That's why I'm increasingly asking Claude to give me 3-paragraph summaries of articles that may contain interesting facts, but are also exercises in creative writing that I am unwilling to commit the time to. Here is an excerpt from the summary Claude gave me:
>The wager the title refers to is that the big labs cannot build everything on top of their own models. Their customers are increasingly sick of paying for tokens, feeding the labs their data, and effectively training the company that may one day replace them, a dynamic one founder describes as not wanting to spend his life drinking Anthropic and OpenAI's water. So Guo bets on the application layer and on the picks and shovels, like Baseten, which lets companies run their own models and has grown revenue twentyfold in a year as token prices spiked.
Does it annoy you to pay an AI to explain what probably another AI wrote? One AI is fluffing up the word count to an unreadable level while another has to trim it back down. If every article was just the text of the prompt that generated it, it would be a much quicker read.