Earthling: Ukraine rolls the dice

Plus: Robocop unplugged, Does AI get the joke?, how Biden helped Chinese protesters, and more!

This week Ukraine carried out drone strikes on Russian territory, killing three servicemen and damaging parked military aircraft. The New York Times drew flak for referring to the attacks as an “escalation,” but they did penetrate Russia much more deeply than Ukraine’s several previous across-the-border attacks. Two of the three airfields targeted were some 300 miles beyond Ukraine’s border, and one was 100 miles from Moscow.

At any rate, the Biden administration seemed worried about the attacks’ escalatory potential. Secretary of State Antony Blinken said the US “neither encouraged nor enabled” the attacks. And anonymous administration officials hastened to leak a story to the Wall Street Journal about how the Pentagon had secretly modified rocket systems sent to Ukraine to disable their long-range capabilities.

The American reaction reflects, in part, a divergence of perspective between Ukraine and the US: Ukraine (as President Zelensky has made clear) wants NATO to enter the war, and the US doesn’t. The Biden administration may have been signaling to Moscow that this week’s attacks don’t signify any expansion of NATO’s war aims and so don’t warrant Russian military or covert action beyond Ukraine’s borders.

But there’s also a subtler sense in which an attack like this could have escalatory consequences. The strikes, by making ordinary Russians feel that their homeland is threatened, could increase their support for Putin’s war, and thus their tolerance for any further troop mobilization he may deem necessary. It wouldn’t be the first time heightened threat perception generated a rally-’round-the-flag effect.

There are signs that Russian officials are trying to foster such a sense of threat. The Russian provinces of Kursk and Belgorod were reported this week to have begun forming territorial defense units that would guard against a supposed ground invasion. And, of course, Moscow has long cast the Ukraine war as an existential struggle against the West—against NATO and, in particular, against a hegemonic America.

Ukraine may be betting that attacks on Russian soil will help convince Russians that Putin miscalculated in invading Ukraine and is now failing to keep the homeland safe from the blowback. And certainly the attacks were an embarrassment for Putin.

The attacks were also, it should be said, against legitimate military targets. The two bases struck on Monday housed bomber squadrons that have helped devastate Ukraine’s energy infrastructure. But the main impact of the strikes is at the psychological, not the military, level, and it remains to be seen whether this Ukrainian tactic will work for or against Ukraine’s interests.

This week social media were abuzz over the latest AI marvel—ChatGPT, a creation of OpenAI, which also brought us the art-making DALL-E and the essay writing GPT-3. And ChatGPT is indeed impressive enough to raise questions about how we can be sure that an online conversation partner is human.

One way may be to probe the partner’s sense of humor. When data scientist Riley Goodside told the AI to write a Seinfeld scene, the results suggested that ChatGPT failed to detect that Seinfeld is supposed to be funny—or at least, that Seinfeld is supposed to be Seinfeld-level funny. To be fair, it did write “audience laughs” every once in a while—if in places where the audience wouldn’t have laughed.   

Meanwhile, Max Roser of Our World in Data looks at the remarkable progress computers and AI have made in their relatively brief time on Earth: 

Senator Bernie Sanders says he’ll try to force a vote—“hopefully next week”—on a bill that would block US support for Saudi Arabia’s war in Yemen, the Intercept reports. The conflict is currently on pause, but the peace is precarious; a ceasefire agreement between Saudi Arabia and Yemen’s Houthis expired in October. Backers of the bill hope to send Saudi Arabia the message that it doesn’t have carte blanche to restart hostilities. Congress passed a similar bill in 2019, but Donald Trump vetoed it.

This week the Chinese government announced a loosening of pandemic rules (including rules like mandatory hospitalization—which means infected Chinese can now quarantine at home). Beijing portrayed the changes as the result of a successful fight against the Covid virus, but the move seemed to come in response to widespread protests against China’s zero-Covid strategy and the oppressive lockdowns it entails.

And that wasn’t this week’s only story filed under “authoritarian government tries to mollify protesters.” The Iranian government announced it was disbanding the country’s “morality police.” That unit had detained the 22-year-old woman whose death in police custody sparked months of demonstrations. But Iran hasn’t rescinded the dress code rules, including the mandatory hijab, that the disbanded unit was charged with enforcing—much less made broader reforms some protesters have demanded.

President Biden—unintentionally, no doubt— turned these two stories into a case study in the efficacy of presidential rhetoric. He issued a pointed statement supporting the Iranian protests, but his comments on the Chinese protests were more restrained. So we have, as they say, a “control group” and a “treatment group”. And if any lesson can be drawn from this “experiment,” it’s this: