Recently, the world was informed that Meta, after several years of strenuous attempts, decided to give up further investments in the metaverse. The $80 billion that was spent on this ambitious project was, of course, not completely wasted, but in the face of the global acceleration of trends related to artificial intelligence and how profitable this industry has now become, the change in priorities in Zuckerberg's empire should not surprise anyone. Meta currently sees the integration of AI in its internal environment as one of the key tasks in the near future and actively encourages all employees not only to develop new solutions for users… but also those that will improve their work.
Will AI replace the CEO? He should be able to perform some duties now
The example should come from the very top, because information obtained by “The Wall Street Journal” shows that for some time Mark Zuckerberg has not only been creating, but also actively using an AI agent built for his own needs, which is to help him run his company. According to people familiar with the matter — the agent is already able to perform some of the CEO's duties completely independently. Zuckerberg is also to use them to obtain information about individual projects in the company – where he would normally have to contact many people in the structure to obtain answers that satisfy him.
At this stage, it is not known what models and environments Zuckerberg uses to develop such a personalized “replacement”; it is also not entirely clear what tasks the AI agent is able to handle, but its presence at the very top of the Meta seems to be the case consistent with the strategy adopted by the company. At the beginning of 2026, while discussing financial results, Mark Zuckerberg directly admitted that he wanted his company to focus on developing this type of tools in the coming months: – We are investing in natively AI-based tools so that everyone at Meta can do more. We strengthen the position of independent employees and flatten the structure of teams.
He also added: – If we accomplish this, I believe that not only will we increase our efficiency, but work will give us much more fun.
“Tokenmaxxing” in Silicon Valley. This is an increasingly common element of company cultures
Whether Meta employees also perceive answering questions asked by the CEO's AI as “fun” is unknown, but Zuckerberg's active use of AI agents is not an isolated case in the company. Many big tech employees are currently racing to use artificial intelligence models as often as possible and to use as many so-called tokens, related to the use of computing power needed for AI work. A phenomenon that in Silicon Valley is currently called tokenmaxxinghas recently become a new measure of performance in some departments. I.e. the more tokens you use, the more value you likely bring to the company.
Recently, he has spoken publicly on this topic, among others: Jensen Huang, the head of Nvidia, clearly indicating that he would be very concerned if someone in a higher position did not use enough tokens. Using a simple example, he said: – At the end of the year I will ask him: “how much did you spend on tokens?” And if this person answers 5 thousand. dollars, then I'll go away. If an engineer earning $0.5 million has not used up at least 250k worth of tokens, I will be deeply concerned.
According to various reports, this culture has also become established at Meta and as a result, at least several intriguing projects have been created, actively used within the company, which currently has nearly 78,000 employees. employees. Internal tools whose existence has recently leaked to the media include: Second Brain, searching and organizing company documentation for all projects or My Claw, used for communication between AI agents of individual employees. Given Meta's recent acquisition of the Moltbook project, which is a social media for bots, the latter in particular should come as no surprise.
We will certainly hear more and more about AI agents in the coming months. More companies are introducing them into their structures, in some places their numbers already (or soon will) exceed the number of human employees, and Mark Zuckerberg himself directly admits that in his vision – each employee is supported by at least one AI agent. His view seems to be shared by all CEOs of key big tech companies, and the belief that increased effectiveness and efficiency in this way will translate into better financial results makes this a path from which it is impossible to turn back in the near future.
See also: I left my high-ranking job at the age of 36 and went on my own. AI agents handle everything from invoices to offers
