The AI Dispatch

AI Industry

Musk Admits xAI “Was Not Built Right” — Orders Full Rebuild as Co-Founders Flee

Only 2 of xAI’s original 11 co-founders remain. Musk acknowledged his AI lab failed to compete with Claude Code and OpenAI Codex, and is restructuring into four divisions while raiding talent from Mistral and Cursor.

Sources: TechCrunch • CNBC

Elon Musk conceded publicly on Thursday what industry observers have suspected for months: xAI, the artificial intelligence company he founded in 2023 with a roster of eleven elite researchers poached from DeepMind, Google, and Microsoft, has failed to produce competitive products and must be rebuilt from scratch. In a post on X, Musk wrote that xAI “was not built right first time around, so is being rebuilt from the foundations up.” The admission came hours after CNBC reported that two more co-founders — Zihang Dai, a former Google Brain researcher, and Guodong Zhang, who had led xAI’s optimization team — departed in recent weeks, leaving only two of the original eleven still at the company.

The restructuring splits xAI into four divisions: a foundational research lab, a consumer products group centered on Grok, an enterprise AI services unit, and a new coding tools division that Musk described as “the most important.” It is the coding division that underscores the nature of xAI’s failure. While Anthropic’s Claude Code and OpenAI’s Codex have become standard tools for professional software development, xAI never shipped a competitive coding product. Musk acknowledged as much, telling employees in an internal memo obtained by TechCrunch that “we lost the coding tools race before we entered it.”

To staff the rebuild, xAI has embarked on an aggressive hiring campaign that is rattling competitors. The company has hired Devendra Singh Chaplot, a co-founder of Mistral, to lead the new research lab, and has recruited at least a dozen engineers from Cursor, the AI-powered code editor. The talent raids extend to Anthropic and Google DeepMind as well, though both companies have reportedly moved to retain key staff with counter-offers. Whether Musk can attract and retain world-class researchers — given xAI’s track record of co-founder departures and his own reputation for mercurial management — remains the central question of the rebuild.

Labor & AI

Meta Weighs Laying Off 20% of Workforce as AI Infrastructure Costs Mount

The company that spent $38 billion on AI last year may cut up to 16,000 jobs — a stark illustration of the industry’s invest-in-AI-while-replacing-humans-with-AI paradox.

Sources: TechCrunch • CNBC

Meta is considering layoffs that could affect as many as 16,000 employees — roughly 20% of its global workforce — according to reports from Reuters and CNBC published Friday. The cuts would represent the company’s largest headcount reduction since the 2022–2023 “year of efficiency” that eliminated 21,000 roles. The timing is revealing: Meta spent approximately $38 billion on AI infrastructure in 2025, a figure CEO Mark Zuckerberg has signaled will grow substantially this year, even as the company concludes it needs fewer humans to operate the systems that spending is building.

The paradox is now impossible to ignore. Meta is simultaneously the largest corporate investor in artificial intelligence and one of the most aggressive practitioners of AI-driven workforce reduction. The roles under consideration span content moderation, human review, and mid-level management — precisely the functions that internal AI systems have been designed to automate. For the broader technology industry, Meta’s deliberations crystallize a dynamic that has been building for two years: the companies building AI are the first to replace their own workers with it, creating a feedback loop in which AI investment justifies AI-driven layoffs, which in turn frees capital for more AI investment.