AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Cost of AGI in the Brain
You can't buy a computer with 100 billion bytes of memory yet, but by the time you can, should we assume it will have AGI? I think this is a problem. Currently AGI phase is why, because in our another paper, we compare the scalability and the cost of the AGI system. Although the currently large language model such as the Church GPD and GPT4 have very powerful performance, but their cosy is also very large. If we create more large neural system, we will pay much for them. And sometimes we can't afford it. But for the brain, although we have thousands of hundreds of neurons, but they are not so expensive. So