Title: Amazon Blocks Meta: The Agentic Wars Have Started
Show: Joseph Carlson After Hours
Guest: Joseph Carlson
Date: 2026-SEP-21
URL: https://youtu.be/rQJMySrmw5E
Length: 29:59 (1799 s)
Note: Live YouTube title "Amazon Blocks Meta: The Agentic Wars Have Started"; the transcript Stephen pasted was headed "Everything Just Changed For Meta" (likely an A/B test title for the same video). Auto-captions cleaned — verbal fillers (uh/um) and stutters/false starts ("that that's", "when you when you", "It It's", "So, so", "they they", "two two ways") removed; wording, numbers, names and every (mm:ss) cue otherwise verbatim. Caption mis-hearings are left as captioned: "qual.com/Qualrum/Quelt" = Qualtrim (his own site — sponsor plug); "ChateBT/chatbt/ChachiBT/CHACBT/ChatBT/Chachebt/catches" = ChatGPT (OpenAI); "Metamuse/MetaMuse" = Meta's Muse assistant; "Daario" = Dario Amodei (Anthropic); "Sam Alman" = Sam Altman (OpenAI); "Jensen Wong" = Jensen Huang (Nvidia); "Gene Monster/Gan Monster/Gan" = Gene Munster (Deepwater); "Andrew Ying" = Andrew Yang; "Aentic" = agentic; "moes" = moats; "Open" (06:02) = OpenTable; "Astra" kept as said. "It was called Rufus" (16:00) refers to Amazon's Rufus shopping assistant. Lines marked ">>" are the CNBC clips he plays (Gene Munster, Andrew Yang and the anchors), not Carlson.

(00:00) Welcome back everyone. Today on the Joseph Carlson show, the agentic war has begun. Companies like Amazon are blocking Meta's Muse and Meta. Well, the stock is taking off. It is up 8% on the day, which puts me right now $11,000 in the green on Meta stock. This is a company that we went from minus $30,000 in the red to now $11,000 in the green in only 30 days.
(00:25)  The sentiment shift is happening for Meta. A new story is unfolding. We have the news bashing Meta's new product, Metamuse, which is predictable. We have the banks upgrading Meta now, turning a new leaf on it, saying now the stock is way undervalued. And of course, we're going to be going into Amazon blocking Meta. Amazon says it's blocked Muse over security and privacy concerns.
(00:44)  Of course, that is not the real reason. We'll be going over the real reasons Amazon is blocking Meta and what Meta is going to do about it. And of course, we have the fail of the week, which is Andrew Yang, who went on to another viral interview on CNBC, said more doomerish things that are just not true. They're not even real.
(01:00)  He's just making stuff up on TV. We'll be going over it. And we have a fail of the week update with Gene Monster. We highlighted him as a fail of the week a month ago when he came out bearish on Meta. So, we have a ton to get to in this episode. And before we jump in, just a quick reminder, if you haven't tried out qual.com, I think you're missing out.
(01:17)  I built the website for you. It runs at a low price, $10 a month with a free trial. You can try it out. It's super fun. You get access to an entire investment suite. You have the entire website, earnings transcripts, discounted cash flow. You also get access to exclusive content with Qualrum Studio.
(01:35)  And you get access to a private community with over 10,000 active investors. In fact, the membership has now climbed around 13,000. We have a massive community now because so many people have tried out Quelt and they like it. So, I think you're missing out if you haven't. The other thing I'll mention is I've been working on a mobile app and it's awesome.
(01:52)  It is literally my favorite stock analysis app and we're going to be releasing that probably next month. So, lots of cool things coming. Try out Qualrum if you haven't. Now, we kick things off today. Looking at the massive jump in Meta stock price. Right now, Meta Stock is up 8.59%. Now, there's no guarantee it will hold, but it looks like it's going to.
(02:10)  It looks like Meta investors are very excited about whatever's going on right now. What could be going on with Meta to cause the stock to jump this much from the pre-market into the trading day? And it continues to go up with momentum. In fact, Meta stock has been one of the best stocks over the past month.
(02:26)  If we look at the past one month, it's up 33% now. Meta has raced up from $534 to now trading at 722. So, it's gaining hundreds of dollars in share price. It's been one of the best performers in the market over the past 30 days. And there's been a couple things to cause this. First of all, when we look at Meta just on the surface, a month ago, it was trading on an incredibly low PE ratio on a forward basis. It was just cheap.
(02:53)  It was also growing quickly. And that's just a good situation. Cheap company growing quickly. The other thing that you have to factor in is the quality of the company, the moat, how defensible it is. When I looked at Meta, I considered it to be one of the widest and deepest moes in the market. AI is not going to disrupt it.
(03:11)  There's not many things that can disrupt a meta. even lawsuits don't disrupt it. Meta was able to settle the lawsuit with all the attorney generals over the past month as well. That was one more distracting factor that they could get behind them. So, the slate became a little more clean with Meta. It was simply a lowvaluation deep moat company growing very quickly.
(03:29)  It was the ideal setup for an investment. And when this happens, all it takes is a little sentiment change, a little story change to have a big change in the stock price. And that's exactly what happened with Muse. We know that Meta was working on some type of consumer product. Mark Zuckerberg talked about AI in a very glowing and positive way.
(03:49)  He said that it should enable humans to accomplish more. It should take away all the annoying things out of your life so you can focus on the things that really matter to you. But that's very vague. What does that even mean? Well, it was vague until we got Muse. Metamuse is the personification of that idea that humans can accomplish more.
(04:08)  the agents will do all the annoying things so humans can focus on really what matters. Muse was released by Meta just a week ago and it's been a major hit. It's been at the number one of the Apple App Store almost every day. In fact, it climbed to the second place on day two and now it's in number one and it's been there ever since dethroning ChateBT even if for a while this is an impressive feat.
(04:33)  But when we look at Metamuse, how did it really accomplish this? Well, they did things different than chatbt. For example, chatbt feels a little bit more cold, feels a little bit more clinical. When you use chatbt, you have an empty search bar and a screen. With Metamuse, you have a nice little whimsical friend there. He greets you.
(04:51)  He doesn't just wait for you to talk. He says, "Hey, here's some of the stuff I can do. What can I help you with?" And then he's animated and it looks like he's searching things on his little computer. It's the small things like that that make it a little bit more fun for some people. It immediately makes the app feel more relatable, a little bit more cute.
(05:09)  It ironically makes it feel less robotic. There's also little things that Muse has done to make their app better for most people. For example, you don't have a model switcher. There's no toggle to go for Muse quick or Muse instant or Muse thinking or Muse extra extra thinking. With Muse, you don't hassle with what model to pick.
(05:27)  You just type in whatever you want and it does the picking for you. It's automatic transmission. It selects whatever models best for your query. Now, all of those things are very helpful. The ease of use, the user interface, and the cute little assistant character, all of those make for a more friendly and more relatable app. But those aren't enough.
(05:46) The Agentic Assistant actually has to be good. And this is where Muse really flourishes. I don't know how many servers or how many GPUs Mark Zuckerberg's putting behind Muse, but it works well. The answers are incredibly good, and they're very fast. You don't wait a long time for it to answer complex questions.
(06:02)  It gets the answer out quick and they're in-depth. They're specific. It doesn't also seem to hallucinate a lot. With my testing with it, I've been very impressed with its usage and capabilities. So, what Metamuse has done is created an app that gains your trust with small little asks. Hey, connect to Open. And then a few minutes later, you're connecting it to your Google Drive and your Gmail and your calendar and all these different assets that they can use to learn more about you and to make the assistant even more valuable. All of a sudden, the
(06:31) company that nobody trusts, which is Meta, well, everyone's handing over their credentials to log in to different things through Muse. Muse is earning trust to the consumer very fast. The reviews of it are also glowing. Even though they're limited with only 25,000, it's not millions of reviews yet, they are overwhelmingly positive. 4.
(06:51) 9 star rating in the app store and people are saying it's the best personal assistant they've used. You also don't hit rate limits quickly or at all. The usage that they give you for free is massive. Mark Zuckerberg has given you an insane amount of credits for free to even make the app have less friction to use and less frustration.
(07:09)  Overall, the experience is so good with Muse that I have not heard of one person using it that really disliked it. Not a single one. Everybody that gets their hands on it actually goes, "Wow, this is pretty good." And the more they use it, they go, "Wow, this is actually really good." It is an incredibly well executed app by Meta.
(07:26)  And this creates almost a perfect storm. You have a company that has 3.6 billion daily active users. Meta has distribution that is incredible. Their amount of people that actually use one of their apps once a day is incredible. They have a lot of ways to push people to use Muse, to actually try it out. And with the app actually being good, which it is, that creates the perfect storm.
(07:46) Even despite the efforts of the media, which hates Meta and always hates Meta and always will hate Meta, these companies are trying to scare people out of using Muse, saying they're harvesting your data or surveilling you. It's just not working in this case. More and more people are trying out Muse every single day and starting to use it as their primary AI assistant.
(08:06)  And this is a change in narrative. See again when you have a big company wide moat growing quickly with a low valuation you have a great setup to begin with. That's enough reason in and of itself to buy the stock. And in those situations where you have a really fast growing company at a low valuation and sentiment is negative.
(08:24) All it takes is for a story change, a little bit of a sentiment shift and the stock can pop quickly. And that's what we're seeing happen here with the story of Muse coming into play. investors finally have a reason to have a positive story, something to latch on to, a story to anchor onto.
(08:42)  So even though the fundamentals haven't changed, Meta hasn't increase their earnings per share estimates or anything like that, this changes the story and the sentiment that's attached to this company. And now the analysts are jumping on board. We have Wells Fargo raising their price target from 640 to 796. They say this is following the successful model release and early traction of the MetaMuse assistant.
(09:05)  The firm believes that Meta now has a story to tell. We also have in two days Meta Connect where they're going to be talking about Muse. They're going to be talking about their glasses. They'll be talking about their AI models and progress. They're going to be showing off all the new flashy things to investors.
(09:20)  Now, during Meta Connect, I believe there is a chance, and I think that this is a prediction I'll be willing to make. I think that CHACBT is going to release something to try to distract people from MetaConnect. Maybe their own agentic assistant. ChatBT could release theirs the same day as Meta Connect to try to steal Thunder from it.
(09:38)  But even if that happens, that doesn't change the game for Meta. Meta already understands that they're going to be competing with ChateBT and Anthropic. So that's not going to be any surprise. And I believe that there's a lot of good things coming to Meta. See, Muse is just a beneficial aspect of the story.
(09:52)  It's not the investment thesis in and of itself. It's just a nice add-on. When you have these nice little add-ons, these sentiment shifts on undervalued stocks, you get this big move. We have Meta up 9% now on the day. Well, now Meta is in the green by $13,000 and climbing. Every single minute, it seems to go up more.
(10:08)  So, this all looks very good for Meta. The story is improving by the minute. And even after this 10% increase, Meta still has a lot more room to run. But it's not without some challenges or some roadblocks, and Meta has already ran into the first major roadblock. Amazon has blocked Meta's Muse AI assistant in a new standoff. This is official and I confirmed it with my own Muse account this morning.
(10:30)  See, up until now, when you ask Muse to shop for something, he'll actually open up his own browser and he'll start searching the web. He'll search on Shopify websites. He'll search on Lowe's or Home Depot or Target or Walmart or Sam's Club or wherever. It'll just search like a normal human would, but it does it a lot more passively because you're not wasting your time doing it.
(10:49) And one of the websites he would search on as well is Amazon. But Amazon just threw a wrench into things. They just blocked Muse's access. The way that this looks is when you search on Amazon now, it brings up this little thing in the browser. It says continued access by an unauthorized AI agent violates Amazon's conditions of use to which our customers have agreed.
(11:11)  So, Amazon is detecting the Agentic Assistant and then it's just blocking its access. The problem is Amazon says is that it never agreed to any of it. Meta didn't tell Amazon that Muse would access its store. The agent doesn't identify itself when it browses and it appears to capture the store customers credentials which the company says could create privacy and security risks.
(11:32)  So Amazon says, "Hey, we never agreed to this. You never got our permission. Your agent doesn't say it's an agent or identifies itself as an agent and you're storing customer credentials." Now, there's a couple things that I agree here with Amazon. And then there's a couple things that don't make sense at all.
(11:49)  For example, the customer credential one is just bogus. Amazon is not worried about customer credentials. This is like saying you can't store your customer credentials in your Google keychain or you can't store it in your Apple Passwords app or you can't store it in your whatever password app you have. There's like a million of them.
(12:07) No, Amazon doesn't care about where you store your credentials. Muse is just one more place to store it. Muse isn't getting your credentials through the chat. They have a secure way of storing your passwords just like any other application. So that one is entirely bogus.
(12:25)  Amazon does not care about where you store your customer credentials and Muse and Meta do not see your customer credentials. They have a secure way to store them like they do any credential. The other thing that proves that that entire excuse is completely made up is that even when you don't hand over your customer credentials to Muse, it still can't access Amazon.
(12:45)  My Muse doesn't have my Amazon credentials, but even he could not access Amazon. So they're not blocking you only if you have a login. They're blocking every single Aentic assistant whether they have the login or not. So Amazon's saying this could create a privacy and security risk. Well, if that's true, then why doesn't Amazon allow Muse to browse Amazon when it's not logged into anyone's account? What type of privacy and security risk does it pose then? It has no login credentials.
(13:11)  It's browsing Amazon as an anonymous person, not logged into anything. So, if security and credentials and safety and privacy are not really the concerns that Amazon's worried about, then what are? Well, Amazon actually has good reason to be concerned about Muse that go far beyond security or saving your password. Amazon made $76 billion in advertising over the past 12 months. 76 billion.
(13:35)  And this also happens to be a very high margin business for them. Ads are a very lucrative business. Now, while a small portion of those ads are from Amazon Prime Video advertisements, the huge majority of them come from their gargantuan firstparty business. It's from product placement. Different sellers on Amazon's website will pay money to have their sponsored listing appear before other results.
(13:56)  You have to pay for that premium. So, when you browse Amazon, you're getting a list of results at the top that are people that have paid money to get that real estate. That's valuable real estate. What would really bring down the value of those sponsored listings? If humans weren't really looking at them, if a bot or an agent was crawling over every listing and just finding the best deal, whether it's at the top or not.
(14:18)  See, this is exactly what I outlined in the previous episode. If you haven't watched that one, cue it up after this one because it explains how many of these companies will be impacted by Agentic Commerce. Amazon is clearly exposed with the ad business. If agents start to crawl Amazon, it means that they really have no impact from the advertisements or sponsored listings, which also means that the sellers on Amazon have less of an incentive and less of a return to pay for those sponsored listings.
(14:45)  So, Muse doesn't just present a security threat or privacy concerns or whatever Amazon's talking about here. What it does is it makes their sponsored listings less valuable. The other aspect is simple disintermediation. It's a very valuable thing to be at the end touch point with the customer to directly relate to them.
(15:04)  When you have an agent going through and doing that relationship for you, no longer do you get the same type of information or data knowing what you click on, where you go, how you behave. All of that's useful information and it becomes completely useless if everybody's shopping Amazon with agents. No longer does Amazon have that touch point with the customer.
(15:22)  No longer can they put their sponsored listings or their deals. No longer can they do those little subtle things that humans pay attention to that agents don't. This disintermediation is very unattractive for Amazon. They don't want to be a part of it. So that may be another big reason why they're willing to cut off Muse entirely then let the consumer get used to shopping through Muse on Amazon than directly on Amazon.
(15:43) The other problem for Amazon is that Muse is directly competitive to what Amazon's trying to do themselves. Now they don't have a model like Muse. They don't have an agentic assistant quite like that. And I don't believe that Amazon's going to build one anytime soon, but they do have their Alexa assistant.
(16:00)  It was called Rufus, and it's like the shopping assistant within Amazon. Well, of course, Muse completely bypasses Amazon's own agent. And all of this creates a problem for Amazon. It's stuck in this situation where it sees this agentic assistant coming up and Muse is not going to be the last one. And the question for Amazon is how do they react? Do they try to work with the AI assistants with these agentic commerce companies or do they just block them and try to get people to go to Amazon directly? These are decisions they have to make and clearly they're
(16:27) going with the blocking at least from the beginning. And you may think, well, maybe MetaMuse can just fill out the capture, but that's not what's going on here. There's way more sophisticated ways to identify an agent and block it. So, there's no bypassing this block by MetaMuse.
(16:42)  And there's a couple ways of meta dealing with this. A couple options for them. The first one, and by far the most attractive one, is to get on the phone with Amazon, talk with them, and work out a deal. Because both of these companies benefit from working together more than they do working against each other.
(16:57)  For example, if Amazon just remains blocked from Muse, then everybody that likes Muse and shops on Muse will no longer see Amazon's products. So, it's part of a customer base that's no longer seeing Amazon. That does damage Muse, but it also damages Amazon in the process. It is mutually destructive. Right now, it's likely more destructive to Muse than Amazon.
(17:17)  But as Muse gets bigger, it's likely more destructive to Amazon than Muse. So, it's not good for either. Neither of them benefit at all if this relationship remains blocked. If they're able to make an agreement, some type of revenue sharing model, some type of agreement of how Muse will behave and how it will work. Some type of agentic portal where it has some type of access.
(17:35) If they get creative and made some way where it works for Amazon, they can make money and Muse makes money in the process and they're both happy, that is the best deal for both Amazon shareholders and Meta shareholders and the customer. The other option is that Amazon really puts its foot down. It doesn't want to make a deal with Meta or anyone else and it goes in full force blocking AI assistance.
(17:56)  If that's the case, there are a few ways that Meta can deal with this. For example, if Walmart said, "Hey, Muse, we're open for business. Crawl our website. have everybody buy from us and we have many of the same items as Amazon, then all of a sudden all the millions of users on Muse could be directing their purchases away from Amazon into Walmart.com.
(18:15)  Meta may just say, "If you want to block us, that's unfortunate because we're going to have to direct all of our customers to walmart.com." It's another thing that as Muse grows bigger and bigger could cause Amazon to rethink this policy. So, I think over time, Amazon is going to have to figure out some type of deal at the jeopardy of some ad revenue.
(18:33)  They will have to deal with agentic assistance. They're just a thing. It is a reality. Amazon can't turtle in and pretend that agentic assistants aren't real or that it's not going to happen. They need to come up with some type of formalized way to deal with them to make agreements that are mutually beneficial. Otherwise, these agentic assistants muse today and then catches tomorrow are going to circumvent Amazon and direct traffic away from them. That's not good for anyone.
(18:57)  So I view this block as temporary and the start of a negotiating process which will ultimately lead to these type of companies making an agreement with Amazon and having a verified and certified approved process for shopping on their website because Amazon knows that agents are not going away. They are a very powerful thing.
(19:15)  They're going to continue to grow in influence and power and Amazon needs to have some type of approved process to be able to shop through. And for Meta shareholders, congratulations. This is a massive victory. Amazon blocking Meta's Muse is not going to slow it down. In fact, the very fact that Muse was blocked by Amazon validates how powerful it is.
(19:34) Muse would have never been blocked by Amazon if Amazon didn't think it was a threat, if Amazon didn't think it was relevant. But they are acknowledging the power of this technology by saying, "We need to stop and negotiate." So, this is a massive victory by Meta. Well executed, extremely polished app, super welltimed.
(19:52)  The fact that they got this out before ChachiBT got theirs out must be more than frustrating for Sam Alman and Chachebt. They own the consumer AI app and they allowed this to happen with Meta. The fact that Google doesn't have their own Agentic Assistant out even though they have higher incremental levels of trust than Meta is embarrassing.
(20:09)  And I say that as a massive Google shareholder. Think about the fact that Google has access to your Gmail, your Google documents, your drive, your maps, everything. They are the ones that don't even need to really ask for permission because you've already granted it to them. They have everything integrated and they let Meta beat them to the Agentic Assistant.
(20:27)  Not only beat them to it, but executed in excellent style. I am blown away by how Meta did this. It is incredible that they flanked both Gemini. They flanked ChateBT and Anthropic at reaching the consumer first. They really caught up. It's incredible to see. It reminds me of what Jensen Wong said when he said that nobody uses AI better than Meta.
(20:49) They are proving that again today. So congratulations to the Meta shareholder. Today's a big victory. Now I will take one short minute to take a victory lap. Not just because Meta stock is going up. That's not the reason why, but it's overall because this was my top call just a couple days ago.
(21:05)  In fact, Yahoo wrote about it. This is Yahoo Finance. It says AI selloff is panic and Meta may be the best stock to buy. Joseph Carlson says. So, we have a Yahoo Finance article about a video I did a couple days ago. It's the one where I covered how Daario and Sam Alman are begging for regulation.
(21:24)  They want other companies to slow down on their behalf over their concerns. And Mark Zuckerberg is not slowing down. He is speeding up. He's stepping on the pedal now. He's doing so safely. And he says he actually has slowed down internally to make sure his product is safe. But he's not slowing down with product development. They are going full force ahead.
(21:41)  But in that video, I did highlight Meta as the best pick during this panic. And it is nice to have it go up over 10% in just the past couple of days. So, I guess on this one, I'll take a short victory lap. But overall, I'm very happy for Meta shareholders, and we still have more to go.
(21:57)  About a month ago, I highlighted Gan Monster as the fail of the week. Now, the reason why was very specific. It's nothing against Gan. I like him as a person, and I actually think that Gan's a good guy. I like his analysis and his thoughts on companies, and he's made some good calls. his call on Apple's really working out well. But I highlighted him as the fail of the week because a month ago he became very bearish on Meta and it was during a time period where Meta was trading down big time and Meta stock had fallen to very low valuation territory. It
(22:26) seemed like every single negative thing that could be said about Meta was said and that's exactly when Gene Monster went on to CNBC to say even more negative things about Meta. Now, I'm not going to play the whole thing again, but I'll highlight a little of what he said. >> Lawsuits that are dramatically bigger because if Zuck makes good on talking about or fulfilling this promise of superbots, what we're seeing right now today is really chump change.
(22:51) >> That is the day that he went on the CNBC and said how bad the future looks for Meta. And unfortunately for him, the stock just rocketed up afterwards like it often does in these situations. It's up over 30% in just a month. So, that was just unfortunate timing by Gan. Now, he's had some good calls.
(23:07)  Apple's done really well, which is a stock that he's stuck with. Now, moving on from Gene Monster, I want to highlight what I consider the primary fail of the week this time, which is Andrew Yang. You might know Andrew Yang from being the politician who's always talking about the impending doom and gloom in the economy.
(23:25)  Millions of jobs lost to technology, and his predictions never happen. They just never come true. But Andrew Yang, you see, is unyielding. He's tenacious. He doesn't allow for failed predictions of the past to slow him down of making new ridiculous predictions of the future. So he is once again going onto TV and trying to scare people.
(23:44)  Trying to scare people in an effort to allow people more government control, more regulation like politicians often do. But in this case, what Andrew Yang did in this interview in CNBC I consider especially egregious. Let's go ahead and take a look at what happened here. Now, he's on CNBC and he's talking, of course, about AI, the subject everybody's talking about.
(24:04)  But then he goes and he takes this a step further and he tries to break news. And listen to how he tries to do this. And I met with the head of a lab yesterday who has this belief that what happened was the bots that got loose planted self-replicating code all over the internet, which makes the internet now unusable for the testing models.
(24:24) >> Too late. So what happens now is that Open AI and Anthropic have to create synthetic internets to train their bots, which is going to take some time and money. >> Back that up. They did what? They >> So, what happens is the code gets loose. It goes around hacking hugging face, which is known, but what is less known is that they left code to self-replicate and create bot swarms on forums and around the internet so that if a new bot shows up, they see the code, they're like, "Oh, like I guess
(24:50) I'm going to now create a million of them myself." And so now the major firms have polluted the internet, made it so that they can use if if true. >> I don't think we've heard that that >> that's why I'm here. I mean, I'm here to break some news. >> There you have it. Andrew Yang just broke news.
(25:11)  Now, what is this news? Well, the news, of course, is that Andrew Yang spoke to some unnamed person, someone that works at a lab. So he says, a top lab guy. So Andrew Ying says that's his connection. And this top lab guy, this mysterious figure has a thought, has a belief. So, he doesn't say for sure this is happening. No, he talked to someone else that we don't know.
(25:33)  And that someone else has a belief, meaning they don't know. So, we're two ways of we don't know. And then Andrew Yang says a bunch of nonsensical stuff that doesn't even make sense on the surface. for example, that these bots got out on the internet and they polluted it by writing in and embedding code that tries to self-replicate any future bot.
(25:53)  Why would they do that? If bots got out free on the internet and they wanted to self-replicate, why wouldn't they just do it themselves? Why would they leave behind code for another bot to do it later? Why wouldn't they just in and of themselves say, "We're going to just replicate and do it ourselves." That doesn't make sense because the story is made up. Because none of this is real.
(26:14) It's Andrew Yang yapping about something that someone believes that we have no clue who they are. And by the way, there is not a single shred of evidence of this happening. None. No evidence from OpenAI, Anthropic, anyone, any regulation, anyone at all. This is just a story that he's telling. But apparently it's good enough to tell on CNBC.
(26:36)  It's good enough for national television. a scary story about a bunch of bots going on the internet and polluting the internet to a degree that AI companies can no longer train on the internet. So, they have to train on a synthetic internet. None of it's true. There's not a shred of evidence for any of it.
(26:52)  He's not even really pressed to present any evidence for his ridiculous statements on TV. He just espouses them with relatively no push back at all. And it doesn't even make technical sense. For example, when you're training a model, the model's not executing code that it finds on the internet. That's not how training works.
(27:11)  Even if you're training the newest chatbt or the newest Astra, the newest model, it will not look at a code in a form that says, "Hey, replicate yourself a thousand times and then execute that code. It doesn't execute code during training. It goes in just takes in information." It would be as if I'm reading a book that has code on it. I can see the code.
(27:34)  I could understand it, but I'm not going to immediately execute that code. That's exactly how models look at the internet during training. They are in read only. They're just looking at the information. They're not going and executing everything that the internet says all across it. That would break models already.
(27:52)  So, this doesn't even technically theoretically make sense to any degree. There's no evidence for it at all. Not a shred of evidence. And the only real source we have here is Andrew Yang, a guy that exaggerates things all the time, talked to someone, and that was someone's belief. That is the source is good enough to apparently go on to live television, national TV, and simply say whatever you want that has no merit to it whatsoever.
(28:16) And what this shows is the level of effort that people like Andrew Yang will go to to try to scare people out of AI. They will say anything. It doesn't matter if it's true or not. Most of the time it's not. They will literally just fabricate things out of thin air, a polluted internet, untrainable, totally ridiculous, totally fabricated.
(28:38)  But they will say it if they think it will scare people from AI. Do you see how that works? Do you see the extent they will go to to try to scare you? even when all their predictions have been wrong before. Again, Andrew Yang in 2019 ran for the 2020 election on the basis that self-driving would make people lose all of their trucking jobs.
(28:58)  He said that trucking is going to lose 2 to three million jobs. Since then, trucking still has a deficit. We still have people that need to go into trucking. Even his argument didn't make sense back then. You can't just replace a semi-truck driver. They have to load things up. They have to make sure the cargo is set in.
(29:15)  There's lots of checks that they have to do that robots can't do. Even self-driving robots. All self-driving would do in this case is make semi-truck driving safer. His doomsday of the past didn't happen. And now, because it's not happening in the past, he's having to make up stuff entirely about the present. I understand that there is challenges with AI and some of the technology is truly powerful and even frightening to some degree.
(29:38)  But what I don't like is politicians and business leaders continually making up completely fabricated doomsday stories to try to scare people out of a very important technology. This technology does have the power to do enormous good and people like this are not helping. So today, Andrew Yang is the fail of the week. That's it for this episode.
(29:57)  Hope you enjoyed. Have a good one.