FT: “World models are trained using data streams of real or simulated environments. They are viewed as an important step in pushing forward progress in self-driving cars, robotics and so-called AI agents, but require a huge amount of data and computing power to train and are considered an unsolved technical challenge. This focus on an alternative approach to LLMs has become visible as several AI groups have unveiled a series of advancements in world models in recent months…One near-term application of world models is in the entertainment industry, where they can create interactive and realistic scenes. World Labs, a start-up founded by AI pioneer Fei-Fei Li, is developing a model that generates video game-like 3D environments from a single image. Runway, a video generation start-up that has deals with Hollywood studios, including Lionsgate, launched a product last month that uses world models to create gaming settings, with personalised stories and characters generated in real time.”
Simon Willison: “LLMs took the ability to predict the next token and turned it into general purpose foundation models for all manner of tasks that used to be handled by dedicated models – summarization, translation, parts of speech tagging etc can now all be handled by single huge models, which are getting both more powerful and cheaper as time progresses. Generative video models like Veo 3 may well serve the same role for vision and image reasoning tasks.”
The Globe and Mail: “Mr. [Mark] Leonard built Constellation into a technology juggernaut with a disciplined approach, buying well over 1,000 software companies. Target companies were typically well run and stable, often led by their founders, having loyal customers, and facing little to no competition catering to specific industry niches. The companies typically generated less than US$10-million in annual revenue. Rolled up together, they generated US$5.5-billion in revenue for Constellation in the first half of 2025 and US$258-million in net income. “What is special about Constellation is not the underlying technology or the businesses, but the business model,” said John Ruffolo, a Bay Street financier who has known Mr. Leonard since the 1990s. “He saw value where no one could. These were not the hot companies, they were in many cases those that people just forgot about or where they didn’t appreciate the value.””
BCG [newsletter]: ““Future-built companies,” representing 5% of respondents, are achieving AI value at scale. Compared with other companies, those that are built for the future achieve five times the revenue increases and three times the cost reductions from AI. They report three-year total shareholder returns that are 3.6 times greater. These results and returns don’t just happen. These companies plan to spend more than twice as much on AI compared with laggards—60% of companies—in 2025. Future-built companies also have a plan. They have moved beyond pilots and automation projects to reshaping entire functions—R&D, sales, marketing, manufacturing. Our research found that 70% of potential value from AI is concentrated in the core business, not support functions. For future-built companies, agents represent the next frontier of advantage, accounting for 17% of AI’s total value creation this year—a share expected to nearly double by 2028. Agents can autonomously reason, break down problems, find ways to solve them, and learn from their mistakes.”
