AI Daily Briefing — 2026-06-27

Artificial intelligence is transforming retail in subtle but profound ways, moving beyond customer-facing gimmicks to overhaul core operations. According to MIT Technology Review AI, the real shift is happening behind the scenes—how products appear in search, how supply chains adjust, and how retailers respond instantly to shopper behavior. Macy’s senior director of engineering Murali Murugan describes this as an “AI-first” philosophy, where intelligence is embedded directly into systems rather than bolted on. “It’s about redesigning how decisions happen so the business moves faster,” he says.

Macy’s early wins came from focused applications like personalized search and customer engagement, which quickly improved conversion rates and built internal support. Once those quick wins were established, scaling became a business decision rather than a technology debate. The company has since expanded into conversational commerce with Ask Macy’s, an AI shopping assistant that acts like a personal stylist, offering curated recommendations based on past purchases and context for events like proms or vacations.

Despite these advances, Macy’s views AI as an invisible layer that augments human judgment, not replaces it. The long-term goal is retail that feels seamless, adaptive, and personalized—powered by systems customers may never notice. “The real transformation comes from continuous improvement,” Murugan notes, “learning from mistakes and adapting quickly to new technology standards.” This approach, as reported by MIT Technology Review AI, reflects a broader industry move from isolated AI pilots toward integrated systems that shrink the gap between signal and action.

Patronus AI, a startup founded by former Meta AI researchers, has raised $50 million in Series B funding to expand its platform that stress-tests AI agents in simulated digital environments. The round, led by Greenfield Partners with participation from Notable Capital, Lightspeed, Datadog, and Samsung, brings the company’s total funding to $70 million. The San Francisco-based firm has seen revenue grow 15-fold over the past year, driven by demand from major AI labs and startups.

Patronus builds what it calls “digital world models”—replicas of websites and internal systems—where AI agents are tested after training. Using reinforcement learning, the platform rewards successful task completion and penalizes errors, helping to ensure agents perform reliably in complex, real-world scenarios. The approach is similar to how Waymo trained autonomous cars in synthetic worlds to handle rare hazards, but with a focus on catching AI agents that take shortcuts or fail to complete tasks correctly.

Currently, Patronus focuses on software engineering and finance, but co-founder Anand Kannappan says the company plans to expand into areas that are harder to verify. “We want to create environments where agents can run for days or weeks,” he said. The startup’s main competition comes from internal evaluation teams at AI labs, but Patronus differentiates by testing agents without human involvement.

Inside General Intuition’s New York R&D floor, a monitor showed what looked like a person playing Fortnite. But the player was an AI agent that had been running for 100 hours straight. The same intelligence controlling the game also powered a four-legged robot roaming the office. Using only a single camera, the bot explored the space, bumping into chairs like a learning toddler. The company says it took just eight minutes of real-world data to fine-tune the model for the robot, and that data was collected on a street, not in the lab.

General Intuition’s core idea is that video game footage can train AI to understand space and time. The startup spun out of Medal, a game-clip sharing platform, giving it access to hundreds of millions of hours of gameplay. Unlike rivals that try to infer actions from video alone, General Intuition uses the actual button presses recorded in those clips. This labeled data, the company argues, is the key to building an agent that can generalize from virtual worlds to physical ones.

The company recently raised $320 million at a $2.3 billion valuation, bringing total funding to $454 million. The model, which powers both game agents and robots, learns from a simulated environment the company calls “the gym.” In tests, the AI understands that walls are solid, ladders are climbable, and shadows shift with the sun. General Intuition plans to sell the agentic model itself, betting that game-trained AI can navigate the real world.

A brutal heatwave scorching Western Europe is doing more than wilting crops and straining power grids—it is also clouding human cognition. London hit a record 36.1°C this week, but felt even hotter. Researchers have long linked rising temperatures to increased irritability and violence, and new studies show firefighters struggle to concentrate after heat exposure. Children and those with mental health conditions face the greatest risks, while lab experiments suggest excessive heat disrupts brain chemical signals. Scientists are racing to understand these mechanisms, as reported by MIT Technology Review AI.

Meanwhile, the Trump administration has imposed unprecedented restrictions on OpenAI, demanding the company vet the first users of its forthcoming GPT-5.6 model before a wider release. Each initial partner must receive government approval, marking the first time a U.S. firm has been ordered to limit an AI model’s launch. The move follows ongoing tensions between Washington and AI developers, with Anthropic also locked in disputes with federal officials.

In other tech news, Apple and Xbox have raised prices on MacBooks, iPads, and consoles by over 20%, blaming surging costs driven by AI chip demand. The shortages, dubbed “RAMaggedon,” have sent Apple shares tumbling. Separately, the U.S. has banned Polestar from selling its electric vehicles due to the company’s majority ownership by China’s Geely, citing rules against connected-vehicle tech linked to foreign adversaries.

Europe is grappling with a record-breaking heatwave that is pushing its power grids to the brink. As residents rely heavily on fans and air conditioning to stay cool, demand for electricity has surged, forcing some power plants offline. According to MIT Technology Review AI, the crisis is compounded by shifting seasonal patterns—Europe's grid historically peaks in winter due to electric heating, but summer outages are now more common. Climate change is expected to worsen these challenges, requiring utilities to rapidly boost supply and adapt to new demand patterns.

In a separate development, IBM has announced a prototype chip containing roughly 100 billion transistors on a fingernail-sized area—double the density of its 2021 technology. This breakthrough could extend Moore's Law for another decade by using a vertical design approach, similar to building upward in crowded cities. As traditional transistor shrinkage nears its physical limits, this innovation promises faster, more energy-efficient computers, offering new hope for the tech industry.

These stories highlight the intersecting pressures of climate change and technological advancement. While Europe scrambles to keep the lights on during extreme weather, IBM's chip design points to a future where computing power continues to grow despite physical constraints. Both developments underscore the urgent need for resilient infrastructure and innovative engineering in a rapidly changing world.