• 2 Posts
  • 565 Comments
Joined 1 year ago
cake
Cake day: July 6th, 2023

help-circle













  • Here are the actual poll results which the article helpfully does not link to.

    Napolitan News surveys ask an initial question to determine the voter preference for each candidate. Then, a follow-up question is asked of uncommitted voters to see which candidate they are leaning towards. The results are then reported “with leaners.”

    On the initial ask– the number without leaners– it was Trump 50%, Harris 47%.

    This Napolitan News Service survey of 774 Likely Voters was conducted online by Scott Rasmussen on September 25-27, 2024. Field work for the survey was conducted by RMG Research, Inc. and has a margin of error of +/- 3.5.

    I think articles like this based on a single poll which appears to be an outlier are uninformative, but I guess they get clicks.




  • I disagree with you, because a modern human could offer the people of the distant past (with their far less advanced technology) solutions to their problems which would seem miraculous to them. Things that they thought were impossible would be easy for the modern human. The computer may do the same for us, with a solution to climate change that would be, as you put it, magically ecological.

    With that said, the computer wouldn’t be giving humans suggestions. It would be the one in charge. Imagine a group of chimpanzees that somehow create a modern human. (Not a naked guy with nothing, but rather someone with all the knowledge we have now.) That human isn’t going to limit himself to answering questions for very long. This isn’t a perfect analogy because chimpanzees don’t comprehend language, but if a human with a brain just 3.5 times the size of a chimpanzee’s can do so much more than a chimpanzee, a computer with calculational capability orders of magnitude greater than a human’s could be a god compared to us. (The critical thing is to make it a loving god; humans haven’t been good to chimpanzees.)


  • I don’t think you’re imagining the same thing they are when you hear the word “AI”. They’re not imagining a computer that prints out a new idea that is about as good as the ideas that humans have come up with. Even that would be amazing (it would mean that a computer could do science and engineering about as well as a human) but they’re imagining a computer that’s better than any human. Better at everything. It would be the end of the world as we know it, and perhaps the start of something much better. In any case, climate change wouldn’t be our problem anymore.


  • The article compares coal and natural gas based on thermal energy and does not take into account the greater efficiency of natural-gas power plants. According to Yale the efficiency of a coal power plant is 32% and that of a natural gas power plant is 44%. This means that to generate the same amount of electricity, you need 38% more thermal energy from coal than you would from natural gas. I’m surprised that the author neglects this given his focus on performing a full lifecycle assessment.

    Natural gas becomes approximately equal to coal after efficiency is corrected for, using the author’s GWP20 approach. GWP20 means that the effect of global warming is calculated for a 20 year timescale. The author argues that this is the appropriate timescale to use, but he also presents data for the more conventional GWP100 approach, and when this data is adjusted for efficiency, coal is about 25% worse than natural gas.

    I’m not an expert so I can’t speak authoritatively about GWP20 vs GWP100 but I suspect GWP100 is more appropriate in this case. Carbon dioxide is a stable gas but methane degrades fairly quickly. Its lifetime in the atmosphere is approximately 10 years. This means that while a molecule of carbon dioxide can keep trapping heat forever, a molecule of methane will trap only a finite amount of heat. This effect is underestimated using GWP20.

    Edit: Also the Guardian shouldn’t be calling this a “major study”. It’s one guy doing some fairly basic math and publishing in a journal that isn’t particularly prestigious.