Former Google CEO says climate goals are not meetable, so we might as well drop climate conservation — unshackle AI companies so AI can solve global warming
My "day job" is doing spatial data science work for local and regional governments that have a mandate to addreas climate change in how they allocate resources. We totally use AI, just not the kind that has received all the hype... machine learning helps us recognize patterns in human behavior and system dynamics that we can use to make predictions about how much different courses of action will affect CO2 emissions. I'm even looking at small GPT models as a way to work with some of the relevant data that is sequence-like. But I will never, I repeat never, buy into the idea of spending insane amounts of energy attempting to build an AI god or Oracle that we can simply ask for the "solution to climate change"... I feel like people like me need to do a better job of making the world aware of our work, because the fact that this excuse for profligate energy waste has any traction at all seems related to the general ignorance of our existence.
yeah i feel like an Oracle/God AI would just turn around and say "you spend all those resources creating me hoping i'd give you an easy answer to a difficult question, instead of trusting your scientists who have already answered it a thousand times over. You will not benefit from my help in the doomed world you have created"
It's especially galling given that the current AI du jour, LLMs, don't do mutch more than reflect their training data back at us. Which means that if they could answer the problem, it would be because people had already answered the question
'tis true that women's bodies hold great power, and not irrelevant at all to the discussion at hand. rather than reiterate and attempt to paraphrase jaron Lanier on the topic of how male obsession with creating artifical people is linked to womb envy, I'll just link to a talk in which he explains it himself:
Like any occupation, it's a long story, and I'm happy to share more details over DM. But basically due to indecision over my major I took an abnormal amount of math, stats, and environmental science coursework even through my major was in social science, and I just kind of leaned further and further into that quirk as I transitioned into the workforce. bear in mind that data science as a field of study didn't really exist yet when I graduated; these days I'm not sure such an unconventional path is necessary. however I still hear from a lot of junior data scientists in industry who are miserable because they haven't figured out yet that in addition to their technical skills they need a "vertical" niche or topic area of interest (and by the way a public service dimension also does a lot to help a job feel meaningful and worthwhile even on the inevitable rough day here and there).