PLUS: Salesforce's AI regret, the Pentagon deploys Grok, and Microsoft's internal AI drama
Good morning
AI is taking a major step beyond generating assets and is now beginning to build entire interactive game worlds from scratch. New 'world models' from giants like Google DeepMind are using simple text prompts to generate complex 3D environments, signaling a major shift for the video game industry.
The technology promises to dramatically lower the barrier to game creation, potentially turning players into world-builders. But as the creation of vast digital worlds becomes more automated, what will this mean for the role of human artists and the quality of future games?
In today’s Next in AI:
DeepMind's new AI builds entire game worlds
Salesforce's AI support replacement backfires
The Pentagon deploys xAI's Grok chatbot
Microsoft's internal AI code rewrite drama
AI's Next Level: Gaming
Next in AI: New AI "world models" from tech giants like Google DeepMind and startups are set to reshape the $190B video game industry by letting creators generate entire interactive 3D worlds from simple text prompts.
Decoded:
Major players are entering the field, with Google DeepMind developing its Genie 3 model and AI pioneer Fei-Fei Li’s World Labs launching a model called Marble.
This technology builds on current AI use cases, where tools are already used to create game assets and interactive characters, with one studio reporting a fourfold increase in development speed.
The long-term vision is to empower individual gamers to build their own worlds while helping professional developers reduce production costs and burnout on massive "triple-A" titles that can cost over $1bn.
Whyt It Matters: This shift promises to dramatically lower the barrier to game creation, enabling developers to prototype faster and players to become active world-builders. However, it also sparks an industry-wide debate about the future roles of human artists and the potential for low-quality, AI-generated content.
Salesforce's AI Reality Check
Next in AI: Salesforce is reversing its decision to replace 4,000 customer support staff with AI after the move led to declining service quality and exposed the gap between AI hype and operational reality.
Decoded:
The company cut its support team from 9,000 to 5,000 employees, with executives now admitting they were too confident in AI's ability to handle complex customer issues without human oversight.
AI systems struggled with nuanced problems and escalations that required judgment calls, leading to longer resolution times and the loss of valuable institutional knowledge from experienced staff.
CEO Marc Benioff has shifted his messaging from full automation to human-AI collaboration, with the company now focusing on augmentation rather than replacement as its core strategy.
Whyt It Matters: This high-profile reversal demonstrates that AI still has significant limitations in customer-facing roles that demand empathy and complex problem-solving. Companies rushing to cut costs through AI-driven headcount reductions may face similar quality and credibility challenges.
Pentagon Deploys Grok
Next in AI: The U.S. Department of War announced it is integrating xAI's Grok chatbot into its systems. The partnership will allow military and civilian personnel to use the model for handling controlled unclassified information starting in early 2026.
Decoded:
The deployment will allow personnel to use Grok for Impact Level 5 (IL5) data, which is defined as controlled unclassified information.
The agreement also gives the Department of War access to X (formerly Twitter) to gain real-time global insights for a decisive information advantage.
This deal moves forward despite Senator Elizabeth Warren previously warning the Pentagon that Grok generated misinformation and offensive content.
Whyt It Matters: This partnership signals a major push by government agencies to adopt commercial AI tools, even those with a history of public controversy. It sets a significant precedent for how AI models are vetted and deployed in national security contexts.
Microsoft's AI Rewrite
Next in AI: A top Microsoft engineer's controversial post claiming AI would rewrite Windows sparked major backlash online. The company quickly clarified the statement refers to an internal research project, not a plan to overhaul the operating system's core code.
Decoded:
The engineer's original goal was to eliminate every line of C and C++ from Microsoft by 2030 with a north star of 1 engineer, 1 month, 1 million lines of code.
This aligns with Microsoft's broader vision, as CEO Satya Nadella has stated that AI already writes up to 30% of the company's code in some projects.
Following the public reaction, the engineer edited his post to clarify the project's research focus and explicitly state that Windows is NOT being rewritten in Rust with AI.
Whyt It Matters: This incident reveals the fine line between ambitious internal AI goals and the public perception of engineering stability for critical products. It confirms that major tech firms are aggressively pursuing AI-driven code generation, even if rewriting legacy systems from scratch is not yet feasible.
AI Pulse
NVIDIA licensed inference technology from AI chip startup Groq in a deal reportedly valued at $20B, acquiring the company's assets and key leadership to enhance its low-latency processing capabilities.
xAI drew criticism for its Grok chatbot, as analysis suggests the model's perceived ideological bias often reflects the personal viewpoints of founder Elon Musk.