In this weeks news Ralph and Emile cover updates from ascross the AI industry and share insights into new tools, concepts and techniques.

Key Topics at a Glance

• OpenAI “code red” vs Anthropic’s calmer strategy
• DeepSeek’s open-source frontier model
• AI-driven revenue at Salesforce & Microsoft
• Meta’s WhatsApp walled garden vs EU regulators
• Data-center debt bubble risk around OpenAI
• Suno’s acquisition of Songkick (AI music → live shows)
• Replit’s AI-caused prod deletion & agent safety
• Apple’s AI leadership shuffle
• Nvidia’s China-targeted servers and export politics
• Anthropic’s DeFi exploit detection agent
• Amazon Rufus & AI-boosted Black Friday conversions
• GPUs vs TPUs & green AI scaling
• Prompting pattern: “5 responses + probabilities”
• Affinity going free forever as an Adobe alternative
• Mistral’s edge models, HSBC deal & EU sovereignty
• Robo-judges, hallucinations & legal risk
• Tokenomics: model pricing, token thresholds, and cost optimisation
• DeepSeek & small language models as ultra-cheap options
• Gemini glasses doing real-world tasks (like oil changes)
• Persona-driven voice agents & Trump-style recipe voice
• Aura + Gemini for stunning AI-built websites
• OpenAI age verification & adult mode vs over-guardrails