There Will Be No More Search: How AI Is Changing the Internet and People's Decisions
What replaces the search bar when ChatGPT, Gemini and Claude no longer look for links but solve the whole task — and what does that change for Google, for business and for an ordinary person?
Search as we know it is leaving not because Google or Yandex will disappear, but because the habit of typing a query and reading links is disappearing: ChatGPT, Gemini and Claude already solve the whole task. The host shows it in numbers — Google AI Mode and ChatGPT have a billion users each, Perplexity tens and hundreds of times fewer because it builds search on other people's models — and in his own case: ChatGPT built his daughter an SAT preparation plan, explained that 1250 points is the top 20 of test takers in the US and 1650 the top 1, and worked as a team of consultants rather than a search engine. Google holds on to search for the sake of contextual advertising, but Alexander Volchek himself has cut his use of it a hundredfold; Google's new SEO rules are two years late, and the "new era of AI search" and Alisa AI from Yandex are already the past. The upshot: the notion of "googling" disappears, and AI becomes a way to make sense of life, work, study and money.
What to watch for
Key takeaways
Google and Yandex will not vanish tomorrow, but the habit itself will: opening the search bar, typing a question, getting a list of links, reading and comparing yourself. A person often comes not with a precise task but with chaos — "sales are not coming in", "I don't understand where to go" — and a strong artificial intelligence will not simply answer but help formulate the real task: the cause, the constraints, the needed result, the risks, what to check and what to come back to in a week or a month. It is a new layer of life and work, and for a person a new skill: not asking "find", but working through your own situation together with AI.
The president of the Russian Federation recently said that Russia is definitely in the artificial intelligence race and can be among the leaders: it has people, science and electricity. The host found it telling that he sees AI infrastructure only through electricity. Owning gas, oil and electricity and developing nuclear power provide opportunities and a certain base without which many things are genuinely hard to do, but they do not give real leadership positions — unlike the companies with very serious infrastructure investments, to which the host promises a separate episode.
Perplexity's model always puzzled the host, although the company is cool and he even made money on it. It built a third-party search that uses other people's models, including OpenAI's and Anthropic's, then made a browser and agents, but it does not own the models themselves and remains an external system. It has tens, if not hundreds, of times fewer users than the leaders: Google Gemini has crossed a billion in AI search, AI Mode has more than a billion weekly users, ChatGPT roughly a billion. Perplexity will not get hundreds of millions of users; the giants do not need it, but it will certainly sell for some amount of money.
Just a week ago Google announced new SEO rules. Two years ago, when artificial intelligence appeared, the host, as an adviser and owner of many businesses, was already telling everyone: search engine optimization will work differently, and right now you can seed the internet with information about your companies and about yourselves that artificial intelligence will eat and read and treat as important and necessary. People started listening later, coined special terms for optimizing for AI machines — and are doing it now instead of two years ago, assuming people will still search in search engines.
In the host's view, search and contextual advertising in their current form will disappear incredibly fast — as will a huge number of sites and portals: worldwide consumption of sites with run-of-the-mill information has fallen tens of times, and an education project that traded at around ten billion lost its value entirely because it helped answer questions. Startups that simply want to present information more beautifully are unlikely to do it better than ChatGPT or Anthropic. Even Excel the host now studies and builds through ChatGPT and Claude, almost never opening the files.
Google is the number one search in the world: China has a different infrastructure with super apps like WeChat and portal systems like Alibaba. It is in no hurry to remove its standard search entirely, because it earns insane money on contextual advertising, but it will have to: terminals and plain command lines like ChatGPT are stepping on its toes. The host himself has cut his own use of Google not tenfold but a hundredfold already, and will drop it altogether as soon as ChatGPT takes an interest in building him an interesting interface — for now OpenAI and Anthropic are busy creating the real artificial intelligence of the future.
Google plays in different positions, as does xAI, which is creating a serious artificial intelligence in tandem with Starlink and satellites, Tesla robots and cars — an entire infrastructure. OpenAI and Anthropic can be placed side by side, but Anthropic made a different, incredibly strong bet, so its valuation is almost OpenAI's — lower, but around a trillion dollars. SpaceX has its own story together with xAI, and Google its own unique life. Yandex, in the host's words, is not in this race and will not be: it is a lost world, and whoever is present there will already be a different company.
Google's chief executive was shown a poor answer given by its AI in search and admitted it could have been better. But the problem, in the host's view, goes deeper: search has become a dump — both Yandex and Google are dumps — and yet they can give reasonably good answers, just not to what you specifically need. One person needs to quit his job while another must not; one needs to buy a car while another must not under any circumstances. A person goes to search as if for a small trinket, while even "recommendation" is too small a word: AI helps build criteria, a plan and control and come back to the task.
The host's daughter studies in Silicon Valley and can take the SAT two years before applying. He uploaded all the information about her into ChatGPT, and in January–February the system built a plan: what to do in spring and summer, first exams in August, a test every two weeks, and Khan Academy in between. In February the daughter scored about 1250; MIT or Stanford need around 1650–1700, and the system replied: it is possible to prepare, but she will have to push — right now she is in the top 20 of test takers in the US, and with 1650 she would be in the top 1. That is not search but an unbelievably professional consultant — a whole team that can be double-checked in a new session.
Google's vice president of search Elizabeth Reid called it a new era of AI search and the biggest update to the search bar in more than twenty-five years. For the host these are phrases about nothing: Google and Yandex were afraid to change the search bar for twenty years, and context is no longer needed when Claude and ChatGPT store thousands of your chats. Shivani Mohan writes of a billion monthly AI Mode users, US queries three times longer than usual and planning growing 80% faster, but ChatGPT's billion users are mostly free micro-queries, and in the host's circle only one in ten uses AI decently.
Demis Hassabis, the head of DeepMind at Google, describes Gemini as a future world model that understands how text builds plans and acts through a device — an attempt to build a universal assistant. Even OpenAI is removing the reasoning switches, because all modern models reason. On the eve of filming, Anthropic released an update to Claude Code — Opus 4.8 and 4.8 Max — with a function for working on code or on an entire system in hundreds of parallel threads. That is not an agent and not reasoning search but the global creation of something — like a person living through dozens of events at once in half a day.
Yandex says Alisa AI in search solves complex tasks, holds dialogues and forms answers with sources, but even that is already the past — like parsing PDFs and the size of the context window: the question is access to the systems, and the programmers at Codex and Claude are not limited to a million tokens. xAI, through Macrohard, will inevitably enter software development, because a watch, a phone or a robot needs a way to communicate, not search. Google will not disappear — Google search as such will go, and artificial intelligence is not "asking for a fact" but a way to make sense of life, work, study, money and your own choices.
What this episode is about
A solo ToTheMoon episode about how artificial intelligence is changing the internet and people's decisions. Alexander Volchek opens with a thesis: search as we know it will no longer exist — not because Google or Yandex will vanish tomorrow, but because the very habit of opening the bar, typing a question, getting a list of links and comparing them yourself will vanish. A person often comes not with a precise task but with chaos — "sales are not coming in", "I don't understand where to go" — and a strong artificial intelligence will not answer but help formulate the real task: the cause, the constraints, the needed result, the risks, what to check and what to come back to in a week. For a person it is a new skill; for business it is a question of visibility: the client of the future comes with a task that AI helps him understand.
The host keeps his eye on the leaders — OpenAI with ChatGPT, Anthropic with Claude, xAI with Grok and Google with Gemini — and explains why Perplexity is doomed: it has tens and hundreds of times fewer users than the big systems, whereas Google Gemini has crossed a billion users in AI search, AI Mode has more than a billion weekly users and ChatGPT roughly as many. Perplexity built a third-party search on other people's models, then a browser and agents, but owns no models of its own; it will not get hundreds of millions of users, the giants do not need it, although the company will sell for some amount of money. Along the way the host takes apart the words of the president of the Russian Federation that Russia is in the AI race thanks to its people, science and electricity: gas, oil and electricity provide a base, but not leadership positions.
Next come the new SEO rules Google announced a week ago. Two years ago the host was telling his businesses that search engine optimization would work differently and that information about a company had to be seeded so that artificial intelligence would read it; people are doing it only now and assume search in search engines will remain. Most still use ChatGPT as search — to check the weather — but ChatGPT, Gemini and Claude have long been solving the whole task, for instance planning a trip with the weather and all the circumstances taken into account. Search and contextual advertising in their current form will disappear together with sites: consumption of sites with run-of-the-mill information has fallen tens of times, an education project worth around ten billion has vanished, and the host studies and builds Excel through ChatGPT and Claude, almost never opening the files.
Google is the number one search in the world (China has a different infrastructure with super apps like WeChat and Alibaba), and it holds on to standard search because it earns insane money on contextual advertising, but command lines like ChatGPT are stepping on its toes; the host himself has cut Google a hundredfold. Every player has its own strategy: xAI builds AI in tandem with Starlink, Tesla robots and cars, Anthropic made a different bet and is worth around a trillion dollars, SpaceX has its own story with xAI, and Yandex, in the host's view, is not in this race and will not be. The problem with search is that both Google and Yandex have become a dump: they give good answers, but not to what you personally need, and "recommendation" is too small a word for what AI does.
The main case is the host's daughter, who studies at a school in Silicon Valley and can take the SAT two years before applying. The host uploaded all the information about her into ChatGPT, and in January–February the system built a plan: spring, summer, first exams in August, a test every two weeks, free Khan Academy in between. In February the daughter scored about 1250; to the goal "MIT or Stanford, around 1650–1700" the system answered that it is possible to prepare but she will have to push: right now she is in the top 20 of test takers in the US, and with 1650 she will be in the top 1. That is not search but an unbelievably professional consultant, a whole team that can be double-checked with a new session in ChatGPT or Anthropic; the system likewise helped choose tracks and insisted on advanced physics.
People who use AI seriously and build AI native systems are freed from search: the host himself opens Google only to look at the seats on a Boeing from Mexico City to Paris or to check his sites in the results. To the "new era of AI search" and "the biggest update to the bar in 25 years" that Google's vice president Elizabeth Reid speaks of, he replies that these are phrases about nothing: Google and Yandex were afraid to change the search bar for twenty years, and keeping the context of a dialogue is no longer needed — Claude and ChatGPT store thousands of your chats. Shivani Mohan's data — a billion monthly AI Mode users, US queries three times longer, planning growing eighty percent faster — do not cancel the fact that AI adoption is very relative: even in the host's circle only one in ten uses it decently.
Demis Hassabis describes Gemini as a future world model that builds plans and acts through a device — not search but a universal assistant. OpenAI is removing the reasoning switches because all modern models reason. On the eve of filming Anthropic released Claude Code Opus 4.8 and 4.8 Max, working in hundreds of parallel threads — not an agent and not search but creation, much like a person living through dozens of events in half a day. The notion of "googling" disappears; Alisa AI from Yandex with its documents and sources is already the past, as is the size of the context window: the question is access to the systems, and the programmers at Codex and Claude are not limited to a million tokens.
In the finale the host looks at strategies: xAI, through Macrohard, will inevitably enter software development, because a watch, a pendant, a phone or a robot needs a way to communicate, not search. Google will not disappear — Google search as such will go. To marketers, salespeople, entrepreneurs, investors and ordinary people he suggests understanding that artificial intelligence is not "asking for a fact" but a way to make sense of life, work, study, money and your own choices.
The episode's value is that the host proves the thesis "there will be no more search" not with a forecast but with practice: the billion-user figures for AI Mode and ChatGPT, the fate of Perplexity without models of its own, and his own case, in which ChatGPT replaced a search engine with a whole team of consultants. The main shift he records — from a bar with a query to a system that holds context and solves the task — hits SEO, contextual advertising and sites with run-of-the-mill information all at once. Google does not disappear in the process, but it loses search as an interface, and Yandex, in his view, is no longer in the race.
Episode transcript
The episode is in Russian; below is an English reading guide to the transcript (the full EN transcript is a machine translation). Voice matching applied to 59 segments: 59 identified, 0 mixed, 0 probable, and 0 unresolved.
Read transcript on a separate page
Loading…