4 ms·

> earthquake predictions because we can't currently predict earthquakes? and if we can with AGI we would be saving lives?

dekhn 2y ago
Let's unpack what you're saying. You seem to believe that AGI would be able to predict something that we have been actively researching for decades. I do not see how AGI would suddenly be able to solve a problem that smart people have already been working on, unless the AGI was also "superintelligent" in ways that contributed. AGI proper is merely artificial intelligence, not necessarily superintelligence. Second, I question that earthquake prediction would make a big difference in saving lives. In particular, earthquake prediction would come with a wide range of risks. It would have to balance true positives with false positives- after a few "OMG everybody leave SF" that didn't pan out, people would just tune out the tsunami warning. Also, telling everybody to leave a large urban area is prone to massive other problems such as increased accidents, disruption of delivery of food and emergency services, and increased overall stress. Likely, more people would die in response (increased road accidents) to your intervention, than were saved. It seems far more likely that investing effort into identifying actual risks (such as infrastructure that is likely to fail, or individuals living in places that are highly prone to flooding), and addressing those issues through systemic improvements would be a much more useful way to use the machine learning technology we have. And even that isn't necessary; we know what the problems are, we just don't invest in improving it.
- colesantiago 2y ago
  lets pack this altogether. so you don't believe it would save lives? prevention is better than a cure and with AGI this would significantly a lot of people, not just SF. > smart people have already been working on, If they were so smart why haven't they found it yet? We now have weather prediction and protein folding now solved problems which were extremely difficult problems. I don't see how this doesn't apply to earthquake prediction, if it has to take AGI to solve it then so be it.
  - dekhn 2y ago
    I think it's unlikely having better earthquake prediction would save lives in an economically justifiable way. Neither weather prediction or protein folding are solved problems; that's the press misrepresenting the advances here.
    - colesantiago 2y ago
      I wouldn't have expected Google DeepMind, Science or Nature to be part of the press and of course they are misrepresenting advances in protein folding and super accurate weather prediction with little to no case studies or breakthroughs in this area which are better than what humans can do. https://deepmind.google/discover/blog/gencast-predicts-weather-and-the-risks-of-extreme-conditions-with-sota-accuracy/ https://deepmind.google/discover/blog/gencast-predicts-weath... https://deepmind.google/technologies/alphafold/ https://deepmind.google/technologies/alphafold/ https://deepmind.google/discover/blog/gencast-predicts-weather-and-the-risks-of-extreme-conditions-with-sota-accuracy/ https://deepmind.google/discover/blog/gencast-predicts-weath... https://www.science.org/stoken/author-tokens/ST-1550/full https://www.science.org/stoken/author-tokens/ST-1550/full https://www.nature.com/articles/s41586-024-08252-9 https://www.nature.com/articles/s41586-024-08252-9 But whatever you say.

dekhn 2y ago

Let's unpack what you're saying. You seem to believe that AGI would be able to predict something that we have been actively researching for decades. I do not see how AGI would suddenly be able to solve a problem that smart people have already been working on, unless the AGI was also "superintelligent" in ways that contributed. AGI proper is merely artificial intelligence, not necessarily superintelligence. Second, I question that earthquake prediction would make a big difference in saving lives. In particular, earthquake prediction would come with a wide range of risks. It would have to balance true positives with false positives- after a few "OMG everybody leave SF" that didn't pan out, people would just tune out the tsunami warning. Also, telling everybody to leave a large urban area is prone to massive other problems such as increased accidents, disruption of delivery of food and emergency services, and increased overall stress. Likely, more people would die in response (increased road accidents) to your intervention, than were saved. It seems far more likely that investing effort into identifying actual risks (such as infrastructure that is likely to fail, or individuals living in places that are highly prone to flooding), and addressing those issues through systemic improvements would be a much more useful way to use the machine learning technology we have. And even that isn't necessary; we know what the problems are, we just don't invest in improving it.

colesantiago 2y ago
lets pack this altogether. so you don't believe it would save lives? prevention is better than a cure and with AGI this would significantly a lot of people, not just SF. > smart people have already been working on, If they were so smart why haven't they found it yet? We now have weather prediction and protein folding now solved problems which were extremely difficult problems. I don't see how this doesn't apply to earthquake prediction, if it has to take AGI to solve it then so be it.
- dekhn 2y ago
  I think it's unlikely having better earthquake prediction would save lives in an economically justifiable way. Neither weather prediction or protein folding are solved problems; that's the press misrepresenting the advances here.
  - colesantiago 2y ago
    I wouldn't have expected Google DeepMind, Science or Nature to be part of the press and of course they are misrepresenting advances in protein folding and super accurate weather prediction with little to no case studies or breakthroughs in this area which are better than what humans can do. https://deepmind.google/discover/blog/gencast-predicts-weather-and-the-risks-of-extreme-conditions-with-sota-accuracy/ https://deepmind.google/discover/blog/gencast-predicts-weath... https://deepmind.google/technologies/alphafold/ https://deepmind.google/technologies/alphafold/ https://deepmind.google/discover/blog/gencast-predicts-weather-and-the-risks-of-extreme-conditions-with-sota-accuracy/ https://deepmind.google/discover/blog/gencast-predicts-weath... https://www.science.org/stoken/author-tokens/ST-1550/full https://www.science.org/stoken/author-tokens/ST-1550/full https://www.nature.com/articles/s41586-024-08252-9 https://www.nature.com/articles/s41586-024-08252-9 But whatever you say.