What the techbro fascist makers of generative AI are trying to accomplish is to create a world 100% driven by statistics and their desired 'models of existence'. The people behind AI are poor statisticians, however, as they dont understand the limits of the field. Statistics are only 100% relevant in the exact time and place they were collected; the further away you get from your collection, the more limited they become if you are trying to apply them to the present. This is particularly true in human sciences fields (economics, medicine, government, culture, psychology). This is why AI will always be extremely energy intensive, because its makers have to perpetually sample their dataset by adding new context, while simultaneously resampling all old material, because the algorithym has to reprocess old data in light of new context.
The main problem is that the data AI uses is generated by humans. Collecting and parsing data is extremely time consuming and difficult work for good reason, requiring the expertise of scientists who understand the limits of statistics and the pitfalls of study design. AI is trying to replace human scientists and bureaucrats using data generated by scientists and bureaucrats. But generative AI and its creators dont understand the scientific process - they just process a bunch of data and spit out the most statistically likely outcome (something we were already capable of before). By attempting to replace human scientists and experts, generative AI is destroying its own dataset, the very basis of its existence. In other words, it will effectively eat itself, while its ability to generate new relevant data will be severely hamstrung, as AI is not capable of 'understanding' anything, let alone applying the scientific process. #science #datascience #statistics #ai
Even a small monthly contribution makes a big difference. Support here: #^https://www.patreon.com/FediverseStatistics 
