Musk Begins Replacing Staff with Grok as Massive Layoffs Hit X

Elon Musk is pushing X into a future run mostly by AI, and the first signs are now impossible to miss. A large part of the trust-and-safety engineering team has been replaced by Grok. Grok is an AI model from xAI. This department, which once had more than a hundred people, is down to just a few.

Reports from The Information say Musk let go of roughly half of the engineers. These were staff still working on things like spam, fake profiles, harmful material, and organized manipulation. And since there were only about 20 people left to begin with, losing that many makes the impact feel even bigger. When Musk bought Twitter in 2022, there were more than six times that number.

This fits his earlier plan from last year, where he said X’s feed would switch to an AI-powered recommendation system built around Grok instead of the old algorithm. If the AI can decide what users see and how posts are ranked, Musk seems to feel he doesn’t need many humans doing the same work.

The idea stretches beyond X, too. Musk is working on something called Macrohard, a sort of Microsoft rival built around AI from the start. The message is clear: automation is replacing people, whether the world is ready or not.

Running much of this shift are two engineers from xAI, twin brothers Dima and Ievgin Soboliev. They’re 33, originally from Ukraine, and now sit close to the top of the organisation.

After several senior AI engineers left for OpenAI, the twins stepped forward and became deeply involved in building and managing Grok. They’ve worked at places like Meta, Google, Apple, and various AI startups. Inside the company, people just call them the twins, because they almost move and think as one.

Workers say they demand long hours and fast results. Many roles they consider unnecessary are cut without hesitation. Around one hundred engineers are still spread across different teams, but it’s unclear how long that number will hold.

Musk wants X to run mostly on Grok. Ideally, users could talk directly to the AI to customise what shows up in their feed. If that happens, Grok becomes more than a chatbot; it becomes the system that runs the platform itself.

But the plan has cracks. The smaller safety team already struggles with harmful posts, and Grok is being trained to be more creative, which makes its behaviour harder to predict. No one seems fully in charge of balancing freedom and safety.

The layoffs have also slowed down projects like X Money, Musk’s plan to bring banking and payments into the app. Regulators still haven’t approved it, and the shrinking workforce has made progress even slower.

AI might make things cheaper and faster, but it also raises the question: who’s watching the machine?

Earlier this year, xAI also laid off over 500 labellers who helped train Grok. After that, photos showed mostly East Asian employees inside the company. One engineer even claimed that he was the only white employee left before being laid off, although that hasn’t been proven.

Musk is building something bold, maybe even historic. But cutting people this quickly could turn into a gamble that either pushes X forward or sends it off the rails.