1,247 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.