All-In with Chamath, Jason, Sacks & Friedberg

Trump: Send National Guard to SF, China Rare Earths Trade War, AI's PR Crisis

October 17, 2025

Key Takeaways Copied to clipboard!

  • San Francisco's crime statistics show significant improvement under the new administration, contrasting with the narrative suggesting an immediate need for the National Guard as proposed by Donald Trump. 
  • The US government may need to implement price floors or strategic reserves for critical rare earth minerals to counteract China's mercantilist tactics and secure domestic supply chains. 
  • The recent cancellation of planned data center builds by major tech companies (Google, Microsoft, Amazon) signals the beginning of a trend where local opposition over energy costs and water consumption is forcing AI infrastructure projects to pause or relocate. 
  • The public perception of AI is currently negative due to concerns over job displacement and rising electricity costs, necessitating tangible economic benefits to gain acceptance. 
  • AI is currently a significant driver of US economic growth, contributing substantially to GDP growth, and historical precedent suggests technological revolutions shift, rather than eliminate, human labor to higher-skilled roles. 
  • The AI industry needs hyperscalers to proactively use their free cash flow to mitigate local concerns (like electricity costs) and needs more eloquent spokespeople to counter negative media narratives and 'doomerism'. 

Segments

Dreamforce and SF Shenanigans
Copied to clipboard!
(00:00:00)
  • Key Takeaway: David Sachs was interviewed by Mark Benioff at Dreamforce, leading to critical coverage from the San Francisco Standard.
  • Summary: The hosts recount attending Mark Benioff’s Dreamforce event, noting Matthew McConaughey’s presence at their table. David Sachs’s interview with Benioff was framed negatively by the San Francisco Standard, suggesting a continued ‘cancel culture’ environment regarding political associations. The hosts also humorously referenced a concurrent, alternative conference called ‘SlutCon’ discovered via social media algorithms.
San Francisco Crime and National Guard
Copied to clipboard!
(00:03:35)
  • Key Takeaway: Despite calls for federal intervention, San Francisco’s crime rates are demonstrably decreasing under the new mayor and DA.
  • Summary: The discussion contrasted Donald Trump’s suggestion to send the National Guard to San Francisco with recent positive local statistics, including a 30% citywide crime reduction and record lows in homicides. Mayor Daniel Lurie and DA Brooke Jenkins are credited with improving enforcement, evidenced by high felony conviction rates in narcotics and robbery cases. The segment highlighted that Gavin Newsom previously demonstrated the city could be cleaned up quickly for high-profile events.
SF Homelessness Funding Absurdity
Copied to clipboard!
(00:16:22)
  • Key Takeaway: San Francisco spends an estimated $52,000 annually per homeless person, funding NGO ‘flocks’ that incentivize maintaining the population rather than solving addiction.
  • Summary: The hosts criticized the city’s $2 billion annual budget for homelessness services, noting that the per-person spending far exceeds reasonable estimates. A specific example of program absurdity cited was the $5 million Managed Alcohol Program, which provides free beer to participants. This system creates a magnet effect, attracting unhoused individuals from outside the city by funding their addictions.
US-China Rare Earths Trade War
Copied to clipboard!
(00:20:48)
  • Key Takeaway: China’s export controls on rare earths necessitate US government intervention, like price floors or strategic reserves, to prevent domestic industry collapse from Chinese price dumping.
  • Summary: China imposed export controls on 12 critical rare earth minerals, escalating trade tensions ahead of a Trump-Xi meeting. The hosts debated the US response, with some favoring government price floors to incentivize domestic reshoring (like MP Materials) against China’s subsidized market manipulation. China’s long-term mercantilist strategy, dating back to Hu Jintao, aimed to dominate strategic inputs like rare earths, batteries, and pharmaceutical APIs.
US-China Geopolitical Context
Copied to clipboard!
(00:41:03)
  • Key Takeaway: China views its current economic rise as a ‘reascending power’ returning to its historical dominance, not merely a rising power challenging the US order.
  • Summary: The discussion framed China’s actions through the lens of historical dominance, noting it held the world’s largest GDP for 70% of the years since 1500. This perspective informs their aggressive state-directed capital allocation via five-year plans, which creates national champions in strategic sectors. The failure of the ‘End of History’ theory—that wealth would lead China to liberal democracy—resulted in a powerful geopolitical competitor.
AI Data Center PR Crisis
Copied to clipboard!
(00:53:00)
  • Key Takeaway: Local opposition over rising electricity prices, water consumption, and noise pollution is causing major tech companies to pull planned US data center investments.
  • Summary: Google, Microsoft, and Amazon have recently withdrawn data center plans in locations like Indianapolis and Wisconsin due to anticipated local council rejection. This resistance stems from community concerns over increased utility costs and water usage, indicating that AI’s tangible benefits are not yet outweighing perceived local burdens. The hosts argue that clear federal regulation and visible economic benefits are necessary to gain public acceptance for AI infrastructure.
AI Public Perception & Benefits
Copied to clipboard!
(00:56:50)
  • Key Takeaway: Tangible economic benefits must be demonstrated to the public to counter the narrative that AI will eliminate jobs while simultaneously increasing costs like electricity.
  • Summary: The public does not understand what a ’token’ is, but they understand that AI tokens are not jobs. AI adoption requires showing clear, tangible economic benefits, especially in regions feeling left behind, like Wisconsin and Indiana. The current situation risks public backlash if AI adoption leads to higher electricity prices without visible abundance.
AI Cash Flow for Community Support
Copied to clipboard!
(00:57:50)
  • Key Takeaway: Big tech companies should leverage their free cash flow to cushion local impacts of AI build-out, such as subsidizing electricity costs or funding local energy solutions.
  • Summary: Free cash flow from large companies can act as a cushion to pay for higher electricity tariffs or directly subsidize local residents’ utility bills. Ideas include funding solar and storage solutions for communities near data centers. This market intervention is necessary for these companies to gain public support for their expansion.
AI GDP Impact vs. Job Loss Narrative
Copied to clipboard!
(00:59:47)
  • Key Takeaway: AI contributed 40% of the recent 3.8% quarterly economic growth, making it the difference between 4% growth and modest 2% growth, challenging the media’s focus on theoretical job loss.
  • Summary: The job loss narrative is largely theoretical and anecdotal, contrasting with the measurable positive impact AI has on GDP growth rates. Historically, technological revolutions shift rote jobs to more sophisticated, higher-paying roles, as seen by the massive drop in agricultural employment since 1900. Humans remain essential for end-to-end tasks like prompting, validation, and setting objectives, making AI synergistic rather than purely replacement technology.
Recruiting Cycle Precedes Job Elimination
Copied to clipboard!
(01:03:11)
  • Key Takeaway: In technological shifts, a recruiting cycle for new, higher-paying jobs precedes the elimination of old jobs, leading to a net positive income improvement for workers who transition.
  • Summary: Instead of focusing on job loss, the focus should be on the recruitment cycle where employers offer higher pay for new capacity needed in emerging industries like manufacturing build-outs driven by AI and robotics. Workers will naturally move to jobs offering double the pay for more interesting work, causing demand for old jobs to decline years later after the transition is complete. This historical pattern, exemplified by the Model T displacing buggy drivers, is hard for the media to capture because new industries are unseen in the short term.
Legitimate vs. Illegitimate AI Skepticism
Copied to clipboard!
(01:08:43)
  • Key Takeaway: Legitimate skepticism regarding AI centers on infrastructure needs like electricity generation, while illegitimate fear stems from media-driven narratives about mass job loss or Terminator-style AI takeover.
  • Summary: The need to stand up more electricity power generation is a legitimate concern that could cause a backlash if residential rates rise due to data center demand. Illegitimate fears are fueled by unproven studies predicting massive job loss or ‘doomer’ scenarios, leading to knee-jerk legislative reactions. The industry must fix the factual issues (like power) and improve communication to prevent essential projects from being canceled.
Hyperscalers’ Role and Industry Spokesmen
Copied to clipboard!
(01:10:30)
  • Key Takeaway: Hyperscalers must deploy their currently unvalued free cash flow to solve local issues (electricity, noise) and the industry requires better, more eloquent spokesmen to articulate a positive vision.
  • Summary: CEOs should use their massive free cash flow, which public markets currently do not value, to invest in local infrastructure and address community concerns like electricity prices. The industry also suffers from poor communication, exemplified by focusing on ’navel-gazing technologies’ or getting caught up in controversies like OpenAI’s erotica push. Better spokesmen are needed to articulate a reasonable, buy-in-worthy vision for AI’s future.
Tech Headcount Stagnation Evidence
Copied to clipboard!
(01:12:23)
  • Key Takeaway: Data from major tech companies (Alphabet, Meta, Uber, Amazon) shows static or decreasing headcount despite profit increases, suggesting job displacement may be happening faster than new job creation.
  • Summary: The speaker’s concern about displacement is based on observing that major tech companies, which deploy AI fastest, are not growing their employee base; for example, Alphabet’s headcount decreased from its 2022 peak while profits rose. This trend of static team sizes, coupled with increasing unemployment among new computer science graduates, suggests new jobs may not materialize quickly enough to offset losses in knowledge work and automation.