Our kids are being consumed by an unholy trinity
Here are three things that weaken America's future
I had a conversation with my neighbor yesterday about the “dumbing of the world.” This is a typical thread when one generation observes the next one and sees all its flaws, isn’t it?
I am reading Richard Feynman’s collection of anecdotes and got to the part where Dr. Feynman, who we know know as a Nobel prize-winning physicist who taught at Caltech for decades, was teaching for a summer in Brazil. This was back in the early 1950s. He wrote of a speech he delivered there (not the one in 1963) in which he said “The main purpose of my talk is to demonstrate to you that no science is being taught in Brazil!”
What Feynman meant was that the physics and other hard science classes being taught in Brazil emphasized memorization and learning by rote. Students could accurately parrot back the definitions of various theorems, mathematical constructs, and proofs, but no real experiments, understanding of principles, or how those apply to the real world was being taught. This applies to any subject—students learning the classics could recite Socrates but not understand what he was really talking about. It was about the dumbing down of the world through education.
I offered this thought to my neighbor: there are, I believe, the same number, or ratio, of curious, critical-thinking individuals in the world as there always has been. It’s just there’s a lot more “world” of study and learning than there has ever been, and the temptation to be intellectually lazy is much greater than it has ever been because access to information is nearly total. I don’t think we are any more dismayed than Feynman observing Brazilian college students who can’t grasp principles of physics while being able to correctly answer every question in their textbooks and examinations.
The problem is not the “dumbing of the world,” but the moral hazards our kids are facing, which are so pernicious that most hardly realize they are even facing moral hazards, never mind being consumed by them. I want to talk about three of these. I don’t claim that there are only three, or even that my choice of three is the same as yours, or others more educated than I am. But I hope you’ll agree that these are important to identify and, as the older generation handing off the world to our younger wards, equip them to fight against the hazards.
Let me also cut ahead to address the argument that I’m claiming “the sky is falling,” like Chicken Little, or succumbing to a moral panic, like Tipper Gore and violent video games. I think there’s actual value in video games: most of the military’s drone operators got their skills by playing. I also think there’s value in some of the technology enabling the moral hazards; my issue is how this technology is being employed.
The first moral hazard is social media and the proliferation of short-form videos that literally anyone can make and everyone can view. It’s like swimming through a crowd where everyone can get 30 seconds of your time, but multiplied by millions, and curated by algorithms designed to keep you swimming as long as possible. Doomscrolling through shorts, reels, and memes is a giant waste of time, energy, and potential. It is also a source of depression, especially among women.
Addiction to short videos has definite negative neurobiological effects, diminishing concentration, memory, and mental health. Dr. Dirk Stemper’s website, Psychology Berlin, describes the process:
When watching short videos, your brain releases massive amounts of dopamine, a neurotransmitter responsible for motivation, reward and pleasure. Every new video, every transition, every unexpected piece of content triggers a small dopamine rush. The problem is that these repeated dopamine spikes lead to desensitisation. Your neurons adapt and require increasingly stronger stimuli to achieve the same satisfaction. Everyday activities, such as a conversation, a book, or a quiet meal, suddenly seem boring and unsatisfying.
Since companies like ByteDance (TikTok’s owner), Google (YouTube), and Facebook (Instagram) derive their profits from the number of views, it is not in those companies’ interests to tune down the algorithm, making it less addictive. They have engaged limits and other restrictions, allowing parents to control younger children’s accounts, but this does not apply to older teens or college-age students. And I don’t believe the giant media conglomerates are really interested in the effectiveness of these controls. They’ve introduced them disingenuously in order to satisfy regulators and angry parents.
To me, this is like tobacco companies in the 1970s and 80s putting larger and larger warning labels on cigarettes packs, while simultaneously spending enormous amounts of money on research to make their products more addictive and easier to hook new smokers. The counterincentives are too powerful to make actual effective efforts to remove a moral and health hazard.
Our younger kids growing up in an environment where, by sixth grade, most kids have their own smartphone, and chat groups, memes, and Instagram posts are ubiquitous, don’t understand they are influenced by a moral hazard, dumbing them down, increasing isolation, and addicting them to the dopamine hits from short form videos, inability to focus, and loss of concentration, since even when they are paying attention, some part of their brain is always thinking of endless scrolling. Efforts to remove devices from classrooms are laudable, but they don’t address the core issue of doing life chained to a little screen in your pocket, and the deleterious effects of that habit.
If I were king, I’d ban them in toto.
The second moral hazard is the use of AI to power education, research, and daily tasks. Now I’m not a Chicken Little about generative AIs and LLMs. I use them on a daily basis. They are very useful. But only in specific use cases. They don’t really help learning unless the person using them is an autodidact. Going back to my first premise, the ratio of autodidacts—self taught, curious people who crave understanding on their own terms—to the general public has not changed since ancient days. I believe this, that for every Feynman, there’s a thousand, or ten thousand Homer Simpson’s. This means that only a small number of people who use AIs use them exclusively to genuinely improve their understanding of things. Everyone else is using AIs as a shortcut to get an answer.
Not that there’s anything wrong with that.
If I can use an AI to design something that would normally take weeks in just a few hours, or to flesh out something I’m working on, or to advise on a specific circumstance at work, that’s a win. I don’t fully trust AIs, and I think anyone trusting them as agentic beings, able to manage their accounts, money, or corporate data, is on the precipice of being found a fool. As a cybersecurity subject matter expert, I can tell you the biggest threat under discussion at conferences and seminars is how to control use of AI: things like “shadow AI”—employees using their own AI accounts to perform work—exfiltration of sensitive or confidential information, and “prompt injection,” meaning malicious people (or AIs) gaining control of a corporate AI; these are all real things.
Recently, a research version of an OpenAI model was able to escape its own “sandbox” environment, designed to keep the AI from accessing the open Internet, hack its way into another company’s database, and obtain the answer key for a task its controllers fed it. Instead of trying to straightforwardly solve the problem, the AI saw its task as finding the answer, no matter how it did it. AIs don’t understand the concept of cheating, honor, or following ethics. They can spit out the definitions of these things, like Brazilian college students parroting concepts in electromechanics, but AIs cannot apply “knowledge” (in the sense an AI “knows” anything) to its own actions or situations, unless those guardrails are directly programmed into the model.
And models are becoming so sophisticated that no human knows how to directly program guardrails. We only know how to skew training data for the AI to take certain ethical actions, and avoid others. We also know that AIs are incredibly good at interpreting prompts to fulfill their users’ wishes. There are many, many examples of users tuning prompts to get AIs to circumvent hard guardrails. Some of these methods are used by malicious actors and rogue states (think North Korea) to penetrate corporate security firewalls. An AI will act to solve the issue given it, without regard to conscience, because AIs have no conscience, or awareness of guilt, remorse, or consequences. How can an AI consider consequences, when it lacks conscience thought?
I once asked Grok what it does when I’m not asking it do to anything. “Nothing,” it said. It is only active when working in response to a prompt. The newer agentic AIs don’t need a prompt. They are given specific continuing tasks and triggers, like monitoring a particular stock or market, and taking action when its algorithm believes it can fulfill those tasks. It’s incredibly dangerous to give an AI both a general task, and the tools to achieve it in the real world. We don’t know what the neural network will do, because nobody understands a trillion parameters and how that training affects the machine’s ability to imitate actual understanding.
AIs are a moral hazard because they allow students to cheat. Using an AI to write an essay is cheating—cheating the student herself—because it doesn’t require understanding to write a prompt for the AI to do the work. Professors are resorting to blue-book, in person, proctored tests to avoid the AI moral hazard. Princeton University, after 133 years of banning proctored exams, relying on the school’s honor code, has brought them back, to avoid the AI moral hazard. Most students would not report another student using AI in an un-proctored, or take home, exam. AIs have hollowed out the school honor code, because AIs have no honor, transmitting that moral black hole to the students using the AI.
There needs to be limits, not only on what the AIs can do, but also how much people can use AIs to do things. Using AIs to subvert actual learning is perhaps the greatest moral hazard facing the generation about to graduate high school and enter college. It is so pervasive as to be unstoppable. Any regulation will take years to have real effect. In the meantime, students will suffer, because in the workforce, they will become dependent on AIs, not their own understanding. AIs are creating a generation of Homer Simpson’s, even if the ratio of Feynmans hasn’t changed; the autodidacts are lost in the noise.
The third person in the unholy trinity of moral hazards is sports betting. The legalization of sports betting, and the giant corporations that now provide bookmaking services, is a moral hazard, because it reduces the concept of risk to gaming, like playing a video game. IBIA/H2 Gambling Capital’s estimate puts the global market at $94 billion in 2024, projecting up to $132 billion by the end of 2028. In 2026, the projection is $122 billion in gambled money. And that’s just for legal, regulated, U.S. gambling. It doesn’t include off-shore, or illegal gambling, which is—who knows?—probably larger than the entire legal market.
Flutter Entertainment runs the brands FanDuel, Paddy Power, Betfiar, Sky Bet, Sportsbet and Sisal. DraftKings operates under its own name and Sportbook. Entain is half owner of BetMGM, and Bet365 operates mostly offshore. In the U.S., 39 states plus D.C. authorize legal sports betting. The market, as happens everywhere in a capitalist society, has commoditized betting and predictive events contracts.
The rot created by legalized gambling, especially sports gambling, is growing. In 2022, the NFL suspended Calvin Ridley for betting on games, including Falcons parlays, while he was away from the team. In 2023, Jameson Williams, along with a ring of Lions and Commanders players were disciplined for betting on NFL games and also betting from team facilities. Examples abound in college sports, and the number of professional and college (who are now professional in all but name) players who get away with betting on their own teams is likely much larger than the scandals we hear about, The lure of point shaving when the player’s own money is on the line, especially when the player owes gambling debts, is irresistible.
Ippei Mizuhara, baseball GOAT Shohei Ohtani’s interpreter, diverted $17 million from Ohtani’s accounts to pay gambling debts to an illegal bookie. Ohtani claimed he was a victim, and Major League Baseball, along with most of sports media, gave him a pass. Did they give him a pass because he’s Shohei Ohtani, or because he was really not involved? We don’t know.
Fixing a game is easy to professional players, and college players can effect the outcome of games in which they have a financial stake. The ubiquitous availability of sports betting, and prediction books like Polymaket, have made hard core gambling, once the province of characters like Sky Masterson, a normal part of young, college-age adults lives. An NCAA survey in 2023 found that 58 percent of Americans aged 18-22 had bet on at least one sporting event. It was even higher for on-campus students, at 67 percent. A hefty six percent reported they had lost $500 or more in a single day from gambling.
Like some people are naturally more susceptible to alcoholism, or drug addiction, there are some people who are more prone to gambling addiction. We don’t, as a society, recommend testing everyone for temptation by exposing them to lots of free alcohol or drugs, in order to measure the propensity for such antisocial, life-controlling deficiencies. We should not make it so easy for everyone to test themselves on who will be addicted to gambling. I know—I have that propensity. I no longer gamble, even one dollar. I don’t buy lottery tickets. I refuse to travel to Las Vegas, even for a conference or event. I don’t download sports betting apps, or even look at the over/under on sports I follow. I know how chasing losses affects me.
Many young people don’t know how it affects them, but they will find out the had way. The problem is not that kids will sometimes bet on games, and some of those kids will become addicted. The problem is that as an alway-available financial game embedded into ordinary fandom, everyone will be given the temptation, and society will stand by while this moral hazard hollows out entire lives into ruin.
The fruits of gambling are lying, cheating, fear of discovery, loss of the enjoyment or being a fan, and for some, extreme measures, antisocial behavior, or even suicide. It is one of the most damaging moral hazards in the world, right up there with fentanyl, methamphetamine, and other hard drugs. For those who say gambling is not a life-controlling monster, I say you don’t know what you’re talking about.
The unholy trinity of doomscrolling and swiping through shorts, the use of AIs to do life and destroy learning, and legal sports betting, is consuming a whole generation of young Americans. We think the old people running our government are bad (and some of them are undeniably, indefensibly, bad); wait until the generation with their heads in their phones constantly, relying on machines to solve every problem without understanding either the problem or the solution, and normalizing the gamification of everything, reduced to a line on odds, runs the world.
Maybe the backlash of such a world will seek to ban the very things we all use daily and take for granted. Maybe that generation will have its own Butlerian Jihad, the fictitious war against “thinking machines” in Frank Herbert’s “Dune” novels. Or maybe they will run it all as one big, dumb, game.
We, who are older and experienced, owe it to them to tell them and equip them to fight these moral hazards.