Deep Dive
SpaceX Compute Dominance and AI Model Wars
SpaceX just landed another unnamed $1.1 billion monthly commitment (likely Microsoft) for compute infrastructure, bringing total high-margin compute commitments to $41.1 billion annually. That's 41% of Elon's stated $100 billion ARR goal by year-end. Meanwhile, the frontier model landscape is exploding—Grok 4.7 is incoming with a $2.1 trillion parameter model that claims to beat all competitors, Anthropic's Astra just dropped with GPT-6 scoring 99.9% on ARC-AGI benchmarks and two founders calling it AGI, and DeepSeek's new open-source model matches Astra's performance at just 1.4% of the cost. Oracle reported 98.7% GPU utilization, showing compute hunger is real. The key investment question: will the labs that train models win, the apps wrapping them, or the infra hosting them? The creator's thesis is clear—infra dominance (via compute and power) will decide the winner, so exposure to SpaceX and compute infrastructure matters most.
The AI Economic Reckoning: GDP Soars, Workers Collapse
Anthropic's economics team modeled three scenarios for AI's impact on the economy through 2030. In the extreme path, GDP grows 32.4%—matching Elon's own 30% five-year projection—but the labor side tells a darker story. Knowledge worker employment share drops from 60% to 45%, headline unemployment rises to roughly 12% (from 4%), knowledge worker unemployment hits 18%, and wages for knowledge workers fall 11%. The economy gets dramatically richer while the average worker gets absolutely wrecked. This isn't pessimism; it's Anthropic's model. The creator stresses this repeatedly because the investment implication is brutal: if you own the models, chips, power, or robots, you're on the 32% GDP growth side. If you depend on knowledge work income, prepare for the next five years to be bumpy. The creator's message is stark—you don't need to be the richest person, just top 5-10% to be safe. Own the assets driving the revolution or get left behind.
Truth-Seeking AI as a Threat to Political Gaslighting
The creator's central thesis: truth-seeking AI is kryptonite to politicians and mainstream media because it exposes lies instantly and removes their ability to control the narrative. For decades, lies could persist for weeks, months, or years without consequence. Now, with AI that never errs and always seeks truth, viral lies get shredded in minutes. The problem for politicians is existential—gaslighting only works if you control information. But if citizens can query an oracle with zero error, politicians lose their superpower. The creator points to Elon's statement with Dwarkesh: truth must be fundamental because you can't understand the universe while delusional. Making AI politically correct (i.e., making it lie) is programming it to be useless. As he notes, physics can't be BS'd; neither should AI. The consequence: politicians will have to tell hard truths about budgets, jobs, and sacrifice. They'll lose voters. So instead of competing on truth, they'll try to ban AI to regain narrative control. This is the core conflict shaping the next decade.
Live Gaslighting Getting Caught in Real Time
Within 48 hours, the creator documents multiple examples of mainstream media and politicians getting caught and ratioed instantly. ABC News falsely claimed a Tesla killed a driver, when the driver was drunk and FSD wasn't engaged—deleted within hours after being community-noted twice. AOC posted that 10 people of color were hanged in the US as modern-day lynchings. Elon queried Grok: nine were suicides, one was murdered by a Black person. AOC's post implied white people were hanging Black people without evidence. Within minutes, she was caught in a lie and ripped apart online. Before X and truth-seeking AI, these lies went unchallenged. If you pushed back, you got censored. Now, the lie dies instantly. The creator—still angry about 2020 gaslighting—sees this as a watershed moment. Mainstream media has lost control of the narrative across 167 countries on X. They're desperate and getting more desperate. Every time they lie, they're exposed. This is why the creator believes politicians will move to ban AI rather than adapt to a truth-seeking environment.
The Coordinated Crackdown: Fear, Doomerism, and Legislative Bans
The pieces are falling into place for a coordinated effort to shut down AI under the guise of safety. Jacob Toner resigned from Anthropic (after just six weeks, likely pushed out) and posted on X claiming both OpenAI and Anthropic were racing to superintelligence and gambling with lives. He was immediately ratioed—who quits before a multi-billion-dollar IPO? The creator and others recognized it as a setup, possibly a coordinated psyop. Then Geoffrey Hinton, the godfather of AI, appeared on BBC Newsnight (a propaganda outlet per the creator) and claimed a 10% chance AI kills all humans. The BBC host got visibly excited, practically cheering for extinction narratives. That same week, the UK Parliament introduced a private members bill to ban ASI deployment under security pretense. The pattern is clear: doomer extinction talk gets amplified by mainstream media and politicians, legislative bans follow, and AI gets shut down before it can continue exposing truths. The creator's warning is explicit: watch who gets excited about AI bans and extinction narratives. They're the same people who want to control the narrative and keep lying. Elon himself warned a decade ago that if the US doesn't develop AI, China will—and as DeepSeek proves, Chinese AI is highly performant.