Linus believes that the current AI boom is a massive wealth transfer from software companies to hardware companies.

“You know, there there's been a lot of economic focused people talking recently about how it's an interesting wealth transfer from these like software companies and software focused investors to hardware companies.”

Luke believes that LLMs are not the path to AGI and that the fundamental idea of predicting the next token will not get us to true machine intelligence.

“But I just think the the fundamental idea of predicting the next thing is not going to get us to AGI.”

AI models currently suffer from a 'sameness' issue because they are essentially averages of their training data, limiting their long-term utility for creative applications.

“every video looks the same and AI is the average of averages right? So if you think about like all the emails that have been written by chat GPT the images the videos they all have a feeling of sameness because the model is just outputting sort of the average of what it's trained on”