Strategic vision.
Proven execution.

It's been quite a week for those following the WBD-Netflix-Paramount cage match. Trips to DC and back, escalating bids, stripped contingencies, and no shortage of "assurances" delivered to every constituency with a stake in the outcome. Bloomberg consistently does some of the best work covering the business of entertainment, and this piece captures the frenetic behind-the-scenes maneuvering that ultimately led Paramount — and David Ellison — to outlast the competition and win the prize.

Is the RSN tumult finally over after seven long years? For Diamond Sports Group — sorry, Main Street Sports Group — it's ending in a whimper, with MLB rights reverting after this fall's season and NBA and NHL rights following when their seasons wrap next spring. The real question is what happens next for affected teams' D2C aspirations — something I explored in Cutting the Cord? back in 2022. Simply put, the teams don't have the expertise to capitalize on those audiences themselves — but the leagues absolutely do. MLB Commissioner Rob Manfred has been clear for years that he wants to control a national-local rights package, and NBA Commissioner Adam Silver has quietly built the business and technical infrastructure to do the same. That leaves the NHL to decide which path they want to take — and the huge success of hockey at the Milano-Cortina Olympics just might be the boost they need.

If you're not already watching 3Blue1Brown, fix that. Grant Sanderson has a rare gift for making deep mathematics not just accessible but genuinely beautiful — and this video on the Hairy Ball Theorem is a perfect example. Beyond the elegance of the proof itself, topology like this has real applications in signal processing and error coding, particularly in how we think about convolutional codes and signal mapping on continuous surfaces. But mostly, I just think anyone working in technically complex fields should have this channel in their rotation. You'll come away smarter every time.

Another gem from the 3Blue1Brown universe, this time in collaboration with Welch Labs. What makes this video worth your time is how clearly it explains the structural parallels between how AI handles language and how it handles images. LLMs encode language into high-dimensional vector spaces where similar concepts cluster together and the complex relationships between them can be represented mathematically. Image generation works on a strikingly similar principle — training images are progressively noised into random states along high-dimensional paths, and generation learns to reverse that process from a random seed back to a coherent image. In both cases, proximity in the space reflects similarity, and the mathematical structure of the space itself captures relationships that would be impossibly complex to define by hand. Understanding these shared foundations is increasingly important for anyone in media and entertainment thinking about where automated content creation, real-time highlight generation, and AI-assisted production workflows are heading.