Sulaiman Ghori, a former engineer at Elon Musk’s xAI, departed the company just four days after giving an extensive interview on the “Relentless” podcast, where he disclosed numerous internal details about the AI startup’s operations. While neither xAI nor Musk has commented on whether the departure was related to the interview, prominent figures like MrBeast have speculated about the connection.
During the wide-ranging podcast interview, Ghori revealed several eyebrow-raising details about xAI’s operations. Perhaps most surprisingly, he disclosed that xAI’s data centers are built on temporary “carnival” leases — special permits typically granted for temporary events like carnivals, which allowed the company to expedite permitting and construction. “It was the fastest way to get the permitting through and actually start building things,” Ghori explained.
Ghori also provided insights into xAI’s use of AI agents as virtual employees. He described how one team rebuilding core production APIs consists of just one human and 20 AI agents. The company’s reliance on AI workers has caused organizational confusion, with Ghori receiving messages asking about employees who turned out to be AI agents. Even without AI employees, xAI keeps teams remarkably small — the iOS team had only three people during the Grok Imagine launch.
The former engineer revealed that each code commit to xAI’s repository is valued at approximately $2.5 million, claiming he completed five in a single day worth $12.5 million. He also discussed Elon Musk’s hands-on role as a problem-solver, particularly when new hardware from companies like Nvidia doesn’t work as expected. Musk personally gets on the phone to resolve issues that would otherwise take weeks.
Ghori described xAI’s intense work culture, including sleeping pods, bunk beds, and tents for employees working overnight. He shared an anecdote about Musk offering a Cybertruck as an incentive to an engineer who could get a training run on new GPUs within 24 hours — a bet the engineer won.
The interview also covered xAI’s ambitious “human emulator” project (developed by the “Macrohard” team), which aims to create AI agents that perform digital human actions. Ghori suggested that dormant Tesla vehicles could potentially power up to one million of these emulators, leveraging the fact that Tesla cars sit idle 70-80% of the time. Finally, he revealed that xAI plans models far in advance, with Grok-5 already designed before he joined in March 2025, while consumers currently use Grok 4.
Key Quotes
It was the fastest way to get the permitting through and actually start building things. I assume that it will be permanent at some point.
Ghori explained how xAI uses temporary carnival permits to build data centers quickly, revealing an unconventional regulatory workaround that allows the company to bypass traditional permitting processes and accelerate infrastructure development.
Multiple times I’ve gotten a ping saying: ‘Hey, this guy on the org chart reports to you. Is he not in today or something?’ It’s an AI. It’s a virtual employee.
This quote illustrates how xAI has already integrated AI agents so deeply into its workforce that they appear on organizational charts alongside human employees, causing confusion and demonstrating the company’s aggressive adoption of AI-as-workers.
Elon’s like, ‘OK, you can get a Cybertruck tonight if you can get a training run on these GPUs in 24 hours.’
Ghori shared this anecdote about Musk’s unconventional incentive structure, highlighting the CEO’s hands-on management style and the intense, deadline-driven culture at xAI where extraordinary rewards motivate rapid technical achievements.
There are 4 million Tesla cars in North America alone. They’re sitting idle for 70-80% of the time. Why not pay owners to lease time off their cars and run the emulator on them?
This quote reveals xAI’s ambitious plan to create a distributed AI infrastructure using idle Tesla vehicles, potentially transforming millions of cars into a massive computing network for running human emulator AI agents.
Our Take
Ghori’s sudden departure after this revealing interview sends a chilling message about transparency in the AI industry. While his insights are invaluable for understanding how cutting-edge AI companies operate, the apparent consequences highlight a troubling trend: the most innovative AI development happens behind closed doors, with employees facing potential retaliation for disclosure.
The technical revelations are equally significant. XAI’s integration of AI agents as virtual employees isn’t just experimental — it’s operational reality, suggesting the AI-replacing-workers narrative is already happening internally at AI companies themselves. The human emulator concept, combined with the Tesla vehicle infrastructure plan, represents a bold vision for ubiquitous AI agents that could fundamentally change how we interact with digital systems.
Most concerning is the regulatory arbitrage — using carnival permits for data centers exemplifies how AI companies are finding creative ways around traditional oversight. This aggressive approach may deliver speed, but it raises questions about whether existing regulatory frameworks can keep pace with AI development.
Why This Matters
This story provides rare insider insights into one of the most secretive AI companies in the industry, revealing xAI’s unconventional operational strategies and aggressive approach to AI development. The departure of an engineer shortly after disclosing internal information highlights the intense confidentiality culture at Musk’s companies and raises questions about transparency in the AI industry.
The revelations about xAI’s use of AI agents as virtual employees and plans to deploy human emulators at massive scale demonstrate how quickly AI companies are moving toward autonomous AI systems. The proposal to leverage idle Tesla vehicles as distributed computing infrastructure represents a novel approach to scaling AI capabilities that could reshape how computational resources are utilized.
For the broader AI industry, this story illustrates the extreme pace of development and unconventional methods companies are employing to gain competitive advantages. The use of temporary carnival permits for data centers and the valuation of individual code commits at millions of dollars underscore the massive financial stakes and regulatory creativity in the AI race. These insights matter for investors, policymakers, and competitors trying to understand how leading AI companies operate and the lengths they’ll go to accelerate development.
Related Stories
- Elon Musk’s XAI Secures $6 Billion in Funding for Artificial Intelligence Research
- OpenAI Lost Nearly Half of Its AI Safety Team, Ex-Researcher Says
- Google Workspace’s Robert Frati Left for AI Startup Gemini Amid Tensions Over AI Strategy
- OpenAI’s Competition Sparks Investor Anxiety Over Talent Retention at Microsoft, Meta, and Google
- What Zuckerberg Risks by Following Musk’s Lead