Productivity Enterprise Film
Celebrity NDAs, AI Film Shots, and the Question the Bull Case Must Answer
Taylor Swift and Travis Kelce married at Madison Square Garden, generating enough cultural coverage to qualify as its own media event — and then generating a secondary story when MSG fired several employees for allegedly violating nondisclosure agreements about the wedding. The NDA enforcement action reveals something about how major celebrity events are now managed: with information-security protocols that resemble classified government operations more than traditional entertainment logistics.
The film 'Young Washington' opened this weekend with AI used in 100 shots — a number that will reframe the entertainment labor debate in the coming week. A typical feature film contains 300 to 600 total visual effects shots, meaning AI-generated work could represent 20 to 30 percent of the visual effects budget on this production. That is commercial-scale deployment, not an experiment. Separately, Peter Thiel accused the Pope of 'working for Chinese Communists' on AI — a politically charged framing of the Vatican's longstanding advocacy for AI governance frameworks that include ethical constraints and limits on autonomous weapons. The accusation misrepresents what the Vatican has actually argued but reflects an influential strain of American tech-policy thinking that treats any constraint on AI development as competitive concession to adversaries.
Drone shows are replacing traditional fireworks in more U.S. cities. The Brooklyn Bridge fire during the July 4th pyrotechnics display — a national landmark catching fire during the national celebration — will likely accelerate that trend by making the risk calculus of traditional pyrotechnics harder to defend to insurers and city governments alike.
The dominant consensus running through nearly every major business story of the weekend — Goldman Sachs reversing its dollar call, Foxconn's revenue surge, Micron's nine-billion-dollar groundbreaking, Macron and Modi personally chasing AI commitments — holds that the current AI investment wave is creating durable economic value and that productivity gains will justify the capital being deployed. The bull case has real empirical grounding: Foxconn's revenue happened, Micron's factory will be built, Harvard's cancer model is producing clinically useful predictions, and Alibaba's superconductor candidates were experimentally verified.
But the strongest version of the skeptical argument is not that AI does not work — it is that the infrastructure buildout is overbuilt relative to the deployment timeline. If enterprise AI adoption, meaning businesses actually changing their workflows at scale, takes a decade rather than three years, then data centers built for 2027 demand will operate at partial utilization until 2034. That is a materially different financial story than what current capital markets are pricing. The signal to watch is not chip orders — that is capital expenditure flowing through. The signal is enterprise software revenue from AI-native products: whether Fortune 500 procurement budgets are shifting toward AI tools at rates that reflect genuine workflow transformation, or whether adoption plateaus at the pilot stage. The infrastructure numbers are already in. The adoption numbers are not yet.