S5 Ep. 8 - Our addition to social media lead me to write a Book

S5 Ep. 8 - Our addition to social media lead me to write a Book

Fifteen years after the smartphone revolution, we are finally asking the right question about social media platforms and their impact on human behavior.The debate about whether social media is addictive has effectively ended, Mark Zuckerberg testified in a California federal court in February 2026, facing allegations that Meta deliberately designed features to addict children to its platforms. The trial brings together lawsuits from dozens of school districts and families, centering on whether Meta knowingly exploited psychological vulnerabilities of young users to maximize engagement and profit. Meta CEO and Chairman Mark Zuckerber arrives at LA superior Court (February 18, 2026) Austria has banned social media for children under 14, while the UK is considering similar restrictions and Australia has already implemented them. The European Commission announced in January 2025 that it is opening formal proceedings against TikTok under the Digital Services Act, investigating whether the platform breached obligations to protect minors from addictive design.The question now is what we do with the knowledge that our attention was systematically engineered away from us through deliberate design choices. When we strip away the moral panic and regulatory theater, what remains is a case study in how incentive structures shape product design, how business models determine behavior, and how fifteen years of data reveal patterns we should have recognized sooner.This is not about whether individual users personally feel addicted to Instagram or any other platform, but It's about understanding the deliberate systems built to compete for human attention, and what that competition costs us collectively in terms of mental health, productivity, and wellbeing.The teams you don't seeMeta employs psychologists, as do TikTok, Instagram, and YouTube, though not for the purpose of helping users manage their mental health. These professionals are employed to optimize engagement by identifying exactly which stimuli trigger dopamine release, which interaction patterns create habit loops, and which notification strategies maximize time spent on the platform.Court filings in the current litigation detail how features like infinite scroll, autoplay video and algorithmic content recommendation were specifically designed to exploit psychological vulnerabilities rather than serve user needs. CR Biological Infographic The discussions center not on how to help people connect meaningfully, but on how to increase daily active users. The focus is not on what value is being created for users, but on what the engagement rate metrics show. The metrics that matter are those that serve the business model, which means advertising revenue, which requires sustained attention, which requires time spent on the platform, which requires features that make leaving the platform psychologically difficult.These companies are not consumer technology businesses in the traditional sense. They are attention arbitrage operations that acquire your attention cheaply through free services and sell it expensively to advertisers. The product is not the application itself but rather you as a user, specifically your eyeballs and the time you spend viewing advertisements between content.Every feature decision flows directly from this business model. Autoplay exists because it reduces the friction between watching one video and starting the next one, thereby increasing total viewing time. Infinite scroll exists because stopping points create opportunities for users to leave the platform. Algorithmic feeds exist because chronological feeds have natural endpoints when you have seen everything new from the people you follow. Pull-to-refresh mimics slot machine mechanics because variable reward schedules create stronger habit formation than predictable ones.This is not accidental design or the unintended consequence of building useful products. It represents deliberate optimization by teams of PhDs in behavioral psychology, neuroscience, and addiction research who are specifically employed to maximize engagement regardless of the impact on user wellbeing. DesignLab - Creativity that connects The playbook from gambling and tobaccoThe techniques being used are not novel inventions but rather adaptations of proven methods from other industries. Slot machines use variable ratio reinforcement schedules because psychologist B.F. Skinner proved in the 1950s that unpredictable rewards create stronger compulsive behavior than predictable ones. The pull-to-refresh mechanism on social media feeds implements exactly this psychological principle, where sometimes you receive interesting content and sometimes you do not, creating the compulsion to keep pulling. iStock Tobacco companies employed behavioral scientists for decades to optimize nicotine delivery, package design, and marketing strategies to maximize addiction among their customer base. Internal documents from those litigation cases reveal the same language now appearing in social media lawsuits, including discussions about "heavy users," optimizing for "consumption occasions," and designing for "habit strength."The gambling industry refined these techniques even further through environmental design. Casinos remove clocks and windows to eliminate time awareness among patrons. They use chips instead of cash to psychologically distance spending from the feeling of financial loss. They provide free alcoholic drinks to impair judgment and decision-making. They design reward schedules that create near-miss experiences that feel psychologically similar to wins, thereby encouraging continued play.Social media platforms have systematically adopted all of these techniques. They use infinite feeds without timestamps to eliminate time awareness. They replace meaningful interaction with gamified metrics like likes and hearts. They use algorithmic amplification of content that triggers emotional response because emotional arousal drives engagement. They time notifications to interrupt other activities because the interruption creates anxiety that the platform then provides relief from through its content. Yale Insightsarticle cover Technology companies have hired the same types of researchers that tobacco and gambling industries employed, in some cases literally the same professionals. The playbook was already proven effective in other industries. The platforms simply applied these techniques at unprecedented scale with access to real-time behavioral data and the ability to conduct continuous A/B testing on millions of users simultaneously.What fifteen years of data revealsWe now have longitudinal studies tracking mental health outcomes correlated with social media adoption across populations, and the results present a consistent pattern. A 2023 study in the Journal of the American Medical Association found that adolescents spending more than three hours daily on social media face double the risk of depression and anxiety compared to non-users. Teen suicide rates, which had been declining for decades, began rising sharply around 2010, which coincides precisely with when smartphone adoption and social media use accelerated among adolescent populations. Statista.com Critics correctly note that correlation does not prove causation, and this is an important methodological point. However, when Meta's own internal research from 2019 found that 32% of teen girls reported that Instagram made them feel worse about their bodies, and when the company chose not to act on that research or disclose it publicly, the pattern becomes more significant. Correlation becomes pattern, pattern becomes evidence, and evidence becomes the basis for legal action.The UK's House of Commons Science and Technology Committee heard testimony in 2024 from researchers at Oxford and Cambridge showing that excessive social media use correlates with measurable attention span reduction, sleep disruption, and academic performance decline. These effects are not limited to fringe users or extreme cases but appear in typical usage patterns. According to screen time data, the median teenager now checks their phone 237 times per day, which represents a significant interruption to sustained attention and task completion.Adult outcomes show less dramatic but still significant patterns. Research demonstrates increased anxiety, reduced ability to focus on long-form content, disrupted sleep patterns, and what psychologists describe as "continuous partial attention," which is the state of never being fully present in any single activity because of the constant awareness of potential notifications and updates. Addyosmani design priciples The most revealing data comes from the platforms themselves through their own usage metrics. Average daily usage has increased year over year since these platforms launched. * Facebook users now spend an average of 38 minutes per day on the platform. * Instagram users spend 53 minutes on average. * TikTok users spend an average of 95 minutes daily. These numbers are not stable or plateauing but continue to grow because the optimization for engagement never stops.If these products genuinely served user wellbeing as their primary function, usage would be expected to stabilize at the level that provides optimal value and satisfaction to users. Instead, usage increases perpetually in pursuit of the engagement metrics that drive advertising revenue. This pattern suggests not user value maximization but rather successful implementation of addictive design principles.The world regulatory responseAustria's 2024 ban on social media for children under 14 represents a significant shift in how governments conceptualize these platforms, moving beyond content regulation or moderation requirements to regulation of access itself. The UK is SUBSCRIBE [https://giulioranucci.substack.com/]

S5 Ep. 7 - Why future of computing might not be on earth

Elon Musk just merged xAI into SpaceX with a plan that sounds absurd until you run the numbers: build data centers in orbit.Not small experiments but industrial-scale computing infrastructure floating 340 miles above Earth (same distance of NASAs telescope Hubble), powered by solar arrays, cooled by the vacuum of space, training AI models that would potentially bankrupt terrestrial facilities.The immediate reaction is skepticism. Rockets are expensive. Space is hard. Servers belong in buildings, not orbit.But strip away the science fiction framing and what remains is a serious economic calculation that every executive should understand: sometimes the cost of working around a constraint exceeds the cost of eliminating it entirely. Nature.com - The development of carbon-neutral data centers in space The constraint no one can optimize AI training has hit a wall. Not a technical wall. A physics wall. Training GPT-4 required approximately 50 gigawatt-hours of electricity. The next generation requires exponentially more. Frontier models now consume energy equivalent to small countries (es. Belgium), running for weeks or months continuously.Data centers already eat 1-2% of global electricity. AI is accelerating that dramatically. And here's the killer: in a typical hyperscale facility, cooling accounts for 38-40% of total energy consumption. For every watt your servers burn doing computation, you spend nearly another watt just preventing thermal failure.This isn't inefficiency but physics. Dense computing generates massive heat. Remove it continuously or the hardware dies.The industry has optimized relentlessly. Better airflow. Liquid cooling. Chip-level thermal management. But optimization has limits, and those limits are approaching faster than most realize. You can only pack so much compute into a building before cooling becomes impossible regardless of technology.Meanwhile, AI's appetite keeps growing. Meta, Google, Microsoft, and others are each planning to spend over $60 billion annually on AI infrastructure. Much of that goes to power and cooling.This is the constraint that changes everything. Not because it's insurmountable today, but because it's exponentially worsening and the traditional solutions don't scale. Haysfluidcontrols.com Math that makes orbit interestingMoving data centers to space sounds expensive until you compare it to the alternative.A 100-megawatt terrestrial AI facility costs $1-1.5 billion to build. Over ten years it consumes roughly 8,760 gigawatt-hours of electricity. At $60 per megawatt-hour, that's $525 million just for power. You spend 38-40% of that, approximately $200 million, on cooling alone. Add construction, land, taxes, maintenance and you're approaching $2 billion total cost of ownership.Now consider orbital economics: Historically (up to year 2000s), this calculation was laughable. Launch costs of $10,000-20,000 per kilogram made space infrastructure prohibitively expensive for anything except satellites and scientific missions.SpaceX's Starship changes this fundamentally. Current projections target $100 per kilogram to low Earth orbit, potentially as low as $10 per kilogram at full operational scale. That's not incremental improvement. It's two to three orders of magnitude cheaper.A server rack weighs 1,000-2,000 kilograms. At $100 per kilogram, launching one costs $100,000-200,000. At $10 per kilogram, it's $10,000-20,000. For comparison, the servers themselves often cost more than their launch at these prices.But weight adds up. A meaningful orbital data center requires thousands of racks, solar arrays, thermal radiators, communication equipment, structural elements. Even at optimistic weights, you're launching hundreds of thousands of kilograms.Here's where the calculation gets interesting. At $10-100 per kilogram, putting 100,000 kilograms of infrastructure in orbit costs $1-10 million for launch. The equipment itself might cost $50-100 million. Total upfront: $51-110 million. Made with Claude Pro Operational costs nearly disappear. Power is free once solar arrays are deployed. Cooling requires zero energy because thermal radiation to space is passive. No property taxes. No grid fees. No atmospheric corrosion. Maintenance becomes the primary ongoing expense.Compare that to terrestrial alternatives where you're spending $200 million on cooling energy alone over a decade, and the economics start shifting.Problem 1: Maintenance The most legitimate objection to orbital data centers isn't launch costs or power but maintenance.Terrestrial data centers have technicians on-site twenty-four hours a day. When a server fails, someone swaps it within hours. When a cooling system malfunctions, engineers fix it immediately. When software needs updating, IT teams access the hardware directly. This operational flexibility is taken for granted, but it's fundamental to how data centers work.In orbit, none of this exists. You cannot send a technician to replace a failed drive. You cannot physically access servers for diagnostics. Hardware failures that would take hours to fix terrestrially might render orbital infrastructure permanently degraded.This isn't theoretical. The challenge is well understood from decades of satellite operations. And it's exactly where major tech companies have been quietly building expertise.Amazon's Project Kuiper isn't just about internet connectivity. It's a testbed for operating thousands of satellites in low Earth orbit with minimal ground intervention. The project requires automated health monitoring, predictive failure analysis, and graceful degradation when components fail. These are precisely the capabilities needed for orbital data centers.Microsoft has been experimenting with underwater data centers through Project Natick, deliberately choosing environments where physical access is difficult and expensive. The goal wasn't really about underwater deployment. It was about learning to operate sealed, autonomous data center modules that could run for years without human intervention. The operational lessons translate directly to space.Google has been operating large satellite constellations and investing heavily in machine learning for infrastructure management. Their data centers already use AI to optimize cooling and predict hardware failures before they become critical. Extending this to orbital environments where prediction becomes essential rather than merely beneficial is a natural progression.SpaceX itself has been solving these problems at scale with Starlink. Managing over 6,000 satellites in low Earth orbit requires automation, redundancy, and the ability to route around failures seamlessly. When a Starlink satellite fails, the constellation reconfigures automatically. Traffic routes to healthy satellites. The network continues functioning. This is exactly the operational model orbital data centers would require.The pattern across all these initiatives is the same: build redundancy at the system level rather than relying on rapid physical intervention. Design for graceful degradation. Use software to route around hardware failures. Monitor predictively and replace proactively during scheduled maintenance windows rather than reactively when things break.In practice, this means orbital data centers would likely operate differently from terrestrial facilities. Instead of repairing individual failed servers, you might design for 20-30% redundancy and simply route computation away from failed units. Instead of frequent small maintenance interventions, you might plan for complete module replacements every few years, launched as part of regular refresh cycles.The economics work if component reliability is high enough and redundancy is cheap enough. In orbit, redundancy doesn't cost cooling energy. Adding 30% extra compute capacity doesn't increase your power bill by 30% because power is free. It only increases upfront launch costs, which at $10-100 per kilogram becomes manageable.This is where the major tech companies are placing strategic bets. Not necessarily on orbital data centers specifically, but on the underlying capabilities that make autonomous, high-reliability distributed infrastructure possible. Whether that infrastructure lives underwater, in orbit, or in remote terrestrial locations, the operational challenges are similar.The company that masters maintenance-free or maintenance-minimal data center operations gains strategic advantage regardless of where those facilities physically exist. And if orbital economics do eventually favor space deployment, that advantage compounds. Made with Canva Problem 2: FuelEach Starship launch requires approximately 3,400 tons of propellant, mostly liquid methane and liquid oxygen. Critics rightfully point out this represents enormous energy expenditure just to reach orbit. But context matters.A 100-megawatt data center running continuously for ten years consumes energy equivalent to burning roughly 2.6 million tons of natural gas. The propellant for a few dozen launches to establish orbital infrastructure is a small fraction of the fuel that would be burned powering that facility terrestrially over its operational lifetime.The strategic calculation isn't about eliminating energy use. It's about frontloading energy expenditure to access a location where ongoing energy is free and abundant. Solar power in orbit has no atmospheric losses, no day-night cycles, no weather interference. It's not 15-20% efficient like terrestrial solar. It approaches theoretical maximums. Moreover, SpaceX is developing propellant production from atmospheric CO2 and water using renewable energy. If launch fuel can be produced sustainably, the energy equation shifts even further toward orbital deployment.This is how infrastructure economics change. You spend significant upfront resources to access conditions where ongoing costs approach zero, the am SUBSCRIBE [https://giulioranucci.substack.com/]

S5 Ep. 6 - EU Inc: Scaling or fading in the AI era

Full Article HERE [https://www.linkedin.com/pulse/eu-inc-scaling-fading-ai-era-giulio-ranucci-6ybif/?trackingId=P7Yr59QCTCaBhgq%2Bh3oxPg%3D%3D] Europe has never lacked talent. It has never lacked research excellence, engineering depth, or ambitious founders. What it has consistently struggled to build is something else, a reliable path from innovation to scale.For more than twenty years, the European tech story has followed a recurring arc. companies are founded, validated, and often celebrated locally. But when growth accelerates, capital intensity increases, and global ambition becomes unavoidable, gravity pulls them elsewhere. Most often, to the United States. As of the mid-2020s: * the United States hosts 600+ unicorns * Europe hosts roughly 130–150, spread across multiple hubs (Source: StartupBlink, Atomico State of European Tech). Between 2018 and 2021, Europe experienced a strong acceleration in unicorn creation. But sustaining that momentum proved difficult.The main reason is capital depth. According to Atomico and Crunchbase data: * Europe consistently captures ~15–18% of global VC funding * the US captures 50%+, with a dominant share of late-stage rounds Since 2015, Europe has missed out on hundreds of billions of dollars in growth capital compared to the US particularly in rounds above $50M, where scale is determined (Atomico, Sifted).This is not a failure of founders or ideas. It is the consequence of how the system is designed. Scale means crossing the AtlanticThe list is familiar. Spotify chose a direct listing in New York. Elastic went public on the NYSE. Farfetch and Adyen built their global investor base outside Europe. More recently, conversations around Klarna or Bending Spoons point in the same direction: when scale and liquidity matter, US markets remain the default option.This preference has little to do with patriotism or branding. It has everything to do with market depth. American public markets, especially Nasdaq, offer liquidity, analyst coverage, and a class of institutional investors that understand growth, technology risk, and long-term compounding. European exchanges, still fragmented along national lines, rarely offer the same combination. The result is predictable. Capital shapes outcomes, and exits follow capital.The capital gap in numbersData makes the pattern hard to ignore. Over the past decade, the United States has consistently hosted more than half of the world’s unicorns. Europe, by comparison, has produced a fraction of that number, despite comparable population size and strong academic output. The difference becomes even clearer at later stages. While Europe captures a meaningful share of early-stage venture funding, it systematically underperforms in large growth rounds. Capital above $50 million the kind that determines whether a company becomes regional or global remains far more abundant in the US.Since 2015, European scale ups have collectively raised hundreds of billions less than their American counterparts. This gap does not reflect weaker ideas. It reflects a thinner, more fragmented capital market.And once US investors enter the cap table, strategic gravity shifts. Board composition changes. Exit expectations evolve. Public listings, when they happen, increasingly take place outside Europe. Europe is not short on exits. What it lacks are exits that reinforce scale. Most European tech outcomes still come through acquisitions, often by American companies. Large IPOs are rarer and less repeatable. In the US, by contrast, public markets function as a continuation of the venture ecosystem, allowing companies to raise capital, stay independent, and keep growing after going public.Without this feedback loop, Europe repeatedly trains companies for someone else’s market.Fragmentation as a structural taxThis is where the problem becomes concrete. Europe operates across dozens of legal systems, tax regimes, labor laws, and capital market rules. Each difference is manageable on its own. Together, they slow everything down.Founders face complexity when issuing stock options across borders. Investors face friction when deploying capital at scale. Companies face operational drag precisely at the moment when speed matters most. This fragmentation acts as a hidden tax on ambition.Citi Institute released a report, “Reimagining European Capital Markets: From Fragmentation to Harmonization”, which focuses on post-trade challenges and fragmentation in the context of capital markets in Europe, and the key benefits of a unified European capital market.In a world of geopolitical and macroeconomic volatility, Europe has an opportunity to position itself as an attractive alternative for investment, innovation, and influence. The moment for incremental change has passed. Shahmir Khaliq, Head of Services, CitiThe report outlines how a unified European capital market could add hundreds of billions in annual investment, boost regional GDP and retain European savings in their own region. The research sheds light on the critical need to address the deep-seated fragmentation within European capital markets with findings detailing the high costs, inefficiencies, and lack of innovation stemming from disparate regulations and infrastructure.As the report cites, a unified European capital market is estimated to add €150 billion ($175.5bn) in annual investments and positively impact the GDP by 1.5% over 10 years, driven by higher risk diversification opportunities, higher market liquidity, the availability of a safe asset and more appetite for investing in the EU of both domestic and foreign investors with a higher propensity for risk.Europe does not fail to create startups. It fails to make scaling feel natural. Why AI changes the stakesArtificial intelligence reshapes this equation.AI-native companies are global from inception. They scale through software, models, and data rather than physical infrastructure. They reward research depth, long-term thinking, and institutional credibility areas where Europe is strong.The rise of governance-first AI approaches, such as Anthropic’s Constitutional AI framework, reflects a broader shift. In AI, competitive advantage is no longer just about raw performance. It is also about trust, safety, and the ability to operate across jurisdictions.This should be Europe’s moment. But only if innovation is matched by infrastructure.Europe is rich in initiatives, funds, accelerators, and public programs. What it lacks is an integrated platform that allows companies to move smoothly from idea to scale, from one market to twenty-seven, and from private capital to public markets.Scaling does not fail because of vision. It fails because of friction. Why EU–INC mattersEU–INC is a policy initiative aimed at creating a single, pan-European legal status for startups the proposal seeks to eliminate the regulatory fragmentation of the 27 member states to make Europe a unified market for innovative companies, similar to the United States or China.Key Features of EU–INC: * Single Legal Entity: An optional corporate structure at the EU level that exists alongside national laws without replacing them. * Digital-First Registration: Businesses can register entirely online through a centralized EU registry. * Standardization: It introduces standardized investment documents and a unified framework for employee stock options, making it easier to attract talent and raise capital across borders. * 2027 Implementation: Following the legislative proposal in early 2026, the framework is expected to be fully operational by 2027. The initiative began as a bottom-up movement supported by a petition from over 15,000 startup founders (including leaders from Stripe, Wise, and DeepL) and investors. You can find more information and updates on the EU–INC official website below: 👉 https://www.eu-inc.org/ [https://www.eu-inc.org/] 👈 Rather than adding another program, EU Inc. proposes a structural shift: a pan-European legal entity designed specifically for startups and scaleups. A framework that reduces cross-border friction, standardizes key building blocks, and makes Europe legible and investable as a single market.In an AI-driven economy, where speed compounds and scale arrives early, this kind of infrastructure is not a bureaucratic detail. It is a strategic necessity."A startup from California can expand and raise money all across the United States. But our companies still face way too many national barriers that make it hard to work Europa-wide, and way too much regulatory burden." Ursula von der Leyen, Oct 2024 [https://www.youtube.com/watch?v=1m06dqXo4bw&t=823s]Europe does not need more talent. It needs to stop losing momentum at the moment it matters most.EU Inc. is not the entire solution, but it is a meaningful step toward turning Europe into a place where innovation does not just start it stays, scales, and competes globally.Giulio Ranucci SUBSCRIBE [https://giulioranucci.substack.com/]

S5 Ep. 5 - How 1990s game design created AI blueprint

In October 2024, the tech world marked a historic milestone: Demis Hassabis, CEO and co-founder of Google DeepMind, [https://en.wikipedia.org/wiki/Google_DeepMind] was awarded the Nobel Prize in Chemistry for his work on AlphaFold [https://alphafold.ebi.ac.uk/] (AI system by DeepMind that accurately predicts the 3D structure of proteins solving the long-standing "protein folding problem [https://pmc.ncbi.nlm.nih.gov/articles/PMC2443096/#:~:text=The%20protein%20folding%20problem%20is,first%20atomic-resolution%20protein%20structures.]"). To many, it looked like the culmination of a career devoted to “serious” science. But to anyone familiar with the history of game design [https://en.wikipedia.org/wiki/Game_design], it told a different story. It was a victory for the power of play.Long before founding Google DeepMind, Hassabis was a teenage game designer. At 17 (!), he co-designed Theme Park [https://en.wikipedia.org/wiki/Theme_Park_(video_game)] (1994), one of the most influential simulation games of its era. Today, as we enter 2026 an age where AI world models like Genie 3 generate interactive 3D environments from text prompts It’s becoming clear that many of the core ideas behind modern AI were first explored in games, not labs.The roots of today’s AI don’t begin with neural networks alone. They begin with simulations, sandboxes, and play.The Sandbox Foundation: Theme Park (1994)Theme Park wasn’t just a game about roller coasters. It was an early experiment in emergent systems.Unlike arcade games built on fixed rules and predictable outcomes, Theme Park simulated a living environment populated by autonomous agents visitors with needs, preferences, and reactions.If you placed a salty food stand next to a soda machine, guests became thirsty. If prices rose too fast, satisfaction dropped. If queues grew too long, behavior changed.The game didn’t follow a script. It responded. For Hassabis, this was a formative insight: intelligence could emerge from agents interacting with a complex environment, rather than being explicitly programmed step by step. This idea agent-based simulation inside a world model would later sit at the heart of DeepMind’s philosophy.What looked like entertainment was, in retrospect, an early rehearsal for artificial intelligence.The Grandmaster Benchmark: StarCraft II (2019)If Theme Park was the sandbox, StarCraft [https://en.wikipedia.org/wiki/StarCraft] II became the stress test.In 2019, DeepMind’s AlphaStar reached Grandmaster level, outperforming 99.8% of human players. This mattered not because it was a game, but because StarCraft embodies many of the hardest problems intelligence can face.Unlike chess or Go, StarCraft operates under imperfect information. The “fog of war” hides your opponent’s actions. For fo war on Starcraft To win a game its required: * Long-term planning: early decisions cascade into outcomes an hour later * Massive action spaces: thousands of possible moves at any moment * Real-time adaptation: managing hundreds of units while anticipating an opponent’s strategy This was no longer pattern recognition. It was situated intelligence under uncertainty.AlphaStar wasn’t just playing a game it was learning how to reason, adapt, and strategize in a dynamic world. 2026: From Games to World ModelsFast forward to 2026, and the lineage is unmistakable. AI has moved from narrow benchmarks to generalist systems capable of reasoning across domains. At the core of this shift is a familiar idea: internal simulation.World Models as Internal SandboxesModern AI systems increasingly function like advanced game engines. Models such as Genie 3 [https://deepmind.google/blog/genie-3-a-new-frontier-for-world-models/] simulate environments, physics, and cause-and-effect, allowing agents to “practice” inside virtual worlds before acting in the real one. This is Theme Park, scaled to reality.Strategic Reasoning Beyond GamesThe reinforcement learning techniques refined in StarCraft II now optimize logistics networks, power grids, and supply chains systems that closely resemble real-world strategy games. Efficiency, foresight, and adaptation are no longer about winning matches; they’re about saving energy, time, and resources.The games were never the goal. They were the training ground.Conclusion: Don’t Fear the GamerDemis Hassabis has often encouraged parents to support the creative use of technology. His own trajectory makes the case better than any manifesto.The skills honed through games spatial-temporal reasoning, strategic planning, adaptive thinking, and systems intuition are not distractions from serious work. They are its foundation. The next time you see a complex strategy game, don’t dismiss it as entertainment. You may be looking at the blueprint for the next scientific breakthrough.This Article was inspired by an interesting documentary about Deep Mind path fully and freely available on YouTube: The Thinking Game [https://www.youtube.com/watch?v=d95J8yzvjbQ&t=4840s] https://www.linkedin.com/pulse/power-play-how-1990s-game-design-created-blueprint-2020s-ranucci-mp1mf/?trackingId=Jllyl%2BhpQgOZkoqKXRrveA%3D%3D SUBSCRIBE [https://giulioranucci.substack.com/]

S5 Ep. 4 - YouTube: The rise of a media OS

SUBCRIBE [https://www.linkedin.com/newsletters/digitalstories-exe-7388138546441732096/] Read more here [https://www.linkedin.com/pulse/your-browser-next-tech-empire-giulio-ranucci-eiwcf/?trackingId=rYV2Kq0ISE2P9dPo%2Bcr%2BUg%3D%3D]: https://www.linkedin.com/pulse/your-browser-next-tech-empire-giulio-ranucci-eiwcf/?trackingId=rYV2Kq0ISE2P9dPo%2Bcr%2BUg%3D%3D [https://www.linkedin.com/pulse/youtube-rise-media-operating-system-giulio-ranucci-nbjke/?trackingId=5Y7OVvcmT9uAcH3ejPRT8w%3D%3D] Media competition has long been been framed as a content arms race, who had the best shows, the deepest library, the biggest stars. That framing is now failing.Netflix is reportedly moving toward acquiring [https://variety.com/2025/tv/news/netflix-to-acquire-warner-bros-82-7-billion-deal-1236601034/] Warner Bros., accelerating vertical consolidation.YouTube will stream [https://variety.com/2025/film/news/oscars-youtube-2029-1236610989/] the next Academy Awards a moment unthinkable just a few years ago.Even social media such as Instagram is extending its reach [https://about.instagram.com/blog/announcements/instagram-tv-app] to TV. Regulators and measurement firms now show YouTube overtaking [https://www.wsj.com/business/media/how-youtube-won-the-battle-for-tv-viewers-346d05b8] traditional broadcasters and competing directly with Netflix for TV viewing time. These events are not disconnected. They point to a deeper structural shift in how the media industry is being reorganized.From Platforms to Power LayersHistorically, media worked as a linear stack:Studios → Networks → Distribution → AudienceStreaming disrupted the middle, but kept the logic intact: premium content flowed through controlled pipes to paying subscribers. YouTube breaks this logic entirely, It is not just a distributor. It is an attention infrastructure a system that captures, routes, and monetizes audience habit at scale. YouTube operates simultaneously as: 1. the largest ad-supported TV network 2. the default video app on connected TVs 3. a creator economy engine 4. a search and discovery layer 5. and increasingly, a live event broadcaster When the Oscars move onto YouTube, it’s not about prestige it’s about acknowledging where audience habit already lives.Distribution is the scarce assetOnce again we tend to see the future of media as a content arms race. But the real scarce asset today is distribution with habit. YouTube’s strategic advantage is not just scale [https://www.charleagency.com/articles/youtube-statistics/], it’s default behavior: * Younger audiences turn on their TV and open YouTube first [https://www.theguardian.com/technology/2025/jul/30/youtube-tv-destination-children-ofcom-survey] * Long-form viewing on YouTube continues to grow, [https://www.tubefilter.com/2025/02/06/youtube-long-form-viewership-increase-shorts/] not shrink * It captures attention across [https://www.theatlantic.com/podcasts/2025/12/how-youtube-ate-podcasts-and-tv-with-rachel-martin-ashley-carmen-and-derek-thompson/685230/] formats: shorts, podcasts, live, sports, news, entertainment Instead Netflix’s acquisition of Warner Bros. reflects the opposite pressure: * When distribution is capped, content ownership becomes defensive. * Consolidation becomes a way to preserve pricing power in a closed ecosystem. This is the strategic divergence: * Netflix is consolidating content to defend a subscription model. * YouTube is expanding infrastructure to own attention itself. YouTube as the Media Operating SystemCalling YouTube a “platform” is increasingly inaccurate. Advertisers are following the attention and the money. More streaming is translating into stronger ad performance. According to Google’s Q1 2025 earnings [http://www.thekeyword.co/news/google-ad-revenue-hits-66-9-billion-in-q1-2025], YouTube brought in $8.93 billion in ad revenue, up more than 10% yoy and rising from $8.0 billion in Q1 2024.Marketers are using YouTube to reach audiences across formats, age groups, and content types. The numbers show that advertisers see real value in YouTube’s reach, especially on TV. This shift could be important for advertisers because it changes the context in which ads are seen. Unlike mobile, TV offers larger screens, longer watch sessions, and a more relaxed viewing environment all of which influence ad performance.There’s a changing media reality. More eyes on YouTube streaming mean more premium ad inventory and more pressure to compete in a platform-first ad economy. With YouTube Shorts, long-form content, and YouTube TV all under the same umbrella, the platform offers a mix of formats that brands can explore depending on their campaign goals.A more precise framing: YouTube is becoming the operating system for video. Just as Windows didn’t win by producing the best software but by being the layer everything ran on YouTube is winning by becoming the execution layer for media itself. Creators, studios, advertisers, and institutions now all route through it.Traditional broadcasters no longer compete against YouTube they depend on it. Even public institutions like the BBC rely [https://www.bbc.com/news/articles/c4gzvee78eqo] on YouTube to reach younger audiences at scale.This is why YouTube streaming the Oscars matters. It formalizes a reality that already exists.Netflix vs YouTube Is not a streaming warSo this is not a Netflix-vs-YouTube competition. It’s a clash between two fundamentally different models: Made with Nano Banana (and related allucinations) Netflix must keep feeding the machine to survive. YouTube lets the machine feed itself.From a strategic standpoint, YouTube does not need to replace Hollywood. It needs to out-distribute it.And increasingly, it already has.Media Is Rebundling UpwardWhat we are witnessing is not disruption it is rebundling at a higher layer: * Studios merge to survive distribution pressure. * Streamers consolidate IP to protect margins. * Platforms evolve into infrastructure. The winners of the next decade will not be defined by awards, catalogs, or individual hits but on control of attention, default placement in the living room and ownership of the creator-to-audience relationship.Seen through this lens, Netflix acquiring Warner Bros. would be a defensive move. YouTube streaming the Oscars is an offensive one.One is trying to preserve the old model. The other is quietly becoming the foundation of the new one.Giulio Ranucci SUBSCRIBE [https://giulioranucci.substack.com/]