Are LLMs reasoning or reciting?
The impressive performance of recent language models across a wide range of tasks suggests that they possess a degree of abstract reasoning skills. Are these skills general and transferable, or specialized to specific tasks seen during pretraining? To disentangle these effects, we propose an evaluation framework based on “counterfactual” task variants that deviate from the default assumptions underlying standard tasks. Across a suite of 11 tasks, we observe nontrivial performance on the counterfactual variants, but nevertheless find that performance substantially and consistently degrades compared to the default conditions. This suggests that while current LMs may possess abstract task-solving skills to a degree, they often also rely on narrow, non-transferable procedures for task-solving. These results motivate a more careful interpretation of language model performance that teases apart these aspects of behavior.
Explosive growth from AI automation: A review of the arguments
We examine whether substantial AI automation could accelerate global economic growth by about an order of magnitude, akin to the economic growth effects of the Industrial Revolution. We identify three primary drivers for such growth: 1) the scalability of an AI labor force restoring a regime of increasing returns to scale, 2) the rapid expansion of an AI labor force, and 3) a massive increase in output from rapid automation occurring over a brief period of time. Against this backdrop, we evaluate nine counterarguments, including regulatory hurdles, production bottlenecks, alignment issues, and the pace of automation. We tentatively assess these arguments, finding most are unlikely deciders. We conclude that explosive growth seems plausible with AI capable of broadly substituting for human labor, but high confidence in this claim seems currently unwarranted. Key questions remain about the intensity of regulatory responses to AI, physical bottlenecks in production, the economic value of superhuman abilities, and the rate at which AI automation could occur.
https://arxiv.org/pdf/2309.11690.pdf
See also: Sam Hammond’s critical discussion.
And note that those most bullish on explosive growth typically only put it at 1/3 before 2100.
Frank Ramsey on the view from somewhere
My picture of the world is drawn in perspective, and not like a model to scale. The foreground is occupied by human beings and the stars are all as small as threepenny bits.
A Christmas message from Walter Benjamin
Humanity’s self-alienation has reached such a degree that it can experience its own destruction as an aesthetic pleasure of the first order.
via Curtis.
Tyler Cowen on Malthus
Whether or not you obsess over the particulars of overpopulation, Malthus’s theory is more broadly one of human pressures on the environment, and the lack of suitable equilibrating mechanisms at anything other than extremely high human costs.
The simplest version of Malthus is an account of how the world runs when all essential factors do not grow at the same rate, and in particular those growth rates diverge in a roughly consistent and sustained manner. At some point one of those factors becomes too scarce and the system crashes, leading to a plunge in living standards and possibly a population crash as well. In this sense Malthus is presenting a general rather than a special case, as it would seem that roughly equal rates of growth for the essential factors is the unusual setting, not the default setting.
[…]
For Malthus it could be said that the idea of equilibrium triumphs over that of progress.
[…]
It is also striking that Malthus was a major influence upon both Charles Darwin and Alfred Russel Wallace and their path breaking theories of evolution varying, and groups of people popping in and out of existence, helped them both formulate their theories of natural selection. Malthus thus helped to drive the very existence of modern evolutionary biology.