Rethinking Media as a Structured, Composable System
Most media is still treated as an opaque file. But media is composed of atoms: frames, stems, shots, scenes, dialogue, music, versions. Once those atoms are structured and addressable, AI gets dramatically more powerful. The atoms are the leverage. The assembly is the art.
Media Discovery in an Agentic World: Why Legacy Metadata Alone Won't Save You
Agents won't browse a carousel or scroll a feed. They'll evaluate options against an intent. For media companies, that raises a hard question: can an agent understand your catalog well enough to know when it's the right answer? Right now, for most of the industry, the answer is n
The promise of AI is range not replacement.
AI is not an artist in the human sense. But it is more than a remix machine. The real opportunity is using it to extend human imagination, not replace it.
Media Agent Transformation Demands Domain Depth
The unlock of agentic around media workflows will be domain-specific judgement to rebuild creation, ingest, QC, localization, metadata, and delivery workflows.
Perfection is not the right benchmark for AI
From medicine to autonomous vehicles to media operations, AI is too often judged against an imaginary standard of human perfection. The real test is whether agentic systems, with appropriate oversight, reduce consequential errors versus today’s manual workflows operating under fa
Stanford proved orchestration matters more than the model
Stanford’s Meta-Harness study found a 6x performance gap using the exact same model weights, with just different scaffolding. The biggest gains in AI may no longer come from the model itself, but from the harness wrapped around it.
Simon Willison on agentic engineering, dark factories, and why testing is the new constraint
Software crossed a threshold in late 2025: models stopped being interesting demos and became reliable collaborators. The bottleneck is no longer generation. It’s verification. And the patterns reshaping software are already arriving in media.
Systems of Record Survive, Interface Transforms
The $400B SaaS market isn’t disappearing. What’s dying is the assumption that humans will always be the primary interface. In the agentic era, the winners won’t own the UI. They’ll become the orchestration substrate everything else runs on.
Agents are the new users. Your API isn't ready.
SaaS was optimized for humans who click, interpret, and forgive ambiguity. But when the user is an agent, one that might call your API a thousand times before your team clocks in, the design assumptions shift. What changes when your primary user migrates from humans to agents?
What Happens When AI Becomes a Primary User
We spent 30 years perfecting how humans interact with software. Now we need to design for a different user, one that doesn't need screens and that processes information differently.
Moving from Science Projects to Scalable Platforms
95% of AI projects miss ROI, but that’s not a failure of AI. It’s a hallmark of early stage technology adoption. Learn why early efforts stumble, where AI is working, and how agentic architectures will define the next decade.
Why We're Still in the 'Command Line Era' of AI and What Comes Next
Karpathy framed it well: AI chat is the mainframe era, pre-GUI. LLMs are OSes missing essentials like real UIs, user profiles, app stores, deterministic agents, payments, comms. We're stuck in text, yet ready to book rides, restaurants, and rentals without leaving the assistant.
AI is the New UI
We're witnessing the decline of the App Economy and the birth of something new. When most people expect to order Uber rides and book restaurant reservations through AI assistants rather than dedicated apps within five years, we're looking at a fundamental replatforming moment.
Why Media Companies Must Create AI-Native Metadata Now to Win the AI Discovery Race
As SEO became essential for web visibility, AI Engine Optimization (AEO) is emerging as the next frontier for discoverability. The race to optimize content libraries for AI discovery has begun. Those who establish early expertise will gain advantage in reach and monetization.
How Conversational AI is Replacing Apps, Search, and Traditional Content Discovery
We're witnessing the gradual replacement of traditional apps and search with AI. For Hollywood, this means preparing content for "AI Engine Optimization", a new frontier where discoverability depends on APIs and schemas rather than standalone UIs and SEO tactics.
Consolidating Viacom, CBS, Showtime, and Paramount Supply Chains Into One Consolidated Cloud-First Platform
In 2021, we faced a real challenge: consolidate separate supply chains built for Paramount's distinct brands into one unified system. While launching Paramount+ across 20+ territories, we deployed a unified platform that made complex operations accessible to ops teams worldwide.
How Apple's Supply Chain Genius Offers a Blueprint for Hollywood's AI Transformation
Just as Cook elevated the "unsexy" supply chain into extraordinary business leverage, Hollywood should view content supply chain excellence as a strategic differentiator. In the AI era, it's the key to unlocking massive new monetization opportunities and competitive advantage.
From Chaos to AI Gold: Why Legacy Media Must Unify Their Content Libraries
Legacy media has massive libraries of content fragmented across multiple repositories worldwide. Consolidation is essential for studios seeking to unlock AI's full potential and transform their content libraries into "AI gold."
AI is the Catalyst for Media’s Next Decade of Growth
AI is a transformative platform similar to the PC, internet, and mobile. While most executives are currently reimagining existing processes with AI, innovators are breaking free from conventional thinking to create entirely new paradigms.
Pragmatic writing on agent-experience design, AI-native media workflows, and the craft of building this stack. One-click unsubscribe in every email.