Scott says many strong open-weight models now come from Chinese developers including DeepSeek and Moonshot. He speculates Beijing could subsidize them to 'AI dump' cheap models, slow OpenAI and Anthropic growth, and spook a market where 40% of the S&P is tied to AI. That gives China a geopolitical lever beyond model performance.
Scott says crime is plummeting across the nation because tools like Flock make it harder to get away with crimes. He argues mayors brag about low crime, but the technology deserves credit and the upside is undeniable. The public rarely notices the deterrence because it prevents crimes invisibly.
Scott calls online gambling a crisis: legalization raises bankruptcies 30%, and one in two men under 25 bet on the Super Bowl. His advice is to reduce risk on screens—crypto, sports betting, aggressive posts—and take more risk offline in jobs, dating, and business.
Scott says only about 3% of U.S. households paid for an AI subscription as of February, and OpenAI had 50 million paid subscribers against roughly 1 billion users—about 5%. He argues the larger opportunity, and why Anthropic pulled ahead, is enterprise, making consumer subscription revenue look too thin to support the valuations.
Scott notes Anthropic has cited gross margins above 80%, but that figure excluded model training costs and revenue share with partners. He calls it 'WeWork adjusted EBITDA,' and the Financial Times estimates the effective margin may be closer to 40% to 50%. That gap matters when judging frontier lab valuations.
Scott floats a theory that frontier labs’ calls to pause AI could slow capex, make them more profitable, and let them cut prices against open-weight competitors. He frames it as normal market economics, not pure safety concern. That would help them defend premium margins against cheaper models.
Scott says an L2 analysis of its top 10% found three common traits: they were disproportionately female, came from elite schools, and were often athletes. He recommends work-sample tests because interviews can fool you. A gymnast from Yale was, in his words, an automatic hire.
Scott says the clinic called him back three times a week because he was tall and Jewish—what they called the 'chocolate and peanut butter' of sperm donation. Height was the top request, and his two blonde friends were barely called back. Donor selection reflects customer demand, not just health screening.
Scott compares open-weight AI to Old Navy and Southwest: roughly 80% of the frontier product at a fraction of the price. He cites UK testing where open models narrowed the cybersecurity gap, with GLM at $6 per task versus $15 for Claude Opus. That pricing gap pressures premium labs’ margins and valuations.
Scott argues you’re already surrounded by cameras in phones, ATMs, and streets, so collection is inevitable. The real problem is whether laws and institutions can restrict access, citing a Texas abortion-related lookup and an Indianapolis officer tracking his wife. He wants stringent privacy safeguards around use.
Scott advises the cybersecurity worker to tell his current bank about the sports-betting offer and ask it to match compensation and scope. Companies often don’t value people until they’re leaving, so transparency can win a better role without the ethical tradeoff. If he takes the job, Scott says forgive yourself because it is legal.