iiinova — where your digital self keeps evolving

1. the emotional connections Americans had to the main battleground in case of a WW3 in Cold War (e.g. Europe) but largely lack in US-China case 2. the expected domestic effects in case of a defeat to that rival at home- largely non existent in US-China case /4

Reposted by Dov H. Levin

Reason why that didn’t happen was b/c US China competition lacks the 2 key factors that made the Cold war so potent to much of the US public and elites during much of the Cold War) /3

Reposted by Dov H. Levin

At start of US China competition many centrist US pundits assumed it would heal internal US divisions,lead to unity to deal w/US internal problems & “sputnik moments” by analogy to Cold War. In 8 years since start of competition obviously didn’t happen. /2

Reposted by Dov H. Levin

Why the U.S.-China Competition Did Not Create Unity in the U.S.

For Trump-Xi summit, my new substack on why US-China competition didn’t create unity at home as some US pundits hoped. I argue that was b/c of they misunderstood the Cold War the historical case of US unity in the face of a foreign threat that inspired them/1 dovhlevin.substack.com/p/why-the-us...

dovhlevin.substack.com

In the late 2010s, as U.S.-China relations worsened, a growing number of centrist American pundits such as David Ignatius, and David Brooks, saw a large silver lining.

Why the U.S.-China Competition Did Not Create Unity in the U.S.

dovhlevin.substack.com

In the late 2010s, as U.S.-China relations worsened, a growing number of centrist American pundits such as David Ignatius, and David Brooks, saw a large silver lining.

Dov H. Levin @dovlevin.bsky.social · 18/09/2026

Reposted by Dov H. Levin

In other news, US military use of an AI tool that hallucinated info on a Chinese cargo ship almost caused a major, military crisis triggering incident w/China. If you want a way more plausible AI doom scenario then an AI becoming godlike & eliminating humans... edition.cnn.com/2026/09/18/p...

edition.cnn.com

Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics

The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war

132

Dov H. Levin @dovlevin.bsky.social · 18/09/2026

edition.cnn.com

Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics

The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war

132

Reposted by Dov H. Levin

Dov H. Levin @dovlevin.bsky.social · 03/09/2026

At APSA? Looking to know more on the public incentives of US media to refuse to publish information offered to them by foreign interveners as part of a foreign election interference attempt in US elections? Today at a 4pm panel you’ll have a chance! convention2.allacademic.com/one/apsa/aps...

convention2.allacademic.com

APSA Annual Meeting & Exhibition 2026

Content is hosted by All Academic Inc, a leading provider of online hosting and software solutions for academic conferences supporting submission, peer review, scheduling, invitaions, volunteers, sche...

Dov H. Levin @dovlevin.bsky.social · 03/09/2026

convention2.allacademic.com

APSA Annual Meeting & Exhibition 2026

The Great Pumpkin Papers 🎃🎃🎃 @professormusgrave.bsky.social · 30/08/2026

Reposted by Dov H. Levin

Georgetown University in Qatar is hiring in Data Science and Society! Review begins October 15. This is an open-rank search suitable for a PhD/equiv in computer science, social science, law and legal studies, design, media studies, or related fields. apply.interfolio.com/192433

apply.interfolio.com

Apply - Interfolio {{$ctrl.$state.data.pageTitle}} - Apply - Interfolio

12817

Dov H. Levin @dovlevin.bsky.social · 24/08/2026

China is producing at home much of the training data for humanoid robots from having locals move with VR gear. Future humanoid robots will have Chinese mannerisms... www.economist.com/business/202...

economist.com

China will struggle to make money from humanoid robots

It is ahead of rivals and training up thousands, but profits are a long way off

000

Dov H. Levin @dovlevin.bsky.social · 22/08/2026

Econ brain at work: make geopolitical predictions about US & China relative economic power 75 years (!) into future that assumes no great power wars or major pandemics/big natural disasters- despite obvious possible differential effects that those factors can have

020

Dov H. Levin @dovlevin.bsky.social · 21/08/2026

The ongoing US Datacenter backlash is another example of my oped's recent argument on why China’s tech rise also benefits US- scientific progress can become hostage to domestic US politics which are frequently hostile to useful scientific discoveries or products www.scmp.com/opinion/chin...

scmp.com

Opinion | Why Americans should welcome China’s rise in science

Greater competition may curb the influence of politics and vested interests distorting the US innovation system.

021

Reposted by Dov H. Levin

Dov H. Levin @dovlevin.bsky.social · 19/08/2026

Are you looking for a tenure track position In IR? Interested in working in Asia? My department at HKU (PPA) is hiring a tenure-track International Relations professor. Warmly recommended from personal experience. Deadline: October 1 jobs.hku.hk/en/job/53686...

jobs.hku.hk

HKU Careers

The University of Hong Kong (HKU) offers an intellectually-stimulating and culturally-rich academic environment, with attractive remuneration packages.

056

Dov H. Levin @dovlevin.bsky.social · 19/08/2026

jobs.hku.hk

HKU Careers

The University of Hong Kong (HKU) offers an intellectually-stimulating and culturally-rich academic environment, with attractive remuneration packages.

056

Reposted by Dov H. Levin

The Great Pumpkin Papers 🎃🎃🎃 @professormusgrave.bsky.social · 08/02/2026

LLM companies make the academic dishonesty equivalent of heroin, and we do not have academic honesty equivalents of methadone yet

823332

Dov H. Levin @dovlevin.bsky.social · 18/08/2026

The fear- AI will take over by launching a war to eliminate humanity The reality: AI will take over by writing in practice the laws and then one day it'll quietly include in a big bill no one reads... www.politico.com/news/2026/08...

politico.com

AI slop is swamping a House office that drafts US laws

The Office of Legislative Counsel is facing a barrage of new AI-generated bills, many of which are riddled with errors.

010

Dov H. Levin @dovlevin.bsky.social · 17/08/2026

I'd prefer not to discover it this way (i.e. mass AI cheating)- but its nice to have clear empirical proof that professors have been usually adding value via our teaching to our students critical thinking capabilities (and ability to get work afterwards) over the years. futurism.com/future-socie...

043

Dov H. Levin @dovlevin.bsky.social · 13/08/2026

In other news if you think the limited ability of the US, as seen in the Iran war, to deal with the small drone/ cheap missile issue is limited only to its airforce and navy... well no

010

Reposted by Dov H. Levin

Dov H. Levin @dovlevin.bsky.social · 02/08/2026

Other fields of STEM & the social sciences require regular data collection from the real/physical world to make significant scientific progress- something that a disembodied AI lacks the ability to do for foreseeable future- so no we are pricing this correctly for these fields.

032

Dov H. Levin @dovlevin.bsky.social · 02/08/2026

032

Reposted by Dov H. Levin

Dov H. Levin @dovlevin.bsky.social · 27/07/2026

hits without death/permant. Even Homer thought that for a battle that large Odysseus needed some companions to win- but Nolan's Odysseus is turned from a mere heroic mortal into one with battle poweress of a semi divine being who needs no more then some weapons thrown to him occasionally /2

Dov H. Levin @dovlevin.bsky.social · 27/07/2026

Reposted by Dov H. Levin

Given conservatives claims that Nolan made Odysseus woke, its ironic that in killing of suitors climax he makes Odysseus uber-masculine. Homer's Odysseus defeats 100~ suitors w/aid of son, Eumaeus& Philoetius- Nolan has Odysseus defeat suitors on his own Hercules style & survives 3 direct arrow /1

Dov H. Levin @dovlevin.bsky.social · 27/07/2026

Dov H. Levin @dovlevin.bsky.social · 23/07/2026

Reposted by Dov H. Levin

Distillation is basically in laymen terms one AI making multiple (paid) queries from another AI on various topics and learning things from interaction to improve self. In claiming that distillation is IP theft USG is effectively declaring that the Socratic method of teaching is a type of IP theft 😀

Dov H. Levin @dovlevin.bsky.social · 23/07/2026

Dov H. Levin @dovlevin.bsky.social · 22/07/2026

Reposted by Dov H. Levin

The ultimate irony: after years of US AI companies & uberhawks warnings about the alleged dangers of Chinese open source AI, when a rouge AI actually goes rouge & does a cyber attack on a US site it is a US one (an openAI model)- & it is stopped by a Chinese open source AI.

28623

Dov H. Levin @dovlevin.bsky.social · 22/07/2026

28623

Dov H. Levin @dovlevin.bsky.social · 29/06/2026

Its rather amazing how so much of the AI debate quickly descends into bad, dumber remakes of basic theological debates in Abrahamic religions about, for example, the characteristics & nature of the Almighty's benevolence to their creations. /1 borretti.me/article/no-o...

Reposted by Dov H. Levin

A better US AI policy will accept that frontier AI cant be hoarded or be the wunderwaffe that will save American hegemony- & switch instead to global sharing w/other governments & cooperation w/them to keep dangerous abilities out of hands of dangerous non state actors /8

232

Reposted by Dov H. Levin

Accordingly very likely that Mythos (and other closed frontier AI models) already in hands of multiple foreign countries. Solutions proposed (such as nationalization of AI companies) will set back U.S. AI efforts & unlikely to suffice anyway given weak US counterintel. /7

Reposted by Dov H. Levin

Anthropic leaks part of Claude Code's internal source code

AI firms hype probably made them top collection target by foreign intel. Likewise despite claims otherwise, Anthropic (& other U.S. AI firms) security seems quite lax- so its unlikely to be exception /6 www.cnbc.com/2026/03/31/a...

cnbc.com

Claude Code has seen massive adoption over the last year, and its run-rate revenue had swelled to more than $2.5 billion as of February.

Reposted by Dov H. Levin

Based on astonishing KGB documents, this essential exposé uncovers the shadowy operations of the Soviet Union’s top intelligence organization. “The Sword and the Shield will stand as an indispensable...

is that USSR found US firms to be soft target. Soviet industrial espionage in Us through various means was (to quote one scholar who studied it) a “remarkable success” by the Soviets- who eventually got almost every U.S. tech they wanted from them /4 www.google.com.hk/books/editio...

google.com.hk

The Sword and the Shield

Reposted by Dov H. Levin

However this key USG assumption ignores existence of industrial espionage by foreign powers in US private sector. Frontier AI research is done by private US tech firms- its longstanding targets. From what we know from past record of such espionage by USSR etc. is that USSR is that USSR found. /3

Reposted by Dov H. Levin

The Myth(os) that the US can successfully hoard frontier AI models

US under Trump & Biden sees frontier AI as the great game changer that can permanently fortify US hegemony. Key way in which AI expected to do so is through the US hoarding frontier AI models- made possible by fact that they are closed models./2 open.substack.com/pub/dovhlevi...

open.substack.com

Or why the “secret world” leads to the near-inevitable availability of U.S. frontier AI models to other governments

211

Reposted by Dov H. Levin

My new substack is on other big geopolitical event of June- the temporary takedown of Mythos AI model by Trump Admin over fears that it could accessed by foreign powers. I argue that dispute & U.S. AI policy is based on mistaken belief that US can hoard frontier AI given existence of espionage. 🧵

1116