<?xml version="1.0" encoding="UTF-8"?><feed
	xmlns="http://www.w3.org/2005/Atom"
	xmlns:thr="http://purl.org/syndication/thread/1.0"
	xml:lang="en-US"
	>
	<title type="text">Haydn Belfield | Vox</title>
	<subtitle type="text">Our world has too much noise and too little context. Vox helps you understand what matters.</subtitle>

	<updated>2024-03-14T19:30:47+00:00</updated>

	<link rel="alternate" type="text/html" href="https://www.vox.com/author/haydn-belfield" />
	<id>https://www.vox.com/authors/haydn-belfield/rss</id>
	<link rel="self" type="application/atom+xml" href="https://www.vox.com/authors/haydn-belfield/rss" />

	<icon>https://platform.vox.com/wp-content/uploads/sites/2/2024/08/vox_logo_rss_light_mode.png?w=150&amp;h=100&amp;crop=1</icon>
		<entry>
			
			<author>
				<name>Haydn Belfield</name>
			</author>
			
			<title type="html"><![CDATA[I’m an expert in the end of the world. The Oscar-winning Oppenheimer made me cry in terror.]]></title>
			<link rel="alternate" type="text/html" href="https://www.vox.com/future-perfect/24093819/oppenheimer-oscar-winner-nuclear-war-christopher-nolan-los-alamos-manhattan-project-academy-awards" />
			<id>https://www.vox.com/future-perfect/24093819/oppenheimer-oscar-winner-nuclear-war-christopher-nolan-los-alamos-manhattan-project-academy-awards</id>
			<updated>2024-03-14T15:30:47-04:00</updated>
			<published>2024-03-11T08:00:00-04:00</published>
			<category scheme="https://www.vox.com" term="Awards Shows" /><category scheme="https://www.vox.com" term="Culture" /><category scheme="https://www.vox.com" term="Defense &amp; Security" /><category scheme="https://www.vox.com" term="Future Perfect" /><category scheme="https://www.vox.com" term="Oscars" /><category scheme="https://www.vox.com" term="Politics" />
							<summary type="html"><![CDATA[My wife and I went to see Oppenheimer on opening weekend in July, and I wore my best Los Alamos-themed costume. We opted for the biggest screen we could &#8212; and as the Trinity test explosion kept on rising higher and higher endlessly upward, I was glad we had sprung for the more expensive IMAX [&#8230;]]]></summary>
			
							<content type="html">
											<![CDATA[

						
<figure>

<img alt="" data-caption="The mushroom cloud produced by the first explosion by the Americans of a hydrogen bomb at Eniwetok Atoll in the South Pacific, in 1952. | SSPL via Getty Images" data-portal-copyright="SSPL via Getty Images" data-has-syndication-rights="1" src="https://platform.vox.com/wp-content/uploads/sites/2/chorus/uploads/chorus_asset/file/25325222/GettyImages_90739961.jpg?quality=90&#038;strip=all&#038;crop=0,0,100,100" />
	<figcaption>
	The mushroom cloud produced by the first explosion by the Americans of a hydrogen bomb at Eniwetok Atoll in the South Pacific, in 1952. | SSPL via Getty Images	</figcaption>
</figure>
<p>My wife and I went to see <em>Oppenheimer</em> on opening weekend in July, and I wore my best Los Alamos-themed costume. We opted for the biggest screen we could &mdash; and as the Trinity test explosion kept on rising higher and higher endlessly upward, I was glad we had sprung for the more expensive IMAX ticket for the sheer spectacle. I could guess, even then, that this film was destined to dominate the Oscars, where this past weekend, it <a href="https://variety.com/2024/film/awards/oscar-winners-2024-list-1235932341/">won</a> Best Picture and Best Director, among seven Academy Awards altogether.</p>

<p>But then the final scene rolled around &mdash; showing Los Alamos lab director J. Robert Oppenheimer&rsquo;s vision of rows and rows of intercontinental ballistic missiles (ICBMs), the rockets zooming upward &mdash; and then the Earth from space with mushroom clouds rising above the clouds and fire spreading across the globe</p>

<p>As the strings rose to their final jarring crescendo,<em> </em>I found myself unexpectedly and uncontrollably bawling my eyes out. I sat in my chair at the London Science Museum cinema, my shoulders shaking, speechless with tears, trying to gasp out to my wife between sobs what on earth was going on with me. &ldquo;It&rsquo;s all still real,&rdquo; I told her. &ldquo;The weapons are still there. Every 12 minutes they could kill us and everyone we love. The entire Northern Hemisphere would be gone. We&rsquo;d all starve to death. Five billion people could die.&rdquo;</p>
<img src="https://platform.vox.com/wp-content/uploads/sites/2/chorus/uploads/chorus_asset/file/25325204/Screenshot_2024_03_08_at_11.10.44_AM.png?quality=90&#038;strip=all&#038;crop=0,0,100,100" alt="Oppenheimer wearing goggles, his face lit from the blast of the bomb as he watches through a circular window." title="Oppenheimer wearing goggles, his face lit from the blast of the bomb as he watches through a circular window." data-has-syndication-rights="1" data-caption="Still from &lt;em&gt;Oppenheimer.&lt;/em&gt; | Courtesy of Universal" data-portal-copyright="Courtesy of Universal" />
<p>I&rsquo;ve worked at the University of Cambridge&rsquo;s <a href="https://www.cser.ac.uk/">Centre for the Study of Existential Risk</a> for the past seven years, where we study risks that could lead to human extinction or civilizational collapse &mdash; and how to prevent them. Most of our conversations at the pub across the road from the office are about <a href="https://www.vox.com/climate" data-source="encore">climate change</a>, engineered pandemics, <a href="https://www.vox.com/2023/4/28/23702644/artificial-intelligence-machine-learning-technology" data-source="encore">artificial intelligence</a>, and yes, nuclear weapons.&nbsp;</p>

<p>The most common question I get asked is &ldquo;doesn&rsquo;t working on this make you all depressed?&rdquo; I normally answer with a joke, but the truth is that while those conversations are normally academic, abstract, distanced, and intellectual, <em>Oppenheimer</em> was hitting me emotionally.</p>

<p>The USA and <a href="https://www.vox.com/russia" data-source="encore">Russia</a> each have <a href="https://www.armscontrol.org/factsheets/Nuclearweaponswhohaswhat">around 1,500 warheads deployed</a> &mdash; ready to launch. Each of these is a hydrogen (or thermonuclear) bomb. Many of these, which are 20 times more destructive than those produced by Oppenheimer&rsquo;s Manhattan Project and which destroyed Hiroshima, sit on the top of ICBMs in land-based silos &mdash; in the US in <a href="https://thebulletin.org/2021/02/why-is-america-getting-a-new-100-billion-nuclear-weapon/">Montana, North Dakota, Wyoming, Nebraska, and Colorado</a> &mdash; and are ready for a &ldquo;launch on warning.&rdquo;&nbsp;</p>

<p>What this means is that in the event of the detection of a possible incoming nuclear strike, the president &mdash; while they are being rushed to the White House bunker &mdash; has perhaps 12 minutes to decide whether to launch all of these missiles. Unlike nuclear missiles on submarines, which are designed to evade a first strike, this is a use-it-or-lose-it situation: the silos will get blown up at the end of those 12 minutes. What could happen next is unimaginable &mdash; save for the scientists who have imagined it.</p>

<p>You may remember the <a href="https://www.vox.com/future-perfect/2018/10/19/17873822/nuclear-war-weapons-bombs-how-kill">theory of &ldquo;nuclear winter&rdquo;</a> from the 1980s. <a href="https://www.smithsonianmag.com/science-nature/when-carl-sagan-warned-world-about-nuclear-winter-180967198/">Proposed</a> by Carl Sagan and other US and Soviet scientists, it suggested that the smoke from burning cities could rise up into the atmosphere and block out the sun. Scientists <a href="https://adamtooze.substack.com/p/chartbook-230-burning-hamburg-1943">learned</a> during World War II that if conditions are right, a burning city after a bombing can become a firestorm, a freak weather phenomenon where the heat of the fire sucks in air from surrounding areas in a great wind, fueling the fire like a bellows.&nbsp;</p>

<p>The nuclear winter theory takes this further: Destroying cities with a hydrogen bomb might also lead to firestorms. That great wind could carry millions of metric tons of black carbon soot from these burnt cities far up into the stratosphere. Once up there, up above rainclouds, that soot might stay suspended for up to a decade. If so, it would act like a massive volcanic eruption or asteroid impact, blotting out the sun and reducing crop yields.&nbsp;</p>

<p>But scientists have carried out research with today&rsquo;s much more powerful physics simulations and climate models. Their conclusion?&nbsp;</p>

<p>In a 2022 <em>Nature Food </em><a href="https://www.nature.com/articles/s43016-022-00573-0">paper</a>, researchers led by Lili Xia find that a full-scale nuclear war could reduce crop yields between 50 percent and 90 percent and kill 2 to 5 billion people. At the top end, this would mean north of&nbsp;90 percent of the populations of the US, Europe, Russia, and <a href="https://www.vox.com/china" data-source="encore">China</a> starving to death. Because of global climate patterns much of this would be concentrated in the Northern Hemisphere &mdash; unfortunately where I and most of my family and friends live.</p>

<p>This is not a closed debate: The Future of Life Institute, which also studies existential risks, <a href="https://futureoflife.org/grant-program/nuclear-war-research/">recently awarded</a> about $4 million to 10 university groups to explore this in more detail. But it is a scenario that has to be taken seriously. And it was this scenario that left me sobbing and gasping at <em>Oppenheimer</em>&rsquo;s end.</p>
<div class="youtube-embed"><iframe title="How would a nuclear war between Russia and the US affect you personally?" src="https://www.youtube.com/embed/-xthzy1PxTA?rel=0" allowfullscreen allow="accelerometer *; clipboard-write *; encrypted-media *; gyroscope *; picture-in-picture *; web-share *;"></iframe></div><h2 class="wp-block-heading"><em>Oppenheimer </em>as a warning</h2>
<p>Since <em>Oppenheimer</em>&rsquo;s release, the film&rsquo;s director, Christopher Nolan, has actively drawn attention to the continued dangers of nuclear war. Nolan spoke about just this subject at the Bulletin of the Atomic Scientists&#8217; annual <a href="https://thebulletin.org/support-the-bulletin/conversations-before-midnight-2023/">Conversations Before Midnight</a> just before the main event of the evening, joint headliner yours truly (my co-recipient of the Rieser Award, Christian Ruhl, had talked me out of making too many jokes about Nolan). In his BAFTA acceptance <a href="https://youtu.be/iYlNtbpn7AQ?si=eBTxYffIbeW7i79Z">speech</a> last month, Nolan thanked his cast (standard), his crew (classy), and his producers (shrewd), but ended by thanking a different group. He said:</p>
<blockquote class="wp-block-quote has-text-align-none is-layout-flow wp-block-quote-is-layout-flow">
<p>Our film ends on what I think is a dramatically necessary note of despair. But in the real world there are all kinds of individuals and organizations that have fought long and hard to reduce the number of nuclear weapons in the world. Since its peak in 1967, they&rsquo;ve done it by almost 90 percent. Of late that&rsquo;s gone the wrong way. And so in accepting this I want to acknowledge their efforts and point out that they show the necessity and the potential of efforts for peace.</p>
</blockquote>
<p>In 1991, as the Cold War came to a sudden end, the Doomsday Clock, <a href="https://news.uchicago.edu/explainer/what-is-the-doomsday-clock">designed</a> by the Bulletin more than 70 years ago to draw attention to the threat of nuclear holocaust, was set back to 17 minutes to midnight. The plan for nuclear arms negotiators from there on was clear: keep on negotiating bilaterally to get the US and Russian stockpiles from their<a href="https://ourworldindata.org/nuclear-weapons"> all-time high</a> of over 60,000 combined down to around 200 warheads each, on par with most other nuclear states, then have multilateral negotiations to get the numbers as close to Global Zero as we can while preserving the possibility of deterrence. But as Nolan noted with marvelous English understatement that night at the BAFTAs: &ldquo;Of late that&rsquo;s gone the wrong way.&rdquo;&nbsp;</p>

<p>From the Anti-Ballistic Missile Treaty to the Intermediate-Range Nuclear Forces Treaty to the Open Skies Treaty, the key Cold War pacts that constrained Russia and the US <a href="https://time.com/6334258/putin-nuclear-arms-control/">have been torn up</a> in the 21st century. The only nuclear treaty left, the New START Treaty, <a href="https://www.state.gov/new-start/#:~:text=Treaty%20Duration%3A%20The%20treaty's%20original,force%20through%20February%204%2C%202026.">expires automatically</a> on February 5, 2026. And last month, Russian President Vladimir Putin <a href="https://www.npr.org/2023/02/21/1158463688/putin-tells-russian-parliament-the-west-is-fighting-to-dismember-russia">withdrew</a> from New START&rsquo;s inspections regime.</p>

<p>Nuclear risk experts <a href="https://armscontrolcenter.org/are-nuclear-armed-nations-entering-a-new-arms-race-in-2024-experts-weigh-in/">warn</a> that we are entering a new arms race. China is building hundreds of new <a href="https://www.armscontrol.org/act/2021-09/news/new-chinese-missile-silo-fields-discovered">silos</a> in its northern deserts and could be increasing its number of operational warheads <a href="https://media.defense.gov/2023/Oct/19/2003323409/-1/-1/1/2023-MILITARY-AND-SECURITY-DEVELOPMENTS-INVOLVING-THE-PEOPLES-REPUBLIC-OF-CHINA.PDF">from around 500</a> to <a href="https://www.npr.org/2023/10/19/1207156597/new-pentagon-report-claims-china-now-has-over-500-operational-nuclear-warheads#:~:text=The%20report%20projects%20that%20China,of%20the%20continental%20United%20States.">around 1,000</a>. All nuclear weapon states are in the middle of nuclear &ldquo;<a href="https://www.defense.gov/News/News-Stories/Article/Article/3505989/pentagon-tackling-nuclear-modernization-with-proactive-integrated-approach/">modernization</a>&rdquo;: replacing old warheads, ICBMs, bombers, and submarines and making tweaks like better <a href="https://publicintegrity.org/national-security/future-of-warfare/nuclear-weapon-arsenal-more-destructive-risky/">fuzes</a> to control blast timing in ways that will make the warheads more damaging. <a href="https://www.kcl.ac.uk/news/new-study-analyses-impact-of-emerging-technology-on-nuclear-safety">New technologies</a> such as highly maneuverable hypersonic missiles or <a href="https://www.europeanleadershipnetwork.org/report/ai-and-nuclear-command-control-and-communications-p5-perspectives/">integrating AI</a> into nuclear decision-making could threaten strategic stability.</p>

<p>Meanwhile, the nuclear risk reduction community is <a href="https://www.politico.com/news/magazine/2023/07/28/nuclear-experts-russia-war-00108438">not doing well</a>. The biggest funder in the field, the MacArthur Foundation, <a href="https://www.politico.com/news/2021/07/19/washington-arms-controllers-nuclear-weapons-500126">pulled out</a> in 2021, declaring that its &ldquo;Big Bet&rdquo; had not paid off. MacArthur had provided <a href="https://www.vox.com/2022/3/17/22976981/nuclear-war-russia-ukraine-funding-macarthur-existential-risk-effective-altruism-carnegie">around half</a> of all the non-government funding worldwide on nuclear policy. Other funders such as <a href="https://www.longview.org/fund/nuclear-weapons-policy-fund/">Longview</a> have stepped up, but have not been able to plug this gaping hole.&nbsp;</p>

<p>One major benefit of Nolan&rsquo;s work and <em>Oppenheimer</em>, especially after its many Oscar wins, is that it will continue to draw more attention to the bizarrely neglected area of nuclear risk.</p>
<img src="https://platform.vox.com/wp-content/uploads/sites/2/chorus/uploads/chorus_asset/file/25325226/Screenshot_2024_03_08_at_11.19.47_AM.png?quality=90&#038;strip=all&#038;crop=0,0,100,100" alt="The silhouette of a man working on the atomic bomb." title="The silhouette of a man working on the atomic bomb." data-has-syndication-rights="1" data-caption="Still from &lt;em&gt;Oppenheimer&lt;/em&gt; | Courtesy of Universal Pictures" data-portal-copyright="Courtesy of Universal Pictures" /><h2 class="wp-block-heading">What <em>Oppenheimer </em>gets wrong about the Manhattan Project</h2>
<p><em>Oppenheimer</em> is a film that affected me so strongly and has done great good. Will I really be so churlish as to criticize it? Yes, I will. Not for its depiction of Florence Pugh&rsquo;s <a href="https://www.digitalspy.com/movies/a44628614/oppenheimer-florence-pugh-jean-tatlock-real-story/">Jean Tatlock</a>, the choice to not include Japanese <a href="https://ew.com/movies/christopher-nolan-responds-to-criticism-for-not-showing-hiroshima-bombing-in-oppenheimer/">perspectives</a>, or for it being &ldquo;<a href="https://youtu.be/4jqsVoCS17I?si=SyXr90ryMYqzXIpN">too long</a>,&rdquo; but about two subjects it didn&rsquo;t touch on.</p>

<p>First, the true story of the production of the bomb.</p>

<p>Like most films and TV shows about the Manhattan Project, <em>Oppenheimer</em> focuses far too much on Oppenheimer and his Los Alamos scientists. The film never shows <a href="https://www.newschannel5.com/news/the-oppenheimer-film-is-now-playing-heres-the-role-tennessee-played-in-this-history">Oak Ridge</a>, Tennessee, or <a href="https://www.seattletimes.com/opinion/oppenheimer-overlooks-a-critical-place-in-the-dawn-of-the-nuclear-age/">Hanford</a>, Washington. But these production plants for enriched uranium and plutonium respectively were responsible for more than 80 percent of the <a href="https://www.brookings.edu/the-costs-of-the-manhattan-project/">budget</a> of the Manhattan Project and most of its staff. It is these locations and people that actually produced the bomb. The Nazis had smart scientists. What they didn&rsquo;t have was the US&rsquo;s vast industrial and financial resources. Yet in the film, these remarkable achievements are reduced to marbles that Cillian Murphy&rsquo;s Oppenheimer places in a bowl to track the production of the nuclear fuel.&nbsp;</p>

<p>It was these industrial workers who built the bomb. But after the war, it was the privileged genius scientists who were <a href="https://content.time.com/time/covers/0,16641,19481108,00.html">on the front cover</a> of Time magazine, and it is their stories that shaped how we remember the Manhattan Project.&nbsp;</p>

<p>This misremembering has important modern implications. It is production that determines &ldquo;<a href="https://www.washingtoninstitute.org/policy-analysis/iranian-nuclear-breakout-what-it-and-how-calculate-it">breakout time</a>&rdquo; &mdash; how long until a state can have enough nuclear material for a bomb. Arms control regimes like the <a href="https://www.vox.com/iran" data-source="encore">Iran</a> deal or the Non-Proliferation Treaty focus on monitoring and controlling production facilities, as did the 1981 Israeli air strike on Iraq or the 2010 Stuxnet cyberattack on Iran. It is production that should be our focus.</p>

<p>Second, the film <a href="https://assets.scriptslug.com/live/pdf/scripts/oppenheimer-2023.pdf">presents</a> the US as desperately behind the Nazis who have, according to Oppenheimer, an &ldquo;18-month head start&rdquo; in the nuclear arms race, and suggests that<strong> </strong>&ldquo;in a straight race, the Germans win.&rdquo; The Nazis only lose in the film because German physicist Werner Heisenberg &ldquo;took a wrong turn&rdquo; by choosing heavy water rather than graphite as a moderator, as revealed in the film by Kenneth Branagh&rsquo;s Niels Bohr in Christmas 1943.&nbsp;</p>

<p>But we&rsquo;ve <a href="https://www.cambridge.org/core/books/german-national-socialism-and-the-quest-for-nuclear-power-193949/389AE0447CBA4D207983B0F98E82E20E">known for decades</a> that none of this was true: Hitler had decided against a serious program <a href="https://www.historyextra.com/period/second-world-war/german-nuclear-weapons-program-nazi-atom-bomb/">at a presentation</a> in June 1942.</p>

<p>The Nazis had indeed explored launching their own Manhattan Project: the &ldquo;Uranverein&rdquo; <a href="http://large.stanford.edu/courses/2018/ph241/turchetta1/">nuclear weapons program</a> led by Heisenberg. But Heisenberg and Germany&rsquo;s military planners predicted that it would be a major investment that would only pay off in two to three years. The Nazis could not make such a massive investment of people and raw materials like steel. They had a much <a href="https://www.penguin.co.uk/books/53768/the-wages-of-destruction-by-adam-tooze/9780141003481">smaller economy</a>, huge shortages, and a greater need for shells and tanks. And they couldn&rsquo;t wait for 1944 or &rsquo;45 &mdash; the Nazis needed a breakthrough in the Eastern Front right then.&nbsp;</p>

<p>This was not a technical error, but a strategic decision. The US was at no risk of losing the race, as no other great power &mdash; the Soviet Union, Imperial Japan, or Nazis &mdash; rushed toward the bomb during the war. Moreover in a straight race, there was no possible way for any of them to compete with the vast industrial and financial might of the US.&nbsp;</p>

<p>The film could have explored this tragic mistake while keeping its laser focus on Oppenheimer. The official historian of the Manhattan Project stated that at the end of 1943 Oppenheimer was explicitly told by Gen. Leslie Groves, the director of the atom bomb program, that the Nazis had abandoned their early program &mdash; and Oppenheimer just shrugged.&nbsp;</p>
<h2 class="wp-block-heading">The Manhattan Project’s lost conscience</h2>
<p>Most egregiously for me, the film doesn&rsquo;t show Joseph Rotblat. Rotblat was the only scientist to <a href="https://ahf.nuclearmuseum.org/voices/oral-histories/joseph-rotblats-interview/#:~:text=Rotblat%20left%20the%20Manhattan%20Project,and%20professional%20relationships%20with%20J.">resign</a> from the project, even though he was a Polish refugee whose wife Tola Gryn was murdered in a Nazi concentration camp. When D-Day made clear that the Nazis would lose, and Groves told him that the focus of the project had always been the Soviets, Rotblat resigned.&nbsp;</p>

<p>In 1957, he <a href="https://www.nobelprize.org/prizes/peace/1995/pugwash/facts/">organized</a> a conference in the small lobster fishing village of Pugwash in Nova Scotia. The Pugwash Conferences would go on to spread key ideas for arms control agreements on nuclear testing, limiting warheads, and banning biological weapons. Rotblat and Pugwash <a href="https://www.nobelprize.org/prizes/peace/1995/rotblat/facts/">shared</a> the 1995 Nobel Peace Prize &mdash; one more Nobel than Oppenheimer ever received.&nbsp;</p>

<p>Again, this misremembering has important modern implications. The lesson of this tragic mistaken arms race shouldn&rsquo;t be &ldquo;never race.&rdquo; It should be &ldquo;make sure you know whether or not you&rsquo;re actually in a race.&rdquo; This is a lesson we have consistently failed to learn &mdash; the US mistakenly thought there was a &ldquo;<a href="https://www.britannica.com/topic/missile-gap">missile gap</a>&rdquo; in the late 1950s, and the Soviets mistakenly thought they were in a biological weapons arms race in the early 1970s. In the coming years, states may mistakenly believe they are in a race to develop powerful advanced artificial intelligence systems.</p>

<p>The &ldquo;individuals and organizations that have fought long and hard to reduce the number of nuclear weapons&rdquo; that Nolan paid tribute to &mdash; like the Bulletin of Atomic Scientists, Joseph Rotblat, and the Pugwash Conferences &mdash; weren&rsquo;t really shown in his remarkable, magnificent, affecting film. Indeed, they were never supported by Oppenheimer. We&rsquo;re in a tough spot, facing the possibility of a new nuclear arms race. We need to learn from these successful arms controllers, rather than from Oppenheimer&rsquo;s failures.</p>

<p><em><strong>Update, March 11, 8:00 am: </strong>This story was published on March 10, 2024, and has been updated to reflect news of </em>Oppenheimer<em>&rsquo;s Oscar wins.</em></p>
						]]>
									</content>
			
					</entry>
			<entry>
			
			<author>
				<name>Haydn Belfield</name>
			</author>
			
			<title type="html"><![CDATA[“Cry baby scientist”: What Oppenheimer the film gets wrong about Oppenheimer the man]]></title>
			<link rel="alternate" type="text/html" href="https://www.vox.com/future-perfect/2023/7/22/23803380/j-robert-oppenheimer-oscar-winning-film-nuclear-weapons-manhattan-project-christopher-nolan" />
			<id>https://www.vox.com/future-perfect/2023/7/22/23803380/j-robert-oppenheimer-oscar-winning-film-nuclear-weapons-manhattan-project-christopher-nolan</id>
			<updated>2024-03-10T21:04:33-04:00</updated>
			<published>2023-07-22T08:00:00-04:00</published>
			<category scheme="https://www.vox.com" term="Culture" /><category scheme="https://www.vox.com" term="Future Perfect" /><category scheme="https://www.vox.com" term="Movies" />
							<summary type="html"><![CDATA[One would be tempted to describe J. Robert Oppenheimer as a tragic figure &#8212; that&#8217;s certainly how Christopher Nolan portrays him in the biopic Oppenheimer. The father of the atomic bomb who spent the rest of his life agonizing over what he had helped birth; the ultimate insider who was humbled and brought low; the [&#8230;]]]></summary>
			
							<content type="html">
											<![CDATA[

						
<figure>

<img alt="" data-caption="American theoretical physicist and professor of physics J. Robert Oppenheimer at the University of California, Berkeley | Photo 12/Universal Images Group via Getty Images" data-portal-copyright="Photo 12/Universal Images Group via Getty Images" data-has-syndication-rights="1" src="https://platform.vox.com/wp-content/uploads/sites/2/chorus/uploads/chorus_asset/file/24801980/GettyImages_815208138.jpg?quality=90&#038;strip=all&#038;crop=0,0,100,100" />
	<figcaption>
	American theoretical physicist and professor of physics J. Robert Oppenheimer at the University of California, Berkeley | Photo 12/Universal Images Group via Getty Images	</figcaption>
</figure>
<p>One would be tempted to describe J. Robert Oppenheimer as a tragic figure &mdash; that&rsquo;s certainly how Christopher Nolan portrays him in the biopic <em>Oppenheimer</em>. The father of the atomic bomb who spent the rest of his life agonizing over what he had helped birth; the ultimate insider who was humbled and brought low; the hopeful scientist who started the nuclear arms race. But then, tragic figures don&rsquo;t generally spend their retirement yachting around the Caribbean. Or maybe he was a tragic figure in the mold of Lord Byron &mdash; interestingly dark and mystical, remarkably pretty, and rich as Midas.</p>

<p>Oppenheimer grew up in privilege, and remained swaddled in it for his whole life. His father immigrated to New York with nothing, and rose up to become a wealthy textile company executive. His parents spoiled their little genius. When he started a childhood rock collection, it grew to cover every surface in their apartment, which itself covered an entire floor overlooking the Hudson River. The Oppenheimers had a chauffeur, a French governess, three live-in maids and three van Gogh paintings. He <a href="https://people.com/oppenheimer-movie-true-story-7563024">corresponded</a> with the New York Mineralogical Club, but when they invited him to speak they were surprised and delighted when he turned out to be only 12. His 16th birthday present was a 28-foot yacht (to go with the family&rsquo;s 40-foot <em>Lorelei</em>) which he <a href="https://content.time.com/time/subscriber/article/0,33009,860615-2,00.html">called</a> <em>Trimethy</em>, after a chemical compound. As Oppenheimer <a href="https://www.smithsonianmag.com/history/exclusive-behind-scenes-look-los-alamos-lab-where-robert-oppenheimer-created-atomic-bomb-180982336/">remarked</a> when he bought his first holiday home in New Mexico, the state where he would later spearhead the development of the atomic bomb: &ldquo;hot dog!&rdquo;&nbsp;</p>
<div class="wp-block-vox-media-highlight vox-media-highlight"><h2 class="wp-block-heading"><strong>Oppenheimer, explained</strong></h2><ul class="wp-block-list"><li><a href="https://www.vox.com/future-perfect/2023/7/19/23799375/oppenheimer-movie-trinity-test-atomic-bomb-ethics-existential-risk">What’s the true story behind the atomic test?</a></li><li><a href="https://www.vox.com/culture/23800888/oppenheimer-review-physics-donne-trinity-christopher-nolan-fission-fusion">How does the movie use fission and fusion timelines?</a></li><li><a href="https://www.vox.com/videos/2023/7/20/23801640/oppenheimer-los-alamos-secret-city-new-mexico">Why was Los Alamos chosen for the Manhattan Project?</a></li><li><a href="https://www.vox.com/future-perfect/2023/7/22/23803380/j-robert-oppenheimer-film-movie-nuclear-weapons-manhattan-project-world-war-ii-christopher-nolan">What does the movie get wrong about the man?</a></li><li><a href="https://www.vox.com/culture/23806019/oppenheimer-fashion-hat-menswear-movie">Is Oppenheimer fashion on the way?</a></li><li><a href="https://www.vox.com/politics/2023/7/24/23800777/oppenheimer-christopher-nolan-atomic-bomb-true-story-los-alamos-manhattan-project">How the movie forces us to the ask new questions about the nuclear arms race.</a></li><li><a href="https://www.vox.com/culture/23789864/barbenheimer-barbieheimer-barbie-oppenheimer-release-memes-double-feature">What is Barbenheimer?</a></li><li><a href="https://www.vox.com/culture/23808552/atomic-bomb-manhattan-strangelove-oppenheimer-pop-culture">Why the atomic bomb is all over pop culture now.</a></li></ul></div>
<p>Oppenheimer was a slightly odd student. He was a nerd at Harvard, excluded for his introversion and, in the intensely antisemitic environment of the 1920s, for his Jewishness. He was a somewhat troubled youth. At Cambridge University, he <a href="https://www.vanityfair.com/hollywood/2023/07/oppenheimer-poison-apple-true-story">once left</a> a poisoned apple on his tutor&rsquo;s desk; on vacation when a friend told him of his engagement, Oppenheimer tried to <a href="https://www.sciencehistory.org/stories/disappearing-pod/the-real-tragedy-of-robert-oppenheimer/">strangle</a> him; and in Gottingen, where he was a PhD student, his classmates <a href="https://unherd.com/2023/07/we-wouldnt-want-oppenheimer-today/">presented</a> a petition to get him to stop interrupting seminars.</p>

<p>However, he began to come out of his shell as a postdoctoral researcher in Leiden and Zurich, and became positively cool when he moved to California in 1929. He cooked nasi goreng &mdash; his colleagues <a href="https://ahf.nuclearmuseum.org/voices/oral-histories/harold-chernisss-interview-part-1/">called</a> it &ldquo;nasty gory&rdquo; &mdash; and &ldquo;<a href="https://www.google.com/books/edition/American_Prometheus/jfSn2RJZI9EC?hl=en&amp;gbpv=1&amp;dq=%22eggs+a+la+oppie%22&amp;pg=PA96&amp;printsec=frontcover">eggs a la Oppie</a>,&rdquo; made with lots of Mexican chiles. He had a house with a Picasso on the wall, New Mexican rugs on the floor and a view of the Golden Gate Bridge. He fundraised for Republican forces in the Spanish Civil War and flirted with communism. With his blackboard chalk and his cigarettes, he made significant breakthroughs, inspired his graduate students, and <a href="https://www.osti.gov/opennet/manhattan-project-history/People/Administrators/robert-oppenheimer.html">built</a> one of the finest theoretical physics departments in the world. And he was lucky: His father&rsquo;s fortune was unscathed by the Crash of 1929. <a href="https://www.google.com/books/edition/American_Prometheus/jfSn2RJZI9EC?hl=en&amp;gbpv=1&amp;dq=%22natalie+raymond%22+and+cezanne&amp;pg=PA91&amp;printsec=frontcover">Once after a crash</a> of Oppenheimer&rsquo;s own, speeding in his Chrysler while racing a train and knocking unconscious and almost killing his passenger Natalie Raymond, his dad gave her a Cezanne drawing by way of an apology. Hot dog!</p>

<p>After the war, he got the cushiest job imaginable, as director of the Institute for Advanced Study in Princeton, New Jersey. As director, he was given the 265-acre Olden Manor, parts of which dated to 1696. He had no teaching responsibilities, and a $120,000 fund to spend on inviting whoever he liked to spend anything from a few months (T.S. Eliot, whose poem &ldquo;The Wasteland&rdquo; Oppenheimer is depicted absorbing onscreen) to the rest of their career (the diplomat George Kennan, he of the <a href="https://www.wilsoncenter.org/article/remembering-george-f-kennan#:~:text=The%20Policymaker-,George%20F.,sent%20on%20February%2022%2C%201946.">Cold War containment policy</a>). It sounds like a great gig. And if I had it, I also would have essentially stopped producing research, as Oppenheimer did.</p>
<figure class="wp-block-pullquote alignleft"><blockquote><p>Oppenheimer spent much of the 50s and 60s in his holiday home at Hawksnest Bay on the Caribbean island of Saint John or on his yacht</p></blockquote></figure>
<p>Eventually McCarthyism, red-baiting FBI director J. Edgar Hoover, and Oppenheimer&rsquo;s own political mistakes came for him, and he lost his security clearance and his political appointments in 1954, events that serve as the framing device for Nolan&rsquo;s film. But Oppenheimer remained as director of the institute until his death. The sheer ludicrous unfairness of the Republican show-trial security hearing &mdash; puppet-mastered by the banker turned atomic energy adviser Lewis Strauss &mdash; made him a martyr, and when the Democrats got back into the White House they gave him a special award. Oppenheimer spent much of the 50s and 60s in his holiday home at Hawksnest Bay on the Caribbean island of Saint John (where he imported champagne by the case) or on his yacht.</p>

<p>By comparison, his brother Frank became a Communist Party member in 1937 while attempting to <a href="https://www.lrb.co.uk/the-paper/v32/n18/steven-shapin/uncle-of-the-bomb">desegregate</a> his local swimming pool in Pasadena; was an early campaigner at Los Alamos on international arms control; and then was <a href="https://www.nytimes.com/1985/02/04/us/frank-oppenheimer-nuclear-physicist-dies.html">blacklisted</a> from academia, denied a passport, and left to spend a decade as a cattle rancher.&nbsp;</p>
<h2 class="wp-block-heading">Los Alamos’s camp counselor-in-chief</h2>
<p>But the central location in Oppenheimer&rsquo;s life wasn&rsquo;t the Upper West Side, the Bay Shore mansion on Long Island, his bachelor pad in California, the manor in Princeton, or his Caribbean island. The central location was Los Alamos. This scientific base was built from scratch, up in the hills of northern New Mexico. It was Oppenheimer&rsquo;s favorite part of the country; indeed, Los Alamos was a day&rsquo;s horse ride from his holiday home. It was like locating CERN, the massive intergovernmental particle physics lab, in the pleasant English countryside of the Cotswolds.</p>

<p>Los Alamos during wartime sounds like great fun. Married scientists were permitted to bring their families. There were barn dances or piano recitals on a Saturday night, hikes and horse-riding on a Sunday. It had a local cinema, 15 cents a ticket. It had a local theater group: Oppenheimer even <a href="https://www.nps.gov/thingstodo/los-alamos-see-a-play-at-the-los-alamos-little-theater.htm">played</a> a corpse in the comedy <em>Arsenic and Old Lace</em>. And it had large quantities of booze &mdash; Oppenheimer <a href="https://janeliasberg.com/oppenheimers-legendary-martini-or-the-manhattan-project/#:~:text=Owing%20to%20the%20difficulty%20of,well%20as%20lime%20and%20honey.&amp;text=Stir%20the%20gin%20and%20vermouth%20with%20ice%20until%20chilled.">was famous for</a> mixing very strong, very cold martinis, while the tipple of choice for the less well-heeled bachelor scientists was half lab alcohol and half grapefruit juice, chilled with a chunk of smoking dry ice. The <a href="https://ladailypost.com/wartime-baby-boom-left-general-groves-fuming-while-parents-counted-and-counted-and-counted-their-blessings/#:~:text=The%20average%20age%20in%20Los,no%20time%20in%20doing%20so.">average age was 25</a>. And everyone, in between the work of creating the atom bomb, was apparently having sex: 80 children were born the first year, and 10 a month after that. All in all, it makes for a better war than storming beaches in Normandy or Iwo Jima.</p>

<p>The comforts provided to the scientists and their families have been described as &ldquo;army socialism.&rdquo; But the soldiers who emptied the bins and the local Indigenous women who cleaned the houses must have had a pretty clear sense of the pecking order. In the many Manhattan Project memoirs, Los Alamos reminds one far more of the summer camp it was before the war than a top-secret government project to develop a weapon of mass destruction</p>
<figure class="wp-block-pullquote alignleft"><blockquote><p>Oppenheimer’s chief contribution was as camp counselor of Los Alamos</p></blockquote></figure>
<p>Oppenheimer&rsquo;s historic contribution was as scientific director of Los Alamos. But what was the nature of that contribution to the Manhattan Project? Not the science &mdash; the real breakthroughs were from Otto Hahn and Fritz Strassmann, who showed nuclear fission was possible, or specialists like Robert Christy, who designed the plutonium implosion &ldquo;Christy gadget&rdquo; successfully tested at Trinity Site near Los Alamos, and later dropped on Nagasaki. And not the direction &mdash; 90 percent of Manhattan Project director Gen. Leslie Groves&rsquo;s <a href="https://www.brookings.edu/the-costs-of-the-manhattan-project/">budget went</a> to the Fordist feats of administration, logistics, and industrial engineering that were the Oak Ridge and Hanford production plants, churning out the plutonium and enriched uranium that fueled the atom bombs. Oppenheimer&rsquo;s chief contribution was as camp counselor of Los Alamos.&nbsp;</p>

<p>Oppenheimer encouraged them on, and his charisma cast a sort of spell over the campers. It is no coincidence that much of the serious thinking about the bomb &mdash; morally and politically &mdash; happened elsewhere, in Chicago under Leo Szilard or in the giant head of the Danish genius Niels Bohr. Oppenheimer whipped them up with a simple message: we need to get the bomb before Hitler.</p>

<p>As it turns out, this was all mistaken. We now know that the Nazis <a href="http://large.stanford.edu/courses/2014/ph241/wendorff2/#:~:text=Thus%20in%20December%201941%2C%20the,make%20a%20more%20immediate%20impact.">had decided</a> against a nuclear fission program by 1942. Nazi planners needed raw materials and manpower for armaments production, and Nazi scientists thought a bomb couldn&rsquo;t be delivered in time to affect the war in Europe, which very much proved to be the case. So the Manhattan Project did not in fact deter, and did not need to deter, Hitler from developing and using the bomb. The scientists were working based on a mistake.</p>

<p>The main effect of the Manhattan Project was to bring forward in time the era of the bomb and the era of the nuclear arms race. The existential risk researcher Toby Ord <a href="https://theprecipice.com/">calls</a> this era &ldquo;the Precipice&rdquo;: the first period in which humanity can destroy itself. The US would likely not have &ldquo;sprinted&rdquo; to the same extent, <a href="https://sgp.fas.org/crs/misc/RL34645.pdf">spending</a> 0.4 percent of GDP, for a peacetime Manhattan Project. And Oppenheimer&rsquo;s nemesis Lewis Strauss may have been right, if for the wrong reasons, when he accused Oppenheimer of helping the Soviet nuclear program. Quite simply, it would have taken the Soviets years longer if they couldn&rsquo;t just <a href="https://www.osti.gov/opennet/manhattan-project-history/Events/1942-1945/espionage.htm#:~:text=Soviet%20spies%20penetrated%20the%20Manhattan,development%20of%20the%20Soviet%20bomb.">copy the secrets</a> of the Manhattan Project. Szilard and Albert Einstein, whose 1939 letter prompted President Franklin Roosevelt to begin the US nuclear program, later <a href="https://www.haaretz.com/jewish/2016-10-11/ty-article/1939-einstein-makes-his-biggest-mistake/0000017f-db72-d3a5-af7f-fbfe922c0000">described</a> their advocacy for the project as the greatest mistake of their life.</p>

<p>This was not simply an honest mistake. Joseph Rotblat &mdash; the only scientist to <a href="https://ahf.nuclearmuseum.org/ahf/key-documents/rotblat-account/">resign</a> from the Manhattan Project &mdash; got a nasty shock in May 1944 when, at a dinner, Groves said, &ldquo;You realize, of course, that the main purpose of this project is to subdue the Russians.&rdquo; Later, Groves <a href="https://www.american.edu/ucm/news/20150804-kuznick-hiroshima.cfm#:~:text=%22There%20was%20never%20from%20about,%2C%22%20Groves%20would%20later%20say.">testified</a> that &ldquo;there was never, from about two weeks from the time I took charge of this Project, any illusion on my part but that Russia was our enemy.&rdquo; It is hard to reconcile this bloodlessness with Matt Damon&rsquo;s blithe face as Groves in Christopher Nolan&rsquo;s film.</p>
<figure class="wp-block-pullquote alignleft"><blockquote><p>Was the bomb just too “technically sweet” for Oppenheimer to resist?</p></blockquote></figure>
<p>How complicit was Oppenheimer? David Hawkins, Oppenheimer&rsquo;s aide and the Manhattan Project&rsquo;s official historian, claims that Groves told Oppenheimer at the end of 1943 that the Nazis had abandoned their attempt &mdash; and Oppenheimer shrugged. Oppenheimer dominated the ethical discussions among scientists in late 1944, as both the war and the race to the atomic bomb were nearing their end stages, arguing that scientists had no right to a louder voice than other citizens, and that if the war ended without nuclear use, the next war would be fought with nuclear weapons. Was Oppenheimer swept up by the same patriotic fervor that prompted him to have a colonel&rsquo;s uniform tailored for himself? Was the bomb <a href="https://www.oxfordreference.com/display/10.1093/acref/9780191826719.001.0001/q-oro-ed4-00007996;jsessionid=ADC81200791B4E377357A86763004BA6">just too</a> &ldquo;technically sweet&rdquo; for him to resist? It is unclear. Perhaps the best we can say in his defense was that Oppenheimer was chumped into doing it (to some extent), and inadvertently or not, he chumped the other scientists as well.</p>
<h2 class="wp-block-heading">“Would you like to wipe your hands?”</h2>
<p>Oppenheimer&rsquo;s complicity did give him prestige and access. However, he squandered that, and lost four key political battles over the use and future of nuclear weapons: on a demonstration attack, on beginning talks at the Potsdam conference, on arms control proposals after the war, and on not racing for the far more powerful hydrogen bomb.</p>

<p>The two key issues on the agenda at the May 31, 1945, meeting of the &ldquo;Interim Committee,&rdquo; a <a href="https://www.atomicarchive.com/history/manhattan-project/p5s2.html#:~:text=The%20briefing%20summarized%20the%20consensus,Secretary%20of%20State%20James%20F.">government advisory group</a> on atomic research, were how to use the bomb, and how to communicate to the Soviets. Oppenheimer, the vast majority of Los Alamos scientists, and indeed Gen. Dwight Eisenhower, all supported a demonstration attack on an empty island. But Harvard President James Conant instead suggested &ldquo;a vital war plant &hellip; surrounded by workers&rsquo; houses.&rdquo;&nbsp;</p>

<p>At this crucial decision-making meeting, Oppenheimer did not disagree with the targeting of civilians, instead merely noting the visual effect of a bomb and the feasibility of simultaneous strikes. He also stayed quiet when Groves got approval to purge dissenting scientists like Szilard from the project. Oppenheimer thought that he had traded these betrayals for a commitment that the USSR was to be clearly informed of the bomb and its planned use. These discussions would mean that the Soviets would not be blindsided in a frightening manner that would spur an arms race. But instead, in his meeting with Soviet leader Joseph Stalin at the Potsdam conference, just after the successful Trinity test, Truman only casually and vaguely mentioned a new weapon, and had no serious discussion with his opposite number. Oppenheimer had lost on both counts.</p>

<p>The first time he met Truman, after the atomic bombings of Japan, out of frustration and passion Oppenheimer blurted out, &ldquo;There is blood on my hands.&rdquo; Truman would stew on this for years, <a href="https://www.sfgate.com/sf-culture/article/oppenheimer-truman-disastrous-meeting-18230557.php">retelling and embellishing</a> the anecdote, once claiming he pulled out his handkerchief and said &ldquo;Well, here, would you like to wipe your hands?&rdquo; Immediately after he left, Truman called him a &ldquo;cry baby scientist,&rdquo; and would never trust him again.</p>

<p>Oppenheimer&rsquo;s postwar record was just as bad. He was the main intellectual force behind the <a href="https://history.state.gov/milestones/1945-1952/baruch-plans">1946 Acheson-Lilienthal Report</a>, which proposed a single worldwide Atomic Development Agency with a monopoly over all uranium mines, labs, enrichment facilities, and power plants. Control over nuclear technology would be international, rather than national. However, as Oppenheimer later acknowledged, this was infeasible and naive. Stalin would never have agreed to renunciation of sovereignty, to the inspections, or to the depth of cooperation with the capitalist West the plan would have demanded. Bernard Baruch, proposer of the failed <a href="https://history.state.gov/milestones/1945-1952/baruch-plans">Baruch Plan</a>, was a convenient scapegoat.</p>

<p>When the Soviets exploded their first bomb in 1949, Oppenheimer told David Lilienthal, the first chair of the Atomic Energy Commission, that&nbsp;&ldquo;we mustn&rsquo;t muff it this time,&rdquo; meaning the arms race. But they did muff it, and the US stockpile <a href="https://web.mit.edu/fnl/volume/235/wallslides.pdf">grew from</a> 50 warheads in 1948 to 300 in 1950. The next fight was on whether to build a &ldquo;Super&rdquo; or hydrogen bomb, much more destructive than the atomic bomb. Oppenheimer opposed it on scientific, technical, and moral grounds. But when the decision came to Truman, the president had one question: can the Russians do it? The answer was yes. &ldquo;In that case,&rdquo; Truman <a href="https://academic.oup.com/book/12875/chapter-abstract/163201293?redirectedFrom=fulltext">replied</a>, &ldquo;we have no choice.&rdquo; The meeting took 7 minutes. The cry baby scientist&rsquo;s concerns were completely dismissed.&nbsp;</p>
<h2 class="wp-block-heading">How Oppenheimer was outplayed</h2>
<p>The two most notable facts about Oppenheimer&rsquo;s life are that he first sped up the creation of nuclear weapons, and then failed utterly to restrict the nuclear arms race he had helped begin. The arms racers used his scientific credibility to support their reckless buildup, and outplayed him in every important political battle. It would take a further 18 years after his 1954 defrocking before the <a href="https://history.state.gov/milestones/1969-1976/salt#:~:text=Nixon%20and%20Soviet%20General%20Secretary,nuclear%20missiles%20in%20their%20arsenals.">first bilateral arms control agreement on nuclear weapons</a>. This removal of his security clearance can be seen as the final mercy kill of an utterly defanged and defeated political opponent.</p>

<p>It&rsquo;s hard to overemphasize how much the authors of <em>American Prometheus, </em>the book on which the film is based, are on Team Oppenheimer. One author, Martin Sherwin, spent 25 years interviewing Oppenheimer&rsquo;s friends and family. They spend 88 pages on a minute-by-minute account of the mistrial of his hearing. They refer to him frequently as &ldquo;Oppie.&rdquo; And even their assessment is that he &ldquo;won nothing and acquiesced to everything.&rdquo;</p>

<p>How should we remember Oppenheimer: A tragic martyr? Death, the destroyer of worlds? The &ldquo;American Prometheus&rdquo; of the title? Another descriptive phrase comes to mind, one that would be more familiar to one of his father&rsquo;s employees in a New York textile factory: &ldquo;What a schmuck.&rdquo;</p>

<p><em>Haydn Belfield has been&nbsp;</em><a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__twitter.com_HaydnBelfield&amp;d=DwMGaQ&amp;c=7MSjEE-cVgLCRHxk1P5PWg&amp;r=mTM0ruyioL3vFpq5GgUZftnxirsFoCe5-UZtYwnKki8&amp;m=fafNoDRYchktF6CmyLmwtRE1Dt1uZmJeDTqC_-94paSKyG8-WBBDyvzsCPMMphux&amp;s=8wpMbuaRp27Bl2ciP_ieN5e9ac-s3eWypajx66fqlFk&amp;e="><em>academic project manager</em></a>&nbsp;<em>at the University of Cambridge&rsquo;s Centre for the Study of Existential Risk for the past six years. He is also an associate fellow at the Leverhulme Centre for the Future of Intelligence.</em></p>

<p><em><strong>Correction, July 22, 12:15 pm ET:</strong> A previous version of this story mistakenly referred to the author Kai Bird instead of Martin Sherwin.  </em></p>
						]]>
									</content>
			
					</entry>
			<entry>
			
			<author>
				<name>Haydn Belfield</name>
			</author>
			
			<title type="html"><![CDATA[If your AI model is going to sell, it has to be safe]]></title>
			<link rel="alternate" type="text/html" href="https://www.vox.com/future-perfect/2023/3/25/23655082/ai-openai-gpt-4-safety-microsoft-facebook-meta" />
			<id>https://www.vox.com/future-perfect/2023/3/25/23655082/ai-openai-gpt-4-safety-microsoft-facebook-meta</id>
			<updated>2023-03-27T12:55:22-04:00</updated>
			<published>2023-03-25T07:30:00-04:00</published>
			<category scheme="https://www.vox.com" term="Artificial Intelligence" /><category scheme="https://www.vox.com" term="Future Perfect" /><category scheme="https://www.vox.com" term="Innovation" /><category scheme="https://www.vox.com" term="Technology" />
							<summary type="html"><![CDATA[On March 14, OpenAI released the successor to ChatGPT: GPT-4. It impressed observers with its markedly improved performance across reasoning, retention, and coding. It also fanned fears around AI safety, around our ability to control these increasingly powerful models. But that debate obscures the fact that, in many ways, GPT-4&#8217;s most remarkable gains, compared to [&#8230;]]]></summary>
			
							<content type="html">
											<![CDATA[

						
<figure>

<img alt="" data-caption="" data-portal-copyright="CFOTO/Future Publishing via Getty Images" data-has-syndication-rights="1" src="https://platform.vox.com/wp-content/uploads/sites/2/chorus/uploads/chorus_asset/file/24533906/GettyImages_1249183770.jpg?quality=90&#038;strip=all&#038;crop=0,0,100,100" />
	<figcaption>
		</figcaption>
</figure>
<p>On March 14, OpenAI released the successor to ChatGPT: GPT-4. It <a href="https://www.vox.com/2023/3/15/23640640/gpt-4-chatgpt-openai-generative-ai">impressed</a> observers with its markedly improved performance across reasoning, retention, and coding. It also <a href="https://www.vox.com/future-perfect/2023/3/18/23645013/openai-gpt4-holden-karnofsky-artificial-intelligence-ai-safety-existential-risk">fanned fears around AI safety,</a> around our ability to control these increasingly powerful models. But that debate obscures the fact that, in many ways, GPT-4&rsquo;s most remarkable gains, compared to similar models in the past, have been around safety.&nbsp;</p>

<p>According to the company&rsquo;s <a href="https://arxiv.org/pdf/2303.08774.pdf">Technical Report</a>, during GPT-4&rsquo;s development, OpenAI &ldquo;spent six months on safety research, risk assessment, and iteration.&rdquo; OpenAI <a href="https://openai.com/product/gpt-4">reported</a> that this work yielded significant results: &ldquo;GPT-4 is 82% less likely to respond to requests for disallowed content and 40% more likely to produce factual responses than GPT-3.5 on our internal evaluations.&rdquo; (ChatGPT is a slightly tweaked version of GPT-3.5: if you&rsquo;ve been using ChatGPT over the last few months, you&rsquo;ve been interacting with GPT-3.5.)</p>

<p>This demonstrates a broader point: For AI companies, there are significant competitive advantages and profit incentives for emphasizing safety. The key success of ChatGPT over other companies&rsquo; large language models (LLMs) &mdash; apart from a nice user interface and remarkable word-of-mouth buzz &mdash; is precisely its safety. Even as it <a href="https://blog.gitnux.com/chat-gpt-statistics/#:~:text=Chat%20GPT%20has%20secured%20100,word%2Dof%2Dmouth%20marketing.">rapidly grew</a> to over 100 million users, it hasn&rsquo;t had to be taken down or significantly tweaked to make it less harmful (and less useful).</p>

<p>Tech companies should be investing heavily in safety research and testing for all our sakes, but also for their own commercial self-interest. That way, the AI model works as intended, and these companies can keep their tech online. ChatGPT Plus is <a href="https://www.itpro.com/business/business-strategy/369989/openai-launches-chatgpt-plus-greater-revenue">making money</a>, and you can&rsquo;t make money if you&rsquo;ve had to take your language model down. OpenAI&rsquo;s reputation has been increased by its tech being safer than its competitors, while other tech companies have had their reputations hit by their tech being unsafe, and even having to take it down. (Disclosure: I am listed in the acknowledgments of the GPT-4 System Card, but I have not shown the draft of this story to anyone at OpenAI, nor have I taken funding from the company.)</p>
<h2 class="wp-block-heading">The competitive advantage of AI safety</h2>
<p>Just ask Mark Zuckerberg. When Meta released its large language model BlenderBot 3 in August 2022, it immediately <a href="https://www.theguardian.com/technology/2022/aug/10/meta-ai-facebook-blenderbot-3-chatbot">faced problems</a> of making inappropriate and untrue statements. Meta&rsquo;s Galactica was only up for three days in November 2022 before <a href="https://www.technologyreview.com/2022/11/18/1063487/meta-large-language-model-ai-only-survived-three-days-gpt-3-science/">it was withdrawn</a> after it was shown confidently &lsquo;hallucinating&rsquo; (making up) academic papers that didn&rsquo;t exist. Most recently, in February 2023, Meta irresponsibly released the full weights of its latest language model, LLaMA. As many experts predicted would happen, it <a href="https://www.theverge.com/2023/3/8/23629362/meta-ai-language-model-llama-leak-online-misuse">proliferated to 4chan</a>, where it will be used to mass-produce disinformation and hate.&nbsp;</p>

<p>I and my co-authors warned about this five years ago in a 2018 report called &ldquo;<a href="https://arxiv.org/abs/1802.07228">The Malicious Use of Artificial Intelligence</a>,&rdquo; while the Partnership on AI (Meta was a founding member and remains an active partner) had a great <a href="https://partnershiponai.org/paper/responsible-publication-recommendations/">report</a> on responsible publication in 2021. These repeated and failed attempts to &ldquo;move fast and break things&rdquo; have probably exacerbated Meta&rsquo;s trust problems. In <a href="https://www.jair.org/index.php/jair/article/view/12895/26701">surveys</a> from 2021 of AI researchers and the US public on trust in actors to shape the development and use of AI in the public interest, &ldquo;Facebook [Meta] is ranked the least trustworthy of American tech companies.&rdquo;</p>

<p>But it&rsquo;s not just Meta. The original misbehaving machine learning chatbot was Microsoft&rsquo;s Tay, which was <a href="https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist">withdrawn 16 hours after it was released</a> in 2016 after making racist and inflammatory statements. Even Bing/Sydney had some very erratic responses, including <a href="https://www.nytimes.com/2023/02/16/technology/bing-chatbot-transcript.html">declaring</a> its love for, and then threatening, a journalist. In response, Microsoft limited the number of messages one could exchange, and Bing/Sydney no longer answers questions about itself.&nbsp;</p>

<p>We now know Microsoft based it on OpenAI&rsquo;s GPT-4; Microsoft invested $11 billion into OpenAI in return for OpenAI running all their computing on Microsoft&rsquo;s Azure cloud and <a href="https://news.microsoft.com/2019/07/22/openai-forms-exclusive-computing-partnership-with-microsoft-to-build-new-azure-ai-supercomputing-technologies/">becoming</a> their &ldquo;preferred partner for commercializing new AI technologies.&rdquo; But it is unclear why the model responded so strangely. It could have been an early, not fully safety-trained version, or it could be due to its connection to search and thus its ability to &ldquo;read&rdquo; and respond to an article about itself in real time. (By contrast, GPT-4&rsquo;s training data only runs up to <a href="https://platform.openai.com/docs/models/gpt-4">September 2021</a>, and it does not have access to the web.) It&rsquo;s notable that even as it was heralding its new AI models, Microsoft recently <a href="https://www.theverge.com/2023/3/13/23638823/microsoft-ethics-society-team-responsible-ai-layoffs">laid off</a> its AI ethics and society team.</p>

<p>OpenAI took a different path with GPT-4, but it&rsquo;s not the only AI company that has been putting in the work on safety. Other leading labs have also been making clear their commitments, with <a href="https://www.anthropic.com/index/core-views-on-ai-safety">Anthropic</a> and <a href="https://www.alignmentforum.org/posts/a9SPcZ6GXAg9cNKdi/linkpost-some-high-level-thoughts-on-the-deepmind-alignment">DeepMind</a> publishing their safety and alignment strategies. These two labs have also been safe and cautious with the development and deployment of <a href="https://www.anthropic.com/index/introducing-claude">Claude</a> and <a href="https://www.deepmind.com/blog/building-safer-dialogue-agents">Sparrow</a>, their respective LLMs.</p>
<h2 class="wp-block-heading">A playbook for best practices</h2>
<p>Tech companies developing LLMs and other forms of cutting-edge, impactful AI should learn from this comparison. They should adopt the best practice as shown by OpenAI: Invest in safety research and testing before releasing.</p>

<p>What does this look like specifically? GPT-4&rsquo;s <a href="https://cdn.openai.com/papers/gpt-4-system-card.pdf">System Card</a> describes four steps OpenAI took that could be a model for other companies.</p>

<p>First, prune your dataset for toxic or inappropriate content. Second, train your system with reinforcement learning from human feedback (RLHF) and rule-based reward models (RBRMs). RLHF involves human labelers creating demonstration data for the model to copy and ranking data (&ldquo;output A is preferred to output B&rdquo;) for the model to better predict what outputs we want. RLHF produces a model that is sometimes overcautious, refusing to answer or hedging (as some users of ChatGPT will have noticed).&nbsp;</p>

<p>RBRM is an automated classifier that evaluates the model&rsquo;s output on a set of rules in multiple-choice style, then rewards the model for refusing or answering for the right reasons and in the desired style. So the combination of RLHF and RBRM encourages the model to answer questions helpfully, refuse to answer some harmful questions, and distinguish between the two.&nbsp;</p>

<p>Third, provide <a href="https://www.governance.ai/post/sharing-powerful-ai-models">structured access</a> to the model through an API. This allows you to filter responses and monitor for poor behavior from the model (or from users). Fourth, invest in moderation, both by humans and by automated moderation and content classifiers. For example, OpenAI used GPT-4 to create rule-based classifiers that flag model outputs that could be harmful.</p>

<p>This all takes time and effort, but it&rsquo;s worth it. Other approaches can also work, like Anthropic&rsquo;s rule-following <a href="https://www.anthropic.com/index/measuring-progress-on-scalable-oversight-for-large-language-models">Constitutional AI</a>, which leverages RL from AI feedback (RLAIF) to complement human labelers. As OpenAI acknowledges, their approach is not perfect: the model still hallucinates and can still sometimes be tricked into providing harmful content. Indeed, there&rsquo;s room to go beyond and <a href="https://time.com/6247678/openai-chatgpt-kenya-workers/">improve upon OpenAI&rsquo;s approach</a>, for example by providing more compensation and career progression opportunities for the human labelers of outputs.</p>

<p>Has OpenAI become less open? If this means less open source, then no. OpenAI adopted a &ldquo;<a href="https://arxiv.org/abs/1908.09203">staged release</a>&rdquo; strategy for GPT-2 in 2019 and an <a href="https://openai.com/blog/openai-api">API</a> in 2020. Given Meta&rsquo;s 4chan experience, this seems justified. As Ilya Sutskever, OpenAI chief scientist, noted to <a href="https://www.theverge.com/2023/3/15/23640180/openai-gpt-4-launch-closed-research-ilya-sutskever-interview">The Verge</a>: &ldquo;I fully expect that in a few years it&rsquo;s going to be completely obvious to everyone that open-sourcing AI is just not wise.&rdquo;&nbsp;</p>

<p>GPT-4 did have less information than previous releases on &ldquo;architecture (including model size), hardware, training compute, dataset construction, training method.&rdquo; This is because OpenAI is concerned about <a href="https://cdn.openai.com/papers/gpt-4-system-card.pdf">acceleration risk</a>: &ldquo;the risk of racing dynamics leading to a decline in safety standards, the diffusion of bad norms, and accelerated AI timelines, each of which heighten societal risks associated with AI.&rdquo;&nbsp;</p>

<p>Providing those technical details would speed up the overall rate of progress in developing and deploying powerful AI systems. However, AI poses many unsolved governance and technical challenges: For example, the <a href="https://www.nist.gov/itl/ai-risk-management-framework/roadmap-nist-artificial-intelligence-risk-management-framework-ai">US</a> and <a href="https://artificialintelligenceact.eu/standard-setting/">EU</a> won&rsquo;t have detailed safety technical standards for high-risk AI systems ready until early 2025.</p>

<p>That&rsquo;s why I <a href="https://www.vox.com/future-perfect/23619354/openai-chatgpt-sam-altman-artificial-intelligence-regulation-sydney-microsoft-ai-safety">and others</a> believe we shouldn&rsquo;t be speeding up progress in AI capabilities, but we should be going full speed ahead on safety progress. Any reduced openness should never be an impediment to safety, which is why it&rsquo;s so useful that the System Card shares details on safety challenges and mitigation techniques. Even though OpenAI seems to be coming around to this view, they&rsquo;re still at the forefront of pushing forward capabilities, and should provide more information on how and when they envisage themselves and the field slowing down.&nbsp;&nbsp;</p>

<p>AI companies should be investing significantly in safety research and testing. It is the right thing to do and will soon be required by regulation and safety standards in the EU and USA. But also, it is in the self-interest of these AI companies. Put in the work, get the reward.</p>

<p><em>Haydn Belfield has been&nbsp;</em><a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__twitter.com_HaydnBelfield&amp;d=DwMGaQ&amp;c=7MSjEE-cVgLCRHxk1P5PWg&amp;r=mTM0ruyioL3vFpq5GgUZftnxirsFoCe5-UZtYwnKki8&amp;m=fafNoDRYchktF6CmyLmwtRE1Dt1uZmJeDTqC_-94paSKyG8-WBBDyvzsCPMMphux&amp;s=8wpMbuaRp27Bl2ciP_ieN5e9ac-s3eWypajx66fqlFk&amp;e="><em>academic project manager</em></a>&nbsp;<em>at the University of Cambridge&rsquo;s Centre for the Study of Existential Risk (CSER) for the past six years. He is also an associate fellow at the Leverhulme Centre for the Future of Intelligence.</em></p>
						]]>
									</content>
			
					</entry>
			<entry>
			
			<author>
				<name>Haydn Belfield</name>
			</author>
			
			<title type="html"><![CDATA[There’s no libertarian approach to preventing the end of the world]]></title>
			<link rel="alternate" type="text/html" href="https://www.vox.com/future-perfect/2023/3/7/23618766/peter-thiel-existential-risk-oxford-union-silicon-valley-technology-artficial-intelligence" />
			<id>https://www.vox.com/future-perfect/2023/3/7/23618766/peter-thiel-existential-risk-oxford-union-silicon-valley-technology-artficial-intelligence</id>
			<updated>2023-03-14T09:21:30-04:00</updated>
			<published>2023-03-07T07:00:00-05:00</published>
			<category scheme="https://www.vox.com" term="Future Perfect" /><category scheme="https://www.vox.com" term="Influence" /><category scheme="https://www.vox.com" term="Peter Thiel" /><category scheme="https://www.vox.com" term="Technology" />
							<summary type="html"><![CDATA[Peter Thiel &#8212; tech billionaire, libertarian polemicist, Trump donor &#8212; recently gave a speech at the Oxford Union, one of the oldest and most prestigious student debating societies in the world, to kick off its 200th year. That&#8217;s hardly news &#8212; we&#8217;ve all heard Thiel&#8217;s spiel many times before on campus conformity and how only [&#8230;]]]></summary>
			
							<content type="html">
											<![CDATA[

						
<figure>

<img alt="" data-caption="Peter Thiel, co-founder of PayPal, Palantir Technologies, and Founders Fund, gestures as he speaks during the Bitcoin 2022 conference at Miami Beach Convention Center on April 7, 2022, in Miami, Florida. | Marco Bello/Getty Images" data-portal-copyright="Marco Bello/Getty Images" data-has-syndication-rights="1" src="https://platform.vox.com/wp-content/uploads/sites/2/chorus/uploads/chorus_asset/file/24467455/GettyImages_1239811817.jpg?quality=90&#038;strip=all&#038;crop=0,0,100,100" />
	<figcaption>
	Peter Thiel, co-founder of PayPal, Palantir Technologies, and Founders Fund, gestures as he speaks during the Bitcoin 2022 conference at Miami Beach Convention Center on April 7, 2022, in Miami, Florida. | Marco Bello/Getty Images	</figcaption>
</figure>
<p>Peter Thiel &mdash; tech billionaire, libertarian polemicist, Trump donor &mdash; recently gave a <a href="https://youtu.be/fQ4rc7npiXQ">speech</a> at the Oxford Union, one of the oldest and most prestigious student debating societies in the world, to kick off its 200th year. That&rsquo;s hardly news &mdash; we&rsquo;ve all heard Thiel&rsquo;s spiel many times before on campus conformity and how only tech can save us.</p>

<p>But my ears pricked up this time as he specifically criticized my field. I&rsquo;m an <a href="https://www.cser.ac.uk/">existential risk researcher at Cambridge University</a>, where my colleagues and I study the risks from nuclear and biological weapons, climate change, and emerging technology such as synthetic biology and artificial intelligence. All of these technologies pose incredibly high risks &mdash; we <a href="https://www.sciencedirect.com/science/article/pii/S0016328717301957">think it&rsquo;s plausible</a> that one or more of them could lead to civilizational collapse or extinction, affecting everyone alive today. As many in the effective altruism community <a href="https://80000hours.org/articles/how-to-reduce-existential-risk/">have argued</a>, I think tackling these risks is a key priority of our time.&nbsp;</p>

<p>Thiel seems to have had a passing interest in these topics a decade ago, <a href="https://youtu.be/KKLDevYyE9I">speaking at some conferences</a> and <a href="https://www.su.org/press/singularity-university-acquires-the-singularity-summit">donating some money</a>. But to my knowledge he has not engaged with the existential risk reduction community for as long as I have been involved. Instead, he seems increasingly interested in <a href="https://www.seasteading.org/31937-2/">seasteading</a> and the <a href="https://www.nytimes.com/2022/02/14/technology/republican-trump-peter-thiel.html">alt-right</a>.&nbsp;</p>

<p>So why was he criticizing the field of existential risk reduction? Thiel seems to suggest we in the community are Luddites, bearing some responsibility for the <a href="https://insight.kellogg.northwestern.edu/article/wage-stagnation-in-america">stagnation in real wages</a> and <a href="https://www.parkersoftware.com/blog/great-tech-stagnation-finally-beginning/">technological progress</a> since the 1970s. He claims a leading cause of stagnation is that scientists effectively have become too scared of their own technology. He told the Oxford Union audience that &ldquo;the single answer as to why it is stalled out on the part of the universities is something like science and technology are just too dangerous.&rdquo;</p>
<h2 class="wp-block-heading">Existential risk isn’t a hoax</h2>
<p>I don&rsquo;t want to minimize the situation. It really is true that real wages for many workers in the UK and US have been stagnant since the 1970s, especially through the last grueling decade of austerity. And too much technical effort and venture capital has been suboptimally invested into e-commerce (like PayPal), social media and online advertising (like Facebook), or surveillance (like Palantir). (What&rsquo;s the connection? All three companies helped to make Thiel&rsquo;s estimated <a href="https://www.bloomberg.com/billionaires/profiles/peter-a-thiel/?leadSource=uverify%20wall">$8 billion fortune</a>.)&nbsp;</p>

<p>But as someone who spends a fair amount of time encouraging technologists to consider their responsibilities for the technology they create, let me say that an overabundance of fear is not typically what I encounter.&nbsp;&nbsp;</p>

<p>Powerful technologies are often &ldquo;dual-use&rdquo;: They can be used to help or to harm people. Take nuclear physics. Nuclear weapons are in many ways the original existential risk, the one that has loomed over the world since 1945. However, nuclear power is also a reliable, zero-carbon source of power.&nbsp;</p>

<p>Advances in biotechnology enabled the <a href="https://www.cgdev.org/blog/covid-19-vaccine-rollout-was-fastest-global-history-low-income-countries-were-left-behind">quickest vaccine rollout in history,</a> as well as further medical breakthroughs. But they also enable <a href="https://www.vox.com/future-perfect/2023/2/1/23580528/gain-of-function-virology-covid-monkeypox-catastrophic-risk-pandemic-lab-accident">&ldquo;gain-of-function&rdquo;</a> experiments where scientists purposely try to make diseases more virulent.&nbsp;</p>

<p>Language models like ChatGPT are wonderful and amazing but in the wrong hands could enable the mass production of disinformation, as we warned about in <a href="https://arxiv.org/abs/1802.07228">&ldquo;The Malicious Use of Artificial Intelligence&rdquo;</a> in 2018. The problem with arguing for &ldquo;more speed&rdquo; is that these dual-use technologies are already moving faster than we can keep up.</p>

<p>Thiel is simply wrong if he thinks that &ldquo;slow it down&rdquo; is the only response. The <a href="https://founderspledge.com/funds/climate-change-fund">Founders Pledge Climate Change Fund</a>, a community of entrepreneurs who pledge to donate a portion of their exit earnings to charity, tries to speed up innovation in low-carbon concrete and steel. <a href="https://www.alvea.bio/">Alvea</a> is a biotech startup aiming to speed up vaccine production. <a href="https://www.anthropic.com/">Anthropic</a> is an AI safety company trying to speed up interpretability and alignment. We call this <a href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4213670">&rdquo;differential technology development&rdquo;</a>: speeding up safe or defensive tech relative to harmful or offensive tech.&nbsp;</p>

<p>But it&rsquo;s also clear that the old Facebook motto of &ldquo;move fast and break things&rdquo; won&rsquo;t work. Just speeding up technology won&rsquo;t be enough to keep us safe. We need sensible domestic regulation: supporting the green transition, raising safety requirements in biological labs, and ensuring that high-risk AI systems go through safety tests. We need international agreements, like the Paris agreement and the nuclear arms control treaties that Donald Trump &mdash; whose presidential campaign Thiel <a href="https://www.nytimes.com/2016/10/16/technology/peter-thiel-donald-j-trump.html">donated to</a> &mdash; ripped up.&nbsp;</p>
<h2 class="wp-block-heading">A libertarian approach to X-risk doesn’t work</h2>
<p>But Thiel doesn&rsquo;t seem to want this. He&rsquo;s an Ayn Rand libertarian. On ideological grounds, he doesn&rsquo;t believe that government action can help. He thinks regulation makes things worse. When asked by an Oxford Union audience member about how he &ldquo;would fix&rdquo; the UK&rsquo;s National Health Service (NHS), he said we need to get over our &ldquo;Stockholm syndrome&rdquo; as a country and privatize it already. Everyone is entitled to their own opinions, even billionaires. But it&rsquo;s a fringe ideology. And it&rsquo;s one that could do a huge amount of harm.</p>

<p>Thiel seemingly adopts this fend-for-yourself mentality in his planning too. He has long had a bolt-hole that he could escape to in case of societal collapse. In 2015, he bought a 193-acre plot of land (bigger than the Disneyland theme park in California) on the South Island of Aotearoa New Zealand for a <a href="https://www.nzherald.co.nz/nz/billionaire-peter-thiels-plans-for-luxury-lake-wanaka-lodge-rejected/TN6SH2XWWNIUIFLWJAM2JKK76Y/">reported</a> $13.5 million. In May 2022, he was <a href="https://www.theguardian.com/world/2022/may/03/setback-for-billionaire-peter-thiels-plan-to-build-luxury-lodge-in-remote-new-zealand">denied planning permission</a> to build a luxury lodge on the plot <a href="https://nypost.com/2022/08/19/peter-thiels-plans-for-dream-home-in-new-zealand-are-gone/">with space</a> for 24 guests, a theater lounge, a spa, and a &ldquo;meditation pod.&rdquo;</p>

<p>A mate of mine was traveling around the South Island a few years ago and had a pint at a pub. One of the locals pointed out a spot on a hill: &ldquo;See there? That&rsquo;s Peter Thiel&rsquo;s house. If anything goes bad, that&rsquo;s where we&rsquo;ll go for food and water.&rdquo;&nbsp;</p>

<p>It is a fantasy to think that existential risk can be reduced &mdash; or survived &mdash; just with individual action, the invisible hand of the market, and a &ldquo;go faster&rdquo; sign. It also needs collective action, wisdom and patience, and sensible and proportionate regulation.&nbsp;</p>

<p>That&rsquo;s the approach that the existential risk and effective altruism communities are taking. But that, unfortunately, is the approach that Thiel appears to disagree with.</p>

<p><em>Haydn Belfield has been </em><a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__twitter.com_HaydnBelfield&amp;d=DwMGaQ&amp;c=7MSjEE-cVgLCRHxk1P5PWg&amp;r=mTM0ruyioL3vFpq5GgUZftnxirsFoCe5-UZtYwnKki8&amp;m=fafNoDRYchktF6CmyLmwtRE1Dt1uZmJeDTqC_-94paSKyG8-WBBDyvzsCPMMphux&amp;s=8wpMbuaRp27Bl2ciP_ieN5e9ac-s3eWypajx66fqlFk&amp;e="><em>academic project manager</em></a> <em>at the University of Cambridge&rsquo;s Centre for the Study of Existential Risk (CSER) for the past six years. He is also an associate fellow at the Leverhulme Centre for the Future of Intelligence.</em></p>
						]]>
									</content>
			
					</entry>
	</feed>
