DOWN TO BUSINESS

🛍️ Retailers race toward AI, but data gaps hold them back

Fragmented data is the biggest barrier retailers face in scaling AI, according to a new Amperity survey of 1,000 retail professionals. The survey highlights that 45% of retailers use AI weekly, and 97% plan to maintain or increase investment in 2025, but only 11% are ready to scale it. Retailers with a customer data cloud (CDP) are far ahead, using AI more frequently, integrating it across business units, and showing greater confidence in acting on customer insights.

🏢 Cohere launches North for scalable enterprise AI

Cohere has launched an agentic AI platform built for security-conscious enterprises to deploy AI agents and workflow automations directly within their own infrastructure. Already adopted by RBC, Dell, LG CNS, Bell, and others, North integrates generative and search models with customizable agents that connect securely to fragmented enterprise data sources like Gmail, Salesforce, and SharePoint.

🚨 Carbyne secures $100M to expand AI-powered emergency response

Carbyne, a leader in cloud-native emergency communications, has raised $100 million to accelerate its AI-driven 911 platform and expand globally. The company’s flagship APEX platform integrates AI directly into call handling, dispatch, and resource coordination. This built-in intelligence helps reduce response times, ease staffing pressures, and deliver real-time support in mission-critical situations.

🦅 Anthropic offers Claude AI to U.S. Government for $1

Anthropic is making its Claude AI model available to U.S. federal agencies for just $1, following similar moves by OpenAI and Google. The offer comes shortly after Claude, ChatGPT, and Gemini were added to the government’s approved AI vendor list. CEO Dario Amodei said the move ensures U.S. institutions have access to “the most capable, secure AI tools available.”

HOT MODELS

🤖 OpenAI launches GPT-5 for expert-level AI in everyone’s hands

OpenAI has released GPT-5, its most advanced AI model yet, delivering major gains in reasoning, accuracy, and versatility. The unified system uses a smart router to decide when to respond instantly or think longer. Other upgrades include stronger coding for complex front-end builds, more nuanced and stylistically aware writing support, and significantly higher scores on health benchmarks.

🎵 ElevenLabs launches studio-grade AI music generator

ElevenLabs has unveiled Eleven Music, a tool that creates studio-quality songs from natural language prompts in any genre, style, or language. Users can generate instrumental or vocal tracks, edit sections or full songs, and craft everything from cinematic scores to retro ballads. Built with input from labels, publishers, and artists, Eleven Music is cleared for most commercial uses, including film, TV, ads, games, and social media.

📝 NotebookLM adds video overviews and upgraded studio

Google’s NotebookLM now features video overviews, turning complex topics into narrated slide presentations with visuals, diagrams, and data pulled from your documents. Users can tailor content to their expertise level, learning goals, and audience. The upgraded Studio panel also lets you create and store multiple outputs: audio overviews, video overviews, mind maps, and reports, all within a single notebook.

Plot Twist

💻 Study finds AI can slow down experienced developers

A METR study found that seasoned developers working in familiar codebases completed tasks 19% slower when using an AI coding assistant despite expecting a 24% speed boost. The slowdown came from time spent correcting AI-generated suggestions, which were often “directionally correct” but not precise enough. While previous research shows AI can speed up coding, this study suggests the gains don’t always apply to complex, established projects.

🤖 Figure brings AI-powered humanoids into homes

Figure, the first company building a general-purpose humanoid robot, is moving from the factory floor to the living room. Designed in a human form to navigate real-world environments, Figure 02 can already perform household chores like laundry, dishes, and unpacking groceries, 100% autonomously. Powered by its in-house AI system, Helix, the robot combines dexterity, mobility, and contextual reasoning to adapt to dynamic tasks.

♟️ OpenAI’s o3 model checkmates Grok in AI chess showdown

OpenAI’s o3 model has won an AI chess showdown, defeating Elon Musk’s Grok 4 in the final of a Kaggle-hosted tournament. The event featured eight large language models from top AI developers, none built specifically for chess, competing over three days. Grok, dominant until the semifinals, faltered in the final with repeated blunders, including losing its queen, while o3 stayed flawless. Google’s Gemini placed third, edging out another OpenAI model.

♻️ Subscribe on LinkedIn and share it with your network!