And I want to say: «Mark, are you in 2023, or where exactly are you? What recipes and grocery orders? What are you even talking about?» Just a few days ago Mark Zuckerberg published a position piece of about six and a half thousand words titled «The Future Belongs to Everyone».
And at the same time Meta presented a new artificial intelligence model, Muse Glimmer. You might think this story is not especially important, particularly in the current technology market for artificial intelligence, when we are seeing the extraordinary leadership of OpenAI, Anthropic and Google Gemini. And Grok from xAI is far better and stronger than anything Meta currently offers in artificial intelligence. But there are a number of nuances, and I very much want to highlight them for you. This subject is extraordinarily interesting, and it is interesting because Zuckerberg does after all own an extraordinarily large social network — both Facebook and Instagram — and the WhatsApp messenger, and he is one of the richest people in the world.
And he is young enough to keep developing the company for decades more. And he clearly plans to do so. The subject is also interesting because when he released this, an article came out — well, he released this publication, and an article came out from Business Insider, which headlined the material roughly this way: Zuckerberg's assumption about artificial intelligence for everyone runs into a big problem. Today in this episode I am going to show you that dividing line between what the world's leaders broadcast about artificial intelligence.
I am not even talking about presidents, or politicians, or economists, or bloggers. What these people broadcast about artificial intelligence, how far what they are telling us is real, and at the same time whether they will be able to do it at all. That is, whether the possibility even exists of building what they are discussing, what they are describing, and whether it is anywhere in reality at all. So, from Business Insider's point of view, why did it say Zuckerberg's proposal runs into a big problem?
Because there are of course a number of very large factors. An enormous number of people do not trust Meta. People do not trust it, and for obvious reasons. In particular, just the storage of all this data and the study of the data, and Meta's own interests. People fear losing their jobs, and people are, of course, very seriously worried about personal data, very seriously. And this piece Zuckerberg has put out was very strongly tied to the construction of data centres. He even said that we will create a one-billion-dollar fund. And people who live in the places where data centres for artificial intelligence are being built — in those places people will receive large, substantial grants.
He even talked about some of the grants. I will show you those examples too. And of course people resist the construction because of problems with electricity, water, the insane noise, the tax problems that can arise, and the load on their local infrastructure generally. And by the way, I want to say that the noise that arises is extraordinary. Quite recently, commenting on the data-centre question, the president of the United States said — and I hope the editors will show his exact words here.
He said that you have to come to terms with it. Come to terms with the fact that data centres for artificial intelligence were being built and will be built, and you have no choice. Because you have to convince your community of how great this is. There is no arguing with it. You simply need to accept it, and you will have all this protection, and we will automatically provide you with protection. But you will make enormous money, you will cut taxes, and you will have what others do not have but everyone wants.
So these are the smartest people. Most of you, most in this room — though there may be a couple of people on the other side who disagree. But I do not understand how they can disagree. Your community and your state will become richer, and that is very little. This is a breakthrough. And again, if you do not take all this money, someone else will. So it is better to do it yourselves. Now about Zuckerberg's main idea. You know that on the ToTheMoon channel we talk about ideas, including other people's ideas.
Obviously ToTheMoon broadcasts a great many opinions and a great deal of knowledge of its own. I personally broadcast an enormous amount of my own opinion and my own knowledge. That said, in one of the episodes a few months ago I talked about the philosophy of Anthropic. It is very interesting to study all those core documents of theirs, and the doctrines, and their constitution. That is genuinely very interesting. So then: Zuckerberg proposes not concentrating the most powerful artificial intelligence in a few corporations and states, and making it available to the widest possible number of people.
And his thesis is that when we talk about a personal superintelligence, every person should have their own personal agent that knows their goals, health, work, family and interests. A very interesting subject right now, in the present day, because all of this has already appeared, hasn't it? It really has genuinely appeared. Of course we can say that ChatGPT today does not know all the data about my health or all the data about my finances. So, for it to know all of that, what would actually have to happen? Do I have to hang myself with sensors, or what specifically has to occur?
Because on the whole ChatGPT already knows a huge amount. And when Zuckerberg talks about this, he is in effect showing that we are not yet in the zone where very serious systems are being developed. And that all of it is still ahead. Let me remind you that Meta lost its leadership in the artificial intelligence race quite extraordinarily. It was present there. That is, back at the start of 2023 Meta was a fairly serious player, especially in the open-source market. And discussing artificial intelligence as an instrument of creation, Zuckerberg says the main value should be not in laying people off but in creating new products, creating medicines, creating companies and ideas.
But if we look at life in general — not even at artificial intelligence but at various technologies — we will see that there are people who are interested in using technology for people's benefit and in bringing them something useful. And there are people who are interested in using technology, or power, or money in their own interests, exclusively in their own interests. And there are no repetitions here. There are an enormous number of rich people who help an endless number of poor ones.
There are rich people who help nobody and calmly buy their forty-eighth boat for half a billion dollars.
Obviously this is something of an abstraction. Speaking of small companies, when the thesis is voiced that small companies will become stronger — we will take that thesis apart now too. What do you think about that thesis, by the way? I am going to tell you a number of things. Today's will probably be a philosophical pace from me, and I would be very interested for you to add your opinions as we go. You can always hit pause, add a comment of your own and carry on watching. Or write something in parallel.
Different people's opinions are very interesting here, because this thesis — «small companies will become stronger» — is a very interesting one. Zuckerberg's logic, as it were, is that small companies, even a single person, will be able to build businesses of significant scale. And we all know this story that is being told endlessly right now, that one person will be able to build a company worth more than a billion dollars. And some examples of that are probably already appearing. I have said many times that a billion-dollar valuation is very abstract, because there have already been millions of people in the world, probably tens of millions of people, who built companies, and at certain moments those companies really were worth a great deal.
And a billion dollars today, for the technology market, is a very relative story. And is that a plus? When someone says small companies will get more opportunity, I can say that small companies will get more competition. I have a clear sense that over the last thirty years — I first went into business twenty-four years ago, though that was obviously kindergarten stuff, trading phones. And professionally I started in 2007, which is nineteen years ago. And I can tell you that back then doing business was far easier than it is now. It was easier precisely with less technology.
Because if you had extra knowledge, if you had the desire, the initiative — while at the time people who also had extra money, as for example in the nineties in Russia after the collapse of the Soviet Union, or had power, or had certain connections — doing business was far easier for them. Today having money says nothing about whether you will build a successful business. Today, if you have knowledge, that says nothing about whether you will build a successful business. Do locations exist on Earth where it would be a plus? They do.
But in developed countries, in developed economies, in a highly competitive environment, it is far harder to do. In my time, in my youth, there were not this many coffee shops. I often say, and I have said to my children, that I do not remember at all whether we drank coffee, whether our parents drank coffee, whether anyone went out for coffee. Now the number of coffee shops is simply extraordinary. Obviously the number of coffee shops grew, demand grew and the number of people who consume that coffee grew.
But the competition is extraordinary, even compared with fifteen years ago. Fifteen years ago opening a beautiful, tidy, excellent coffee shop or a good restaurant was obviously far easier. Or winning the competition for search promotion on a website. What is called search engine optimisation, SEO, was far easier to win. Or launching paid search on Yandex.Direct in 2012 was fairly simple and easy, and gave a decent, good result. Today, launching advertising on Facebook, on Google, on YouTube, on Yandex, on — it does not matter where, on X, on Telegram — the probability that you will make money from a normal, simple launch is very low. Even if you set everything up superbly, it is small; the probability of making money is small.
Obviously in working processes, in a normal economy, with normal competition, it works — but again, in a normal economy, with normal competition, when we are talking about advertising on the internet. Well, that is not competition. Here I am about to go and shoot a video and drive into town. I have a meeting, I will walk around and buy a coffee there. That is not competition between the five coffee shops in the town of Los Altos. And Los Altos is one of the most expensive towns in America, and in some rankings it was the most expensive town in America.
A village, essentially. You would think the richest people in Silicon Valley live there, and yet it is not as though fifty coffee shops have opened, or as though those coffee shops are golden. No, they are broadly exactly the same as everywhere else in America. And there is not a single special coffee shop or special place there, not one. Everything is very ordinary. So, what else does Zuckerberg raise, what other theses, so that we can build the full picture — there is the subject of open-weights models.
And in Zuckerberg's story users and developers will be able to download models, modify them and run them locally. And everything to do with open source generally is an extraordinarily discussed subject lately. Jensen Huang has spoken about it a great deal. Then Mira Murati spoke about it, saying open source is essential, it is absolutely necessary. And what does that mean? It means you can take a model, install it on your computer and, notionally, pay nobody. Notionally. Notionally pay nobody. But we have to understand that a genuinely good product will still cost money.
Do free, high-quality, excellent products exist? That is a big question. Even if you think about Wikipedia — supposedly a free foundation, a free act of charity — today there are an enormous number of accounts saying it is a portal sponsored by certain political and not only governmental structures in the United States. I know that what gets published on Wikipedia has always been moderated inadequately. It is not for nothing, by the way, that Elon Musk began building his own new project — not as a copy of Wikipedia but as an alternative to Wikipedia.
One where the data will be real and honest. And by the way, on open source: Elon Musk wrote literally a week ago that starting in about a month — and I think we are coming up to that time, it is probably the end of August or the start of September — they will be publishing all changes in open source, in open code, relating to X, to the X network, to the X social network.
I do not know whether this will be connected to xAI in some way or not. We shall see. And what does it mean for an ordinary person? Does it give an ordinary person any understanding at all? Or does it give some people the chance, in effect, to study and analyse things a bit more? There is importance in that, of course, on the one hand. On the other hand, if it degrades quality and prevents us from developing and obtaining a genuine, seriously powerful artificial intelligence — which is supposedly being built for the good of humanity, and it seems to me that open source will not allow that to be done, for a whole range of reasons — then it is a big question.
I am sure the amount of open source will grow, but for one simple reason: because open source will always be cheaper. There are still costs for infrastructure, for your own hardware and so on. There is no reason for illusions here. Even if you want to install some open-source model on your own computer, you need at minimum your computer. It will not be a cloud system that keeps all this data turning twenty-four seven, that is built to survive failure and so on. For me, for example, that matters enormously. Where things sit and what hardware you do or do not put up.
Incidentally, the world has had and still has a problem with the internet. And the bulk of the world is not covered by good internet. You could say Starlink exists now, but that is a temporary, or rather current — not temporary but fairly small — solution as of today. The whole world is still not covered. Let us keep going. I want to devote all the blocks to this study — not study but description — from Zuckerberg, so as to build up more of the picture.
There is a piece that concerns the user's personal values. And in Zuckerberg's view, an artificial intelligence agent should help achieve the person's goals rather than impose the values of the company that built the agent. And here there is, it seems to me, a rather chapel-like conclusion, because we live in a society where people impose values on each other endlessly. And what does helping achieve a person's goals even mean? And who defines those goals of the person? They are defined by the state, they are defined by the person themselves. Who defines, who is going to define this person's goals at all?
And to what extent does the person agree with that? Because if a person was born in a country where women are forbidden to drive a car, for example, or where you have to cut your hair a certain way, or where men have to go and fight at a certain age — and if, say, they do not want to — then how does that square with the person's goals? That is, we live where we live in any case, and we are inside the constraints we are inside. And to me, personally, as a human being, it is very important to hear from the creators of artificial intelligence systems some sane reasoning with a sane understanding of where we live, where we exist.
When people start telling us abstract things about raising children — or, by the way, when the head of Meta, and I will keep saying this tirelessly, said he would only have children after excellent chips appear that can be implanted in a child so that the child learns everything quickly. A question arises for me: are those definitely the person's goals? And does Meta actually follow these things? How does that square with itself, then? How does it square at all? And there is one last area, probably, one of the last areas Zuckerberg lists within his focus: the distribution of power.
He says that several competing agent models are safer than one — than a single centralised superintelligence. And here a very interesting story arises. This is August, it happened literally a few days ago. And Zuckerberg states the following: he directly contrasts Meta with other labs which, in his words, mainly build artificial intelligence for companies, governments or, I do not know, some large organisations. Whereas Meta, you see, is trying to take the position of a company creating personal artificial intelligence for billions of ordinary users.
Well, first of all, Gemini reported literally a few days ago that they have a billion users of the Gemini app. For ChatGPT there was information a month ago that they had just approached a billion users. So in any case those two systems are probably the foundational systems creating artificial intelligence for all people. And by the way, let us be honest: OpenAI, and Google with Gemini, and of course Anthropic with Claude — it is just that fewer people use Anthropic, not a billion people.
Anthropic with Claude ran an extraordinary number of initiatives around the world to put their systems in front of people as widely as possible, everywhere possible, in particular for free. A great many serious free things. And what is Meta's model quality today, and where are their free models? In WhatsApp, inside WhatsApp, or inside Instagram, or inside Facebook? They have a separate app, but who actually uses it? Probably only employees of Facebook use it. There are enough of them to drive the downloads or install it. I am not sure they use it themselves.
I have not heard from Facebook employees anyone insisting with froth at the mouth that they use Facebook's apps. And another question arises about Zuckerberg's reasoning. Well, he is a supremely intelligent man. He talks about the goals of artificial intelligence, and he uses the word «agent» there.
Look, this is very interesting. Again, 10 August 2026. And he says the following: that artificial intelligence should work round the clock on the person's behalf, that it should help analyse health and sleep, help with career, business, finances. Think about it — I will try to list all of it. Organise affairs and the household. He uses things like picking recipes and ordering groceries. I am already starting to laugh, but I will still go through the rest of it with you. So: see the surrounding world through Meta's smart glasses.
Here we get a rather specific story. And work with personal documents, messages and files. And I want to say: «Mark, are you in 2023, or where exactly are you? What recipes and grocery orders? What are you even talking about? Especially when you are talking about a billion people. What grocery order? Or does life in Palo Alto, where every house costs from two million dollars — well, he simply lives there, and he has bought up, I think, fourteen houses nearby — does it seem to you that the whole world lives like that, that the whole world orders groceries to the door and that someone has a problem picking recipes?
Do you really think that?» Where are we right now, you and I?
We are right now in August of 2026, when we have systems like Codex from OpenAI and Claude Code that build entire infrastructure solutions and extraordinarily good software. There are systems that solve questions that have nothing to do with picking a recipe or analysing sleep. These systems are starting to work in place of a doctor, starting to save an enormous amount of money, starting to carry out an enormous number of different processes on a person's behalf. The other side of it is that a person becomes dependent on them. I have said it and I say it again: a video will come out on our channel, and it has probably already come out yesterday.
My sister filmed it, and she filmed it on a very interesting subject, in particular about how when artificial intelligence was appearing everyone said that everyone's time would be freed up. I remember Sam Altman saying a phrase to that effect. Well, one wants to lift them up and say: «Were you speaking genuinely, or was that marketing?» If it was marketing, fine. But if it was genuine, I have a big question about who is building these systems. He said artificial intelligence would give people the chance to attend to their spiritual life, because they would have time.
To which I personally objected and said that you should attend to your spiritual life regardless of whether you have time or not. But that is his interpretation, presumably, and his understanding of different positions and states, of what spiritual life is. He probably views it like a gym. And this is exactly what my sister talks about — look at it — about how everything initially pointed to artificial intelligence freeing up an enormous amount of time for people, so people could rest, spend time with family, do all sorts of things.
If you look at the narrative everyone broadcasts now, we see exactly the opposite. So what troubles me? Coming back to the story I was telling: what troubles me is that Mark Zuckerberg is talking about an artificial intelligence that is not a superintelligence, and about an artificial intelligence whose wave has already passed. And right now the mode of what is happening is completely different. Because what we need to think about and watch is when these systems get embedded into various kinds of hardware.
To abstract away and talk about Meta's smart glasses is simply funny. How many people today have smart glasses from, I think, twenty well-known global manufacturers. Starting with Apple, Meta obviously, Samsung. The question is: who uses these smart glasses, and how many people actually wear them at all? Simply micro. Interesting again — look, folks, today's episode is on the one hand philosophical and on the other very much an educational one for many people. And write in your opinion and your view and your perspective. And even where I speak up and say it is strange to me to hear about some person or something, I have no grievance against that person.
I am rather taking things apart, and sometimes in a different tone of voice, or perhaps emotionally taking some things apart. But I am trying to convey and show things to you, including my own opinion. Mark Zuckerberg promises that Meta will have a fully private mode in which even Meta supposedly will not be able to — and it seems to me it is very important to put in that «supposedly» for yourself, and I have put it in — will not be able to read the user's information or present it to anyone.
And yet the technical architecture of this mode, the launch dates, how it will work, what devices will be needed for it — of course he did not publish any of that. And right now it is simply a declared direction and an unfinished system. And we all know in general that they all change their minds, all the leaders change their minds, change them a great deal. There are some people, of course, who go about it quite deliberately and very precisely. And of course you have to filter a lot of things when it comes to what they say.
In Mark Zuckerberg's description there is a large section that concerns helping people around data-centre construction. And also the creation of this one-billion-dollar fund. But when I read about a billion dollars, to be honest, I thought the number said trillion. At first I thought: «Not bad, he has swung big, saying we should create a trillion dollars for people within artificial intelligence projects.» But then, in those materials, there was information about Meta's revenue and information about the fact that they are spending, what, a hundred and fifty or however many billion dollars on infrastructure — specifically on artificial intelligence infrastructure.
And it was written that this billion dollars is nought point seven per cent. At which point I caught myself: «Listen, so this is not a trillion dollars, it is a billion.» And I stopped taking the figure in at all. Because when we are talking about billions of people, a billion dollars is absolutely nothing. And even when we are talking about tens of millions of people, a billion dollars, honestly, is nothing in the United States of America. He was still saying it in — well, there is no description there of whether it is the United States or not the United States.
But when in the United States of America the minimum wage in California is eighteen and a half dollars an hour — well, a hundred dollars of support spread over ten million people. What is there to talk about? You earn that money absolutely easily in five hours. He gives examples there: in Louisiana teachers received one-off bonuses of up to fifty thousand dollars thanks to the construction of an enormous data-processing centre. Though a question does arise for me: the teachers received it, but how do the other people who hear the noise live?
I am not, by the way, against building data centres. I think they should be built. Presumably they should be built more sensibly, more reasonably, with some understanding of all the circumstances. Because everyone rushed to build data centres, Meta in particular, and Google, and Amazon, and of course SpaceX and Tesla. They all started building this very chaotically, it seems to me — very chaotically, very fast and very abruptly. On the new systems Facebook has released: it seems to me this is either the end of an era or the beginning.
They did say their new Muse systems are supposedly superb. That terminology, by the way, is not much used. I am sure it will be used when systems like ChatGPT, Codex, or Claude Opus become your operating system, fully the operating system of your computer, and you essentially use it, and with it you move files and folders, and store photographs, and send mail, and do absolutely everything. That is, you do not need connectors, you do not need browsers. And by the way, the story about connectors, the fact that there are different connectors, is a big question in itself: how it will be rebuilt, what will happen to it.
But this Muse Glimmer was released simultaneously with Zuckerberg's manifesto. It has open weights, obviously, and thirty billion parameters. I do not know whether an understanding of thirty billion parameters gives you anything. Parameter counts tell me absolutely nothing any more. This system was released on the basis of Muse Parque. But I will note that Muse Parque is a larger model, a very serious model, and the bulk of humanity will of course use models over the internet.
For free, in the same way as people use — well, as is happening now, everyone uses WhatsApp or Telegram, or Instagram itself, for free. In the same way people will of course use artificial intelligence models for free. Will there be advertising there or not? That is something we shall see and understand. Because in WhatsApp itself there is no advertising, and there are no paid or unpaid tiers there. I do not know whether advertising on free tiers still exists in Telegram now, but advertising in general bothers me, for instance.
And even the presence of the strip at the top in Telegram bothers me. I do not know. Write in and say who is not bothered when the people you have in your phone — you endlessly see their stories, and you do not want to watch their stories. There is a whole series of things Zuckerberg states in the manifesto.
It seems to me they are the politically-corporately-NASDAQ-correct kind, meant to raise the value of his shares. For example, it says there that intermediate versions of the most powerful models must be handed to the government before training, before or during training, so that specialists can look for vulnerabilities in advance. So the question here is: if the model really is that good, that superb, what specialists will be able to find vulnerabilities? And it seems to me that Mark Zuckerberg is playing certain games here, because the very best, most powerful models happen not to be his.
He says the state will need to be given company engineers to protect energy, critical infrastructure, possibly to analyse these models. That is of course a very good and interesting thing. The question is what happens if we look at the corporate market and the war going on there. And by the way, Meta, like nobody else, also took part in that war of poaching people onto large salaries, offering people something close to more than a hundred million dollars in salary, or compensation, over a year or over five years.
Right now a large number of people have again left OpenAI, and you can see that the people leaving are often in a negative state, and negative discussions are going on there. Apple sued OpenAI. Elon Musk sued OpenAI, in part over the departure of a number of employees. Obviously, finding employees. In short, everyone endlessly sues everyone else. And so, will engineers work with the state? And there are clearly government structures there, secret services, that are interested in getting even more data out of these companies and, as a consequence, more data out of people.
What will happen? As we know, in Europe there is GDPR and there is supposedly protection, the protection of a person's individual rights, very strongly, really very tough. At the same time we repeatedly hear statements from American companies that the governments of European countries are demanding full access to all the messages inside their systems. And that raises a big question for me. Hold on — so if you do not let the corporate market work, and in a CRM system I can barely save a client of my own — well, it is bedlam.
Or those idiotic cookies. It is impossible to build any products properly in the market. And yet you yourselves apparently have access. And these «you yourselves who have access» — who is that? There is another proposal from Meta: not to impose long mandatory government reviews if, because of those reviews, American companies suffer relative to China or fall behind China.
And here, post factum, there is a very interesting question. We know there are radically different opinions from Jensen Huang at NVIDIA and from Amodei at Anthropic; they point in different directions on how to compete with China. Whether to open up to China or not to open up. Again — what do you think about that? Let me note once more that the world market currently registers two waves, the US and China, in terms of players in artificial intelligence in particular. There are always users who write that some other countries also have leadership or have advantages.
I have a big question about that. Even about China I have a big question, because will China be able to build a system that two billion people will use? That is the question. Or even a billion people. Chinese systems today together account for fewer than half a billion users. And I am talking here about use simply as apps, like ChatGPT. And there is no reliably accurate data on how far the Chinese use Gemini or ChatGPT or Claude. A lot is outright blocked, a lot goes through intermediaries, a lot is what systems are trained on. Remember there was a case I talked about, that there are special corporations in China that buy Claude and resell tokens ten times cheaper to other people in order to train on those other people.
Though these are ordinary people — well, not people there, they are simply used as bots. Zuckerberg also proposed allowing models to train on the outputs of other models. And this is fantastical, because Anthropic recently filed a statement and a complaint with the US Senate that Chinese models — in particular, I think, from Alibaba and maybe from Tencent, I do not know who else was there — trained some fabulous number of accounts, trained on Claude models and distilled them.
And it seems to me this position is weak when it comes to distilling other people's models; it is interesting in some ways, but it is weak. And still, everyone has fairly equal access, on the one hand, to open data, and on the other hand we have entered precisely the era where data matters. And if it turns out I am inside ChatGPT and ChatGPT suddenly trains on my data, then Meta will also train on my data, or European companies will train on it, or who? That is a very serious question. I hope it will actually be raised and noticed.
I do not see that everyone has spoken about it. I do not even see that Elon Musk has spoken about it, because the subject is very strange. On distilling other people's models — very strange. If you read the rules: if any of you build on Facebook and Instagram, can other companies train on Facebook and on Instagram? Look, for example — Elon Musk has the X system, and you may not train on it, and you may not download or scrape the system. That is directly prohibited. Fines of tens of thousands of dollars. And connecting via the API is fairly expensive.
That is, if you want to scrape inside it, that is the API, and it costs a decent amount. I needed it for one project and used it. It is not that you take it and can scrape all of X endlessly. Zuckerberg says restrictions on supplying advanced computer chips to various geopolitical competitors should continue. But obviously this story has held and will hold, especially given Facebook's wars with various countries: with Eastern European countries, and there was a big problem with India recently, and with China and so on. And Zuckerberg has a piece about giving independent members of Meta's board the right to approve safety criteria for releasing new models.
A very interesting aspect. I have told you about such bodies existing in various companies, and quite recently I talked about how an independent service that analyses other companies' artificial intelligence did a fairly large study at Meta. Very interesting. Take a look at it.
Once again, importantly: this is not simply a philosophical text by Mark Zuckerberg, and it is not charity in the sense of one billion dollars. Meta has a very concrete, plainly commercial interest. Meta does after all own Facebook, Instagram, WhatsApp, and it owns one of the largest audiences in the world. And so to build this personal assistant into those systems — which is what Meta will in fact do, because if you use their apps, it is built into their systems — is full integration, and that is serious power for the Meta corporation.
Or do you think something different about it? Share this video, subscribe to our channel, watch the other episodes. Bye, everyone.