AI Doom vs Boom, EA Cult Returns, BBB Upside, US Steel and Golden Votes

Sacks argues that extreme claims about imminent mass job loss, bioweapon creation, or guaranteed extinction (X-risk) go far beyond available evidence and are tightly correlated with Anthropic fundraising milestones. He traces funding and personnel links among Effective Altruism (EA), Dustin Moskovitz’s Open Philanthropy, Anthropic, and key Biden AI staffers, contending that this network is pushing for “global compute governance”: tight controls on GPUs, international AI safety regimes, and values-laden regulation that would anoint a few winners and embed ideological biases in models.

Friedberg frames AI as a massive increase in return on invested capital: one engineer can produce 20–50x more code, which makes new projects economically attractive rather than obsolete. Historically, such technology shocks drove higher capital deployment and job creation, not permanent mass unemployment. He expects deflation in many goods and services (e.g., automated food prep dropping a latte from $8 to $2), allowing people to maintain or improve lifestyles with fewer working hours as abundance increases.

Driving and level‑one support are cited as rare cases where entire roles may largely disappear. More commonly, AI will automate “chores” inside complex jobs (sales, engineering, HR) rather than eliminate those jobs wholesale. Chamath notes that GPTs are now effectively replacing what new grads used to do as “glorified autocomplete,” making entry‑level hiring tougher in big firms. His advice to new grads: become AI‑native, master the tools, and gravitate to startups or start companies rather than chase traditional feeder roles.

Sacks argues the only AI dystopia we have concrete evidence for is government and Big Tech using AI to enforce ideological narratives and social control—evidenced by “woke AI” behavior and DEI mandates in the Biden AI executive order. He warns that concentrating AI power in a few heavily regulated, politically aligned firms is more dangerous than today’s “autocomplete” models, and that fearmongering is a classic tactic to stampede the public into ceding more power to government.

Friedberg argues AI, like the Industrial Revolution or the internet, will ultimately enrich multiple countries without a single “winner.” Sacks partially agrees on the infinite-game framing but emphasizes realist geopolitics: AI is a dual‑use technology that will power future militaries (drones, robots), so both the U.S. and China must assume the worst and race to avoid being outclassed. Market share and tech stack dominance (e.g., which country’s chips, data centers, and software underpin global AI) are his practical definition of “winning.”