Why Movies, Tech & Mental Health Feel Broken - Jeffrey Katzenberg & Hari Ravichandran (4K)

The analogy was developed by Hari and Jeffrey: driving requires a learner's permit, supervised practice, and a seatbelt, but the digital world has none of that. Aura's approach provides the equivalent: insights like 'your daughter downloaded a calorie tracking app, which is step one toward an eating disorder' enable early conversations, while instant alerts for severe threats are the digital seatbelt. The system is flexible across ages and parenting styles, allowing parents to dial up or down oversight. Jeffrey stressed that this is not an anti-tech crusade—social media has many positives—but about giving parents the tools they lack.

Jeffrey described being fired from Disney at the peak of his success and treating it as a learning moment. His motto 'never let your memories be greater than your dreams' means he wakes up each day excited about today and tomorrow, not yesterday. In 7.5 years of near-daily contact, Hari has never heard him discuss past glories or turkeys — only what's coming. Jeffrey noted that of ~400 movies associated with him, 80% were flops, yet he rarely thinks about them. Hari observed that Jeffrey handled the failure of Quibi with extraordinary grace, simply owning it and moving on, a lesson Hari now applies in his own life.

Hari's family went through 23 weeks of DBT after his daughter's hospitalisation. He noted that therapists are overwhelmed (6,000+ clinicians with ballooning caseloads), so parents must accelerate their own learning. Aura's companion resources include Q&A with clinicians and CBT/DBT primers. Hari believes that even a 30-minute introduction can give parents tools to have constructive conversations and not feel helpless.

Hari noted that Pinterest pioneered this concept: after a series of similar posts, ask 'Did you actually want to see this?' and if the answer is no, the algorithm shifts. He argues this is a simple way to give back control and that more of these types of features would be beneficial. Chris added that the current user-tech relationship feels adversarial; such nudges could restore a sense of partnership.

Chris Williamson introduced the idea from Stuart Russell's book, noting two ways algorithms improve click prediction: becoming more accurate at guessing preferences, or altering the user's preferences to make them more predictable. Hari elaborated with the example of a child searching for a car crash photo for a school project: the algorithm then floods the feed with more crashes, trapping the user in a self-reinforcing loop. He argued that we have 'hacked ourselves' because 10 million engineers are incentivized to maximize time on site, and kids are especially easy to program. Jeffrey added that humans have dark thoughts, and this feedback loop can push people into scary territory. The implication is that guardrails must account for the algorithm's ability to manufacture new vulnerabilities, not just mirror existing ones.

Hari observed that the 'truth of their lives are not in physical land it's almost on their device.' He and Jeffrey pointed out that kids spend 8+ hours a day online, more than they sleep. Chris referenced writer Freya India's point that because teens spend more time on devices than in the physical world, the digital world is more real to them than real life. Jeffrey illustrated the parenting blind spot: a child sitting across the dining table appears safe, but they might be in a completely different, often toxic world. This explains why emotional distress from online rejection or bullying hits as hard as physical events, and why banning phones feels like taking away a child's entire social universe.

Jeffrey read Aura's numbers: 46% depressed, 35% social withdrawal, 22% scrolling late at night, 30% low self-esteem, 22% with self-harm or suicidal thoughts, 52% with eating issues, and 80% of girls 18 and under don't like their body shape, with over half taking harmful steps. Hari validated that Boston Children's Hospital and others see 10x increases in severe cutting cases requiring medical care. He explicitly tested the 'just better identification' hypothesis by asking hospitals, who confirmed it's not just more reporting—something in the environment broke. He noted the data is from a diverse sample of 2,500 then 10,000 users, spanning geographies and backgrounds, making it robust. The impact is now becoming visible as those children age into middle school.

Jeffrey discovered Walt Disney's archival materials when he joined Disney in 1984, including the villain rule and the related idea that 'there's no such thing as a great story without a great ending.' He listed villains like Ursula, Scar, Jafar, Farquaad, and Tai Lung, noting how their menace elevated the hero's journey. Chris connected this to the psychological 'peak-end rule' (Kahneman): memories are dominated by the peak intensity and the ending, so a strong villain creates the peak challenge and a satisfying ending amplifies the experience. Jeffrey said this rule became his internal compass for evaluating scripts and productions, and remains a universal lesson in storytelling.

Jeffrey explained that he has a 'very weak memory gene'—he rarely reminisces about past successes or failures, adhering to the motto 'never let your memories be greater than your dreams.' He got fired from Disney during the summer of 1994, a period of unprecedented dominance (Lion King, Home Improvement, Beauty and the Beast on Broadway), and he freely says he owns that fiasco. He noted that of the 400+ movies he was involved with, 80% were 'dogs', and he learned more from the misses than the hits. Hari observed this firsthand: in 7.5 years of daily contact, they've never discussed the past—only what's next. Hari added that seeing Jeffrey's grace after the Quibi failure taught him it's far easier to just say 'I messed up' than to spend energy positioning.