Language Log

Daron Acemoglu, "The Simple Macroeconomics of AI":

ABSTRACT: This paper evaluates claims about the large macroeconomic implications of new advances in AI. It starts from a task-based model of AI’s effects, working through automation and task complementarities. It establishes that, so long as AI’s microeconomic effects are driven by cost savings/productivity improvements at the task level, its macroeconomic consequences will be given by a version of Hulten’s theorem: GDP and aggregate productivity gains can be estimated by what fraction of tasks are impacted and average task-level cost savings. Using existing estimates on exposure to AI and productivity improvements at the task level, these macroeconomic effects appear nontrivial but modest—no more than a 0.71% increase in total factor productivity over 10 years. The paper then argues that even these estimates could be exaggerated, because early evidence is from easy-to-learn tasks, whereas some of the future effects will come from hard-to-learn tasks, where there are many context-dependent factors affecting decision-making and no objective outcome measures from which to learn successful performance. Consequently, predicted TFP gains over the next 10 years are even more modest and are predicted to be less than 0.55%. I also explore AI’s wage and inequality effects. I show theoretically that even when AI improves the productivity of low-skill workers in certain tasks (without creating new tasks for them), this may increase rather than reduce inequality. Empirically, I find that AI advances are unlikely to increase inequality as much as previous automation technologies because their impact is more equally distributed across demographic groups, but there is also no evidence that AI will reduce labor income inequality. AI is also predicted to widen the gap between capital and labor income. Finally, some of the new tasks created by AI may have negative social value (such as design of algorithms for online manipulation), and I discuss how to incorporate the macroeconomic effects of new tasks that may have negative social value.

Samuel Bray, "Cruel AND Unusual?", Reason 4/21/2024:

On Monday, the Supreme Court will hear argument in an Eighth Amendment case, City of Grants Pass, Oregon v. Johnson. One thing I will be watching for is whether the justices in their questions treat "cruel and unusual" as two separate requirements, or as one.

The Eighth Amendment (to the U.S. Consitution) says that "Excessive bail shall not be required, nor excessive fines imposed, nor cruel and unusual punishments inflicted."

And the issue in the cited Supreme Court case is "Whether the enforcement of generally applicable laws regulating camping on public property constitutes “cruel and unusual punishment” prohibited by the Eighth Amendment." (More here, here, and elsewhere…)

I got an echo of Saturday's post about chatbot pals, from an article yesterday in Intelligencer — John Herrman, "Meta’s AI Needs to Speak With You" ("The company is putting chatbots everywhere so you don’t go anywhere"):

Meta has an idea: Instead of ever leaving its apps, why not stay and chat with a bot? This past week, Mark Zuckerberg announced an update to Meta’s AI models, claiming that, in some respects, they were now among the most capable in the industry. He outlined his company’s plans to pursue AGI, or Artificial General Intelligence, and made some more specific predictions: “By the end of the decade, I think lots of people will talk to AIs frequently throughout the day, using smart glasses like what we’re building with Ray-Ban Meta.”

Is monolingualism a normal, natural, necessary state of affairs for human beings?

Can you imagine a world in which there were only one language?  How is that even possible?

These are questions that come to mind after reading Gina Anne Tam's deeply thought provoking "Mandarin Hegemony: The Past and Future of Linguistic Hierarchies in China", pulse (4/18/24).

Tam begins with a gripping, hard-hitting scene that we at Language Log were already well aware of last fall:  "Speak Mandarin, not Cantonese, even in Macau" (10/31/23).  Here are the opening paragraphs of her article:

At a concert in Macau in the autumn of 2023, Cantopop superstar Eason Chan used an interlude to talk about his songwriting process. Suddenly, shouts from the audience interrupted his soliloquy, as a few fans demanded that he shift from speaking in his native Cantonese, the majority language in Macau, to Mandarin, the Chinese national language. Chan stopped and quickly launched into a multilingual lecture, reprimanding those who deigned to tell him what to speak. In English, Cantonese, Mandarin, and Thai, he defended multilingualism for the freedom it grants: ‘I love speaking in whatever way and language I want’ (Huang 2023).

Daniel Deutsch wrote:

I had to read this headline a couple of times.

"The pandemic cost 7 million lives, but talks to prevent a repeat stall"

Is the pandemic talking? Is it trying to prevent a repeat stall?

A new feature at the Dalí Museum in St. Petersburg FL:

This is one of the most novel theories on the Voynich manuscript (Beinecke MS408; early 15th c.) that I've ever encountered, and there are many.

The Voynich Manuscript, Dr Johannes Hartlieb and the Encipherment of Women’s Secrets, by Keagan Brewer and Michelle L Lewis, Social History of Medicine, hkad099 (22 March 2024)

Keywords:  Voynich manuscript, Dr Johannes Hartlieb, women’s secrets, sex, gynaecology

Worries about future applications of AI technology focus on many things, including new forms of automation replacing human workers, realistic deepfake media spreading disinformation, and mass killing by autonomous military machines. But there's something happening already that hasn't gotten as much commentary: chatbots designed to be pals or romantic connections.

Most Americans probably know a few Japanese loanwords, especially those who were alive in the two or three decades after WWII, when so many terms from Japan entered the English language — kamikaze, banzai, bonsai, origami, and so forth — with soldiers returning from the war in the Far East.

In the recent two or three decades, Japanese words, continued to enter English but from different avenues — anime, manga, sudoku, karaoke, etc.

In a comment on yesterday's post ("High vowel lenition/devoicing in French"), carveuir wrote:

Ha! As a final-year undergraduate in 2015, I mentioned having come across devoicing of the second /i/ in "université" to my French linguistics tutor and he didn't believe me. Finally I've been vindicated.

My impression is that this is common and perhaps almost categorical in Québecois vernacular, but more gradient (or maybe I should say less complete?) in Parisian French. So I looked from some examples of the word université in a collection of transcribed radio broadcasts and political speeches from France. And I found a few, all of which were consistent with my impression. So my recent series of French phonetic anecdotes continues below.

One of the students in my class — all from China — hit her elbow on the edge of her desk and grimaced.  I asked her, "Did you hit your crazy bone?"

She didn't know what I meant, and none of the other students in the class knew either.  I explained what "hit my crazy bone" signifies (see below for a physiological note), and the entire class thought it was funny.  Lots of giggling and laughing.

One needs to be careful when using a phrasal verb that has a wide range of possible meanings.  For example, if you're corresponding with a woman who travels a lot and you comment, wishing to commend her mobility, "You sure do get around a lot", she may be offended and retort, "Are you saying that I'm sexually promiscuous?"

On a trip to Québec in the 1970s, I asked a passerby for directions (in French), and he gave me an answer that at first I thought was in Polish or some other Slavic language unknown to me. He also pointed to the visible train-track overpass a couple of blocks away, and waved his arm to indicate a right turn, so I got the meaning from his gestures. And after a bit, I realized that his opening phrase, which I heard as something like

[tvɐ.drɛk.tʃsko.trɐk]

was a Québecois vernacular version of "tu vas direct jusqu'au trac", with the [i] and [y] vowels deleted (and the  initial /ʒ/ of "jusqu'au" devoiced). I asked a Canadian colleague about it, and was told that the deletion of high vowels was known to linguists in Francophone Canada, but (as far as he knew at that time) had not been documented.