xAI Rebuilds, Meta Plans 20% Layoffs & AI Dog Vaccine

xAI Loses Co-Founders as Musk Rebuilds 🏗️

Elon Musk is tearing down and rebuilding xAI. Following a wave of employee exits, the billionaire stated that his AI company “was not built right” and is undergoing a foundational restructuring to help its Grok model catch up to the frontier of the industry.

Here are the key details of the xAI shakeup:

• The Departures: Zihang Dai and Guodong Zhang (who led Grok Code) have left the company. This leaves only two of the original 11 co-founders—Manuel Kroiss and Ross Nordeen—remaining with Musk.
• The Coding Deficit: Musk reportedly blamed Zhang for Grok’s coding shortfalls before his departure, publicly admitting that Grok is “currently behind” on coding capabilities.
• The Reinforcements: To bridge the talent gap, xAI has actively begun hiring major coding talent, recently onboarding senior Cursor leaders Andrew Milich and Jason Ginsberg.

Why it matters: Building a frontier model requires deep institutional knowledge and a highly cohesive team. Losing 9 of 11 co-founders in roughly a year is a massive blow to xAI’s momentum. While Musk has the capital to hire top-tier replacements, rebuilding the foundation while simultaneously trying to prepare for an IPO and catch up to OpenAI is an incredibly steep climb.

UrviumAI Take: Money cannot instantly buy a cohesive engineering culture. If you are building an AI product, do not rely on Grok for complex coding tasks just yet. Stick to established coding models like Claude 3.5 Sonnet or OpenAI’s latest models while xAI spends the next few months rebuilding its engineering team and rewriting Grok’s coding capabilities.

Meta Plans 20% Layoffs to Fund $600B AI Push 📉

The transition to an AI-first economy is beginning to show its human cost. Meta is reportedly planning sweeping corporate layoffs that could affect up to 20% of its nearly 79,000-person workforce as the company seeks to aggressively offset its artificial intelligence infrastructure bets.

Here is the financial strategy behind the proposed cuts:

• Massive Spending: Meta plans to invest a staggering $600 billion to build AI data centers by 2028. They are also spending heavily on acquisitions, including $2 billion for Chinese AI startup Manus and social platform Moltbook.
• AI-Assisted Efficiency: CEO Mark Zuckerberg is banking on AI to replace human headcount, noting that projects that previously required big teams can now be accomplished by single, AI-assisted employees.
• Model Struggles: This financial pivot comes amid reported setbacks with Meta’s internal models, including the delayed Llama 4 “Behemoth” and underperforming “Avocado” superintelligence models.

Why it matters: This marks a watershed moment in corporate tech. Meta is explicitly tying mass layoffs to the efficiencies gained by AI tools while simultaneously pouring hundreds of billions into AI hardware. It is the clearest signal yet that major tech giants view generative AI not just as a new product line, but as a direct replacement for human operational overhead.

UrviumAI Take: The “Year of Efficiency” is becoming the permanent reality of the AI era. If you work in tech or digital operations, you must master AI agents immediately. Companies are no longer hiring teams of ten; they are looking for one “10x employee” who can manage a swarm of AI agents to do the work of a full department.

AI-Designed Vaccine Shrinks Dog’s Tumor 🐕

Artificial intelligence is democratizing life-saving biotechnology. Sydney AI consultant Paul Conyngham successfully designed a custom mRNA cancer vaccine for his rescue dog, Rosie, by leveraging a suite of publicly available AI tools to process complex genomic data.

Here is how Conyngham built the treatment:

• The Diagnosis: Rosie was diagnosed with mast cell cancer and given months to live, despite traditional chemotherapy and surgery.
• The AI Pipeline: Conyngham paid $3,000 for genomic sequencing, used ChatGPT to map the biological research, and fed 350 GB of tumor data through DeepMind’s AlphaFold to model Rosie’s specific protein mutations.
• The Final Construct: Working with the UNSW RNA Institute to physically synthesize the formula, Conyngham revealed that the final vaccine blueprint was actually designed by xAI’s Grok model.
• The Results: Following her December injection, one of Rosie’s tumors shrank by half, and Conyngham is currently designing a second vaccine for remaining tumors.

Why it matters: A year ago, it was impossible for a pet owner with zero formal biology training to turn a cancer diagnosis into sequenced DNA data, model the proteins, and output a working vaccine blueprint. This story proves that AI models like AlphaFold and Grok are no longer just academic curiosities; they are practical tools capable of generating real-world, biological interventions.

UrviumAI Take: Personalized medicine is about to become a software problem. Stop viewing LLMs merely as text generators. By chaining models together (like using ChatGPT for research, AlphaFold for physics, and Grok for sequencing), everyday people can now execute complex scientific workflows that used to require entire pharmaceutical labs.

Last AI News: Nvidia’s Nemotron 3 Super, Perplexity’s Mac Mini Agent & Microsoft Backs Anthropic

Other AI News Today:

• Former Anthropic researchers are raising $175M at a $1B valuation for Miraendil, a new startup building specialized AI for scientific R&D.
• A humanoid robot was escorted away by police in Macau after it startled a 70-year-old woman, leading to her brief hospitalization and sparking online amusement.
• ByteDance has suspended the global launch of its highly anticipated Seedance 2.0 video generation model following major copyright disputes with Hollywood studios.
• Elon Musk announced that Tesla’s ambitious Terafab semiconductor project will launch in seven days, signaling a massive push to manufacture custom AI chips.
• A Florida homeowner successfully sold his Cooper City house in just five days using ChatGPT, saving roughly 3% in real estate agent fees.

Jigar Chaudhary

Jigar Chaudhary is the Editor-in-Chief at UrviumAI, where he oversees coverage of artificial intelligence news, tools, and in-depth studies. With over 5 years of experience analyzing AI and robotics, he focuses on maintaining high editorial standards, accurate reporting, and clear explanations to help readers understand how AI is shaping the future.