Google DeepMind has exposed obvious

shafi876@gmail.

Not so with Google. Over the years, DeepMind, Google's AI R&D institution, has used reinforcement learning and deep learning to solve various artificial intelligence problems, and has accumulated a wide range of technologies. For example, the groundbreaking AlphaGo, AlphaFold that completely changed biology, and NLP technologies such as Transformer. This is equivalent to two drivers preparing a car for a competition. OpenAI selects a venue for AGI, such as a formula racing car, and then develops and manufactures the model with language as the


core, and determines the structure, length, width, and Guangdong Mobile Phone Number List length of the car (model). Engines, cylinders, etc. are optimized and transformed (engineered). If he loses every time, he might be able to pay attention to whether there are problems from the sources such as horse breeds, horse pens, and fodder. Going back to the source, Google and Google's DeepMind was not sure which car could end the AGI competition, and it had many technical tools at its disposal, so it tried building formula cars, sports cars, and motorcycles. There is no inherent advantage





or disadvantage between the two routes. However, with the intelligence emergence of large language models, it has been proved that the technical route chosen by OpenAI is more promising to achieve AGI, and the technical route of  shortcomings: The direction is scattered and the cost is high. Pan-innovation invested in various technical directions consumes a lot of money, and the commercial conflict between