<div dir="ltr"><div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr">On Fri, 4 Sept 2026 at 10:37, John Clark via extropy-chat <<a href="mailto:extropy-chat@lists.extropy.org">extropy-chat@lists.extropy.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div><font size="4" face="tahoma, sans-serif"><b>Bernie Sanders has proposed a bill that would ban superintelligence research, give a "corporate death penalty" to any company that violates the ban, and put in prison for 20 years any individual who engages in superintelligence research.  Chinese leaders must be smiling. </b></font></div><div></div><div><div style="color:rgb(80,0,80)"><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default"><snip></div></div></div></div>
_______________________________________________<br></blockquote><div><br></div><div><br></div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default">I asked Qwen AI to review two articles about this proposal:</div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default">Joe Rogero (from MIRI-(If Anyone Builds it, Everyone dies) and Gary Marcus.</div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default">MIRI want a pause to AI development, whiie Gary worries about China continuing development.</div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default">I thought Qwen did a good review of the problem.</div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default"></div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default">BillK</div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default"><br></div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default">Question: </div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default">Can you review and discuss these two articles? - <a href="https://aistop.watch/p/shutting-it-down?open=false#%C2%A7sanders-and-casar-propose-superintelligence-ban">https://aistop.watch/p/shutting-it-down?open=false#%C2%A7sanders-and-casar-propose-superintelligence-ban</a> and <a href="https://garymarcus.substack.com/p/the-new-sanders-casar-ban-artificial">https://garymarcus.substack.com/p/the-new-sanders-casar-ban-artificial</a></div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default"><br></div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default"><br></div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default">Qwen3.8-Max Thinking -</div><div style="font-family:arial,sans-serif;font-size:small;color:rgb(0,0,0)" class="gmail_default"> <div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">Both articles, published on September 3, 2026, address a major breaking event in AI policy: the introduction of the </span><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Ban Artificial Superintelligence Act</span></strong><span class="gmail-qwen-markdown-text"> by Senator Bernie Sanders and Representative Greg Casar. The proposed legislation seeks to outlaw the development and deployment of AI systems that match or exceed human cognitive performance, halt "advanced AI development" until federal standards are met, establish a cabinet-level AI regulatory agency, and push for a global treaty to enforce these rules. </span></div><div class="gmail-qwen-markdown-space"></div><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">While both authors recognize the severe dangers posed by advanced AI and applaud the government's urgency, they reach dramatically different conclusions about whether the legislation is a necessary lifeline for humanity or a naive policy misstep.</span></div><div class="gmail-qwen-markdown-space"></div><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">Here is a review and discussion of both perspectives.</span></div><div class="gmail-qwen-markdown-space"></div><h3 class="gmail-qwen-markdown-heading" dir="ltr"><span class="gmail-qwen-markdown-text">1. AI StopWatch: "Shutting it down" (by Joe Rogero)</span></h3><div class="gmail-qwen-markdown-paragraph" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Stance:</span></strong><span class="gmail-qwen-markdown-text"> Enthusiastically supportive. 
Joe Rogero views the Sanders-Casar bill as a "major policy breakthrough" and a "massive breakthrough in sane AI policy." He writes that the news moved him almost to tears, framing the legislation as a vital step toward ensuring human survival.</span></div><div class="gmail-qwen-markdown-space"></div><div class="gmail-qwen-markdown-paragraph" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Key Arguments:</span></strong></div><ul class="gmail-qwen-markdown-list" dir="ltr"><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">The Precautionary Principle:</span></strong><span class="gmail-qwen-markdown-text"> Rogero argues that it is common sense not to build something smarter than ourselves that we cannot control. He strongly supports the hard ban on superintelligence and AI systems capable of planning the "disempowerment of humanity."</span></li><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Unilateral Leadership:</span></strong><span class="gmail-qwen-markdown-text"> While acknowledging that a unilateral U.S. halt isn't a permanent fix without international cooperation, he views a bold U.S. stand as the best way to force countries like China to treat the issue with the gravity it deserves, akin to nuclear non-proliferation treaties.</span></li><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Accepting Bluntness:</span></strong><span class="gmail-qwen-markdown-text"> He admits the definition of "advanced AI development" might be overly broad and the international policy is currently thin, but he believes that erring heavily on the side of caution is the only rational choice when no serious regulations currently exist.</span></li></ul><div class="gmail-qwen-markdown-space"></div><div class="gmail-qwen-markdown-paragraph" dir="ltr"><em class="gmail-qwen-markdown-em" dir="ltr"><span class="gmail-qwen-markdown-text">(Note: The broader issue of AI StopWatch also covers the bipartisan "Stop Rogue AI Act," which mandates technical security standards for AI agents, a survey showing that 111 national security experts are deeply concerned about AI extinction risks, and NYC's one-year ban on classroom AI).</span></em></div><div class="gmail-qwen-markdown-space"></div><div class="gmail-qwen-markdown-hr"><hr></div><div class="gmail-qwen-markdown-space"></div><h3 class="gmail-qwen-markdown-heading" dir="ltr"><span class="gmail-qwen-markdown-text">2. Gary Marcus: "The new Sanders-Casar Ban Artificial Superintelligence Act – and why I oppose it"</span></h3><div class="gmail-qwen-markdown-paragraph" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Stance:</span></strong><span class="gmail-qwen-markdown-text"> Pragmatically opposed to the bill's specifics, while supportive of its underlying regulatory goals.
Cognitive psychologist and AI critic Gary Marcus agrees with the need for a temporary pause and a federal AI agency (a stance he has held since 2023). However, he strongly opposes the bill as written, arguing it goes "too far."</span></div><div class="gmail-qwen-markdown-space"></div><div class="gmail-qwen-markdown-paragraph" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Key Arguments:</span></strong></div><ul class="gmail-qwen-markdown-list" dir="ltr"><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Permanence and Geopolitics:</span></strong><span class="gmail-qwen-markdown-text"> Marcus argues that a permanent, unilateral ban on superhuman AI is too broad. He warns it is a "guarantee of leaving the US behind," creating a scenario where authoritarian regimes will simply ignore the ban and gain a massive strategic advantage. </span></li><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Benchmarking Naivety:</span></strong><span class="gmail-qwen-markdown-text"> He notes that the bill relies on complex and often flawed benchmarks to define "superintelligence," making a permanent legal ban difficult to enforce accurately.</span></li><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Sci-Fi vs. Reality:</span></strong><span class="gmail-qwen-markdown-text"> Aligning with virologist Dr. Angela Rasmussen, Marcus criticizes the bill for focusing entirely on hypothetical, existential "sci-fi" risks (like AI disempowering humanity) while ignoring the immediate, real-world harms AI is causing right now (such as cyber vulnerabilities, disinformation, and algorithmic bias).</span></li><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">A Conditional Alternative:</span></strong><span class="gmail-qwen-markdown-text"> Instead of a permanent ban, Marcus proposes borrowing language from physicist Anthony Aguirre: a conditional halt. Development of highly general and autonomous AI would only be permitted if companies can demonstrate with "high assurance" to an independent regulatory authority that sufficient alignment and safety controls exist. This shifts the burden of proof onto the tech companies.</span></li></ul><div class="gmail-qwen-markdown-space"></div><div class="gmail-qwen-markdown-hr"><hr></div><div class="gmail-qwen-markdown-space"></div><h3 class="gmail-qwen-markdown-heading" dir="ltr"><span class="gmail-qwen-markdown-text">Discussion & Synthesis</span></h3><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">These two articles perfectly encapsulate the current ideological divide in AI safety and policy discourse. They share several foundational beliefs but diverge sharply on execution.</span></div><div class="gmail-qwen-markdown-space"></div><h4 class="gmail-qwen-markdown-heading" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">The Common Ground</span></strong></h4><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">Both authors agree on three vital points:</span></div><ol class="gmail-qwen-markdown-list" start="1" dir="ltr"><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">The Era of Laissez-Faire is Over:</span></strong><span class="gmail-qwen-markdown-text"> Both agree that private tech companies can no longer be trusted to self-regulate frontier AI models.</span></li><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">The Need for a Federal Agency:</span></strong><span class="gmail-qwen-markdown-text"> Both strongly support the creation of a dedicated, cabinet-level federal body to oversee AI, enforce standards, and monitor the lifecycle of frontier models.</span></li><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">The Danger of Proliferation:</span></strong><span class="gmail-qwen-markdown-text"> Both are deeply concerned by recent events (such as the Hugging Face attack mentioned in both articles) and the rapid release of open-weights models with dangerous cyber capabilities.</span></li></ol><div class="gmail-qwen-markdown-space"></div><h4 class="gmail-qwen-markdown-heading" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">The Geopolitical Prisoner's Dilemma</span></strong></h4><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">The most fascinating divergence is how the two handle the "China factor." </span></div><ul class="gmail-qwen-markdown-list" dir="ltr"><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">AI StopWatch</span></strong><span class="gmail-qwen-markdown-text"> hopes for a "moral contagion" effect—that a brave, unilateral pause by the United States will shame or diplomatically force rival nations into a global non-proliferation treaty. </span></li><li class="gmail-qwen-markdown-list-item" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Marcus</span></strong><span class="gmail-qwen-markdown-text"> views this through the lens of realpolitik. He sees a unilateral ban as unilateral disarmament. If the U.S. makes advanced AI research a criminal offense punishable by 20 years in prison, top talent will either flee to jurisdictions without such bans or work in secret, handing a massive geopolitical advantage to adversarial nations.</span></li></ul><div class="gmail-qwen-markdown-space"></div><h4 class="gmail-qwen-markdown-heading" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">The Burden of Proof: Hard Bans vs. High Bars</span></strong></h4><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">Marcus’s proposed alternative (borrowed from Aguirre) offers a highly pragmatic middle ground that the Sanders-Casar bill lacks. Instead of Congress trying to permanently outlaw a moving target (superintelligence), Marcus suggests establishing a regulatory "high bar." If a company wants to build a system smarter than humans, </span><em class="gmail-qwen-markdown-em" dir="ltr"><span class="gmail-qwen-markdown-text">they</span></em><span class="gmail-qwen-markdown-text"> must prove to a federal agency that it is safe and aligned. If they cannot prove it, development halts. This incentivizes safety research and leaves the door open for beneficial innovation, whereas AI StopWatch embraces a permanent, blunt instrument to guarantee survival.</span></div><div class="gmail-qwen-markdown-space"></div><h4 class="gmail-qwen-markdown-heading" dir="ltr"><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">X-Risk vs. Immediate Harms</span></strong></h4><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">Finally, the articles highlight a philosophical debate about what we are actually trying to regulate. The Sanders-Casar bill—and AI StopWatch’s support of it—is rooted in </span><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">Existential Risk (X-Risk)</span></strong><span class="gmail-qwen-markdown-text">. The primary fear is a rogue, superintelligent agent destroying human civilization. Marcus, however, represents the </span><strong class="gmail-qwen-markdown-strong" dir="ltr"><span class="gmail-qwen-markdown-text">AI Ethics/Immediate Harm</span></strong><span class="gmail-qwen-markdown-text"> camp. He argues that by drafting sweeping legislation around hypothetical "Terminator" scenarios, lawmakers are failing to address the very real algorithmic damages happening today.</span></div><div class="gmail-qwen-markdown-space"></div><h3 class="gmail-qwen-markdown-heading" dir="ltr"><span class="gmail-qwen-markdown-text">Conclusion</span></h3><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">The Sanders-Casar "Ban Artificial Superintelligence Act" clearly represents a watershed moment in how Washington views AI. AI StopWatch is right to celebrate the fact that the existential threat of AI is finally being treated with the legislative gravity it warrants. However, Gary Marcus’s critique highlights the practical flaws of reactionary legislation. Moving forward, the most effective policy will likely need to marry the aggressive urgency of the Sanders-Casar bill with the conditional, burden-shifting regulatory frameworks that Marcus advocates.</span></div><div class="gmail-qwen-markdown-paragraph" dir="ltr"><span class="gmail-qwen-markdown-text">--------------------------------</span></div></div></div></div>