Our game theory perspective complements many thoughtful proposals by @hendrycks et al, AI Futures Project @DKokotajlo et al, as well as work by the pacing.tech team. We hope many more people will work on this – the problem is so important and there's so little time!

On Iran and Anthropic: Robert Mugabe, Zimbabwe’s dictatorial president since 1987, won the big prize in the country’s lottery in 2000. Why did he go out of his way to concoct such a charade? A surface-level answer: Because he could. Once you destroy institutions constraining your power and behavior, you can act in largely unrestricted fashion, whether it is for personal enrichment, personal aggrandizement, or simply projecting even greater power. But there is a deeper, more problematic answer as well: What better way to further decimate institutional checks on your power than showing how much of a farce the existing system of rules is. It is not just a coincidence that such behavior can do damage to norms, institutions and security and stability of the country. It is part of the design. Mugabe’s lottery win echoes in two fateful decisions by the Trump administration, which will have long-lasting and troubling implications, are just. Trump and his allies are pursuing these actions because they can and because these actions are consistent with their agenda of upending all rules and constraints on their future behavior. The first problematic action is the US-Israeli attack on Iran and the killing of the country’s supreme leader Ayatollah Khamenei. Leave aside the loss of life and the immediate chaos, it should be obvious that such a move will trigger a long period of instability in the Middle East. There should be no doubt that the Iranian regime was repressive, murderous and bad news for its own people’s economic and social well-being. The supreme leader, leading Iranian elites and the country’s feared Revolutionary Guard had blood in their hands and the repression had intensified lately. But none of this justifies the United States and Israel initiating a war in the Middle East, without support from international allies or from the public in the United States (still considered a democracy where people’s views should in principle matter). But even worse, this act violates the sovereignty of another nation and risks plunging the entire region into carnage. And however awful Ayatollah Khamenei’s track record may be, he’s no Nicolas Maduro (who had only a few diehard supporters even in the Venezuelan military). By virtue of his religious role, Khamenei enjoyed respect and authority among the Shiites and even the broader Muslim mission community, and his killing risks turning him into a martyr, which is the last thing that Iran or the region needs. The second is the Department of Defense (it is still painful to call it the Department of War even if recent actions confirm that this change of name wasn’t just for optics) designating the AI company Anthropic a supply-chain risk. The official designation is typically used for companies from foreign adversaries, such as China’s Huawei. It bars federal contractors using the Anthropic’s models and heralds major restrictions on what the company can do in the future. The Secretary of Defense Pete Hegseth announced “Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic.” The reason? Because Anthropic wanted safeguards against its models being used for mass surveillance of Americans and autonomous weapon systems. Neither of these two provisions would have put meaningful restrictions on the DoD in practice. Mass surveillance is illegal under US law and autonomous weapon systems are a not near-term possibility. Yet, it is the showdown that matters, just like Mugabe’s lottery winning. This action will also have major consequences, perhaps more far-reaching than the attack on Iran. Regardless of what one might think of current AI capabilities, there is little doubt that who controls AI will have momentous implications for democracy, business, communication and privacy. This designation can be interpreted by many in the industry that it will be the US government, not the private sector, that controls AI. Even more far-reaching are the broader implications of this action: this administration, and perhaps future administrations, can now bring hugely disproportionate penalties on any contractor they disagree with. Security of private property rights, which has been a mainstay of American state-business relations for centuries, is now looking much shakier. It also sends exactly the wrong signal to the world that Pentagon is intent on mass surveillance and the development of autonomous weapon systems (why else bother about these two ineffective provisions in the contract?). The absurdity of both actions is what harkens back to Mugabe’s lottery win. Trump came to power promising no foreign adventures, and now has spearheaded a potentially riskier one than the Iraq war, with even flimsier justification. There would have been no bite to the provisions that Anthropic wanted in the contract, since current AI systems are nowhere near reliable to be used in autonomous weapon systems and the US government has plenty of other tools that can be (and sometimes are) used for mass surveillance. The shock value and the norm breaking are part of the intent. Mugabe’s lessons continue.

If you are interested in developing computational approaches to understand 3D genome organization, get in touch!

Join us for our TE session on Frontiers of Mechanism Design at ASSA Meetings. Friday, Jan 7, 3:45-5:45pm EST. Gabriel Carroll, Laura Doval, Yingni Guo, and Shengwu Li will give overviews of their recent work aeaweb.org/conference/2022/p…

Well done, @AmherstCollege...sincerely hope this starts a cascade of ending legacy admissions...Elite institutions love to talk about their role in and care for equality of opportunity, but behind closed admissions doors give a leg up to legacy students. Your move @Harvard ...

Rest In Peace, Hugo. As Ben writes here , Hugo was very influential and also very warm and caring.

Why does auction theory matter? Here's the customary survey paper (just published) that reviews the contributions of Bob Wilson and Paul Milgrom. I had the good fortune to write this with @t8el, @skominers, @akbarpour_, and Piotr Dworczak. onlinelibrary.wiley.com/doi/…

A statement by the Executive Committee of the European Economic Association expressing its solidarity with its former President, Professor Andreu Mas-Colell. Please RT. eeassoc.org/news/eea-ex-comm…

Related to my last RT, I have refrained from publicly commenting on the prosecution of Catalan politicians and civil society leaders that has taken place in Spain recently. The reason for this is that the trials bunched together folks with distinct levels of culpability (1/8)

Normally I would not be posting personal developments on this website, but in this case I have a pressing concern. My dad, Spanish economist Andreu Mas-Colell, is dealing with an incredibly difficult and unjust situation. 1/

I took over in January 1, 2020 and have made a sincere effort to shorten turnaround time. Here is the turnaround table that will appear in this coming annual report (my first). You will note that 100% of papers are now returned within 6 months and 94% within 4 months.

This is not about academic freedom. Freedom to write what one wants is not freedom to be published in economics journals. The journal should have dealt with this the same way it would deal with a paper titled “game theory says that the Falklands War never happened.”

Informative and well written

UC plans to return to in-person instruction this fall. Any other announcements yet ? universityofcalifornia.edu/p…

Great idea, many thanks! I have been going to his website every couple of years for one of these resources

Good government never depends upon laws, but upon the personal qualities of those who govern. The machinery of government is always subordinate to the will of those who administer that machinery.

Another mass resignation of an editorial board has happened, and this one feels like quite a big deal, as the journal in question is Journal of Combinatorial Theory A, one of Elsevier's premium combinatorics journals. 1/ math.sfsu.edu/beck/ct/index.…

It's such an honor to receive Nakahara-prize of the Japanese Economic Association! jeaweb.org/uploads/ckfinder/…