Good to see fossil fuel's contributing to electricity generation decrease. But shouldn't we want to see increasing electricity generation? It's a decade of horizontal, and electricity is still only a quarter of all energy consumption in the EU..
Absolutely, although I suspect many people underestimate the degree to which high energy costs also drive efficiency innovation. For example, the main thing tech companies with servers care about is performance per watt, and that’s where most of the gains have been in computing for the past ~15 years. A typical laptop lasts 5 times longer than it used to between charges, and not because the batteries have gotten 5 times better.
Not to say there aren’t more depressing reasons for energy consumption flatlining.
I think there is a problem with your reference to the Norwegian radiologists study.
You say that "Applying this across all screenings would reduce total radiologist workload at breast centers by 4.5 percent". Given the change is from using two human radiologists to one, this seems a very small reduction. It is so small for a couple of reasons.
First, the researchers assume that reviewing a scan takes only one minute - I'm not an expert, but this seems a remarkably short average time. This article, for example, suggests significantly longer time - https://pubmed.ncbi.nlm.nih.gov/38852330/
Second, the researchers seem to assume that AI could not do anything else apart from reviewing a scan. But it obviously could do other things, such as all the admin around writing up the assessment, notifying patients and other medics, arranging follow ups etc. So the time savings from using AI should be a lot more than 4.5%.
Of course, within a few years (or less), if it is not there already then AI will be superior to even the best human radiologist at reading a scan.
Good to see fossil fuel's contributing to electricity generation decrease. But shouldn't we want to see increasing electricity generation? It's a decade of horizontal, and electricity is still only a quarter of all energy consumption in the EU..
Absolutely, although I suspect many people underestimate the degree to which high energy costs also drive efficiency innovation. For example, the main thing tech companies with servers care about is performance per watt, and that’s where most of the gains have been in computing for the past ~15 years. A typical laptop lasts 5 times longer than it used to between charges, and not because the batteries have gotten 5 times better.
Not to say there aren’t more depressing reasons for energy consumption flatlining.
I think there is a problem with your reference to the Norwegian radiologists study.
You say that "Applying this across all screenings would reduce total radiologist workload at breast centers by 4.5 percent". Given the change is from using two human radiologists to one, this seems a very small reduction. It is so small for a couple of reasons.
First, the researchers assume that reviewing a scan takes only one minute - I'm not an expert, but this seems a remarkably short average time. This article, for example, suggests significantly longer time - https://pubmed.ncbi.nlm.nih.gov/38852330/
Second, the researchers seem to assume that AI could not do anything else apart from reviewing a scan. But it obviously could do other things, such as all the admin around writing up the assessment, notifying patients and other medics, arranging follow ups etc. So the time savings from using AI should be a lot more than 4.5%.
Of course, within a few years (or less), if it is not there already then AI will be superior to even the best human radiologist at reading a scan.