Idea kernels & writing for the AIs
How LLMs change internet writing
I’ve recently been seeing more of Tyler Cowen's Marginal Revolution blog posts quote OpenAI's o1. I think this reflects the fact that he is earlier than others in understanding and adapting to how LLMs change the nature of blogs & ideas.
On Marginal Revolution Tyler will mention o1 as the side-kick that can provide context and details where needed: He’ll link to an idea and mention: “Here is o1s response <link>; here is its analysis <link>; here is its view on whether this is exaggerated <link>”. If you don't get it, just ask o1!
Tyler’s blog posts then only need to provide the core idea - AI then can do the exposition.
Blogs and content become “idea kernels” - enough context to seed the idea and then the reader can expand with LLMs.
Blog can be like instant cake mixes: “Just add prompt!”
On the other end of the spectrum, Tyler's long-form writing also changes in the world of LLMs. He now writes long-form for the AIs.
As he mentions on Dwarkesh’s podcast when describing his book “GOAT”:
“I’m happy if humans read it but I mostly wrote it for the AIs... my next book is even more for the AI again. Human readers are welcome... but it doesn’t matter anymore [if human review it.]... the AIs will trawl it and will change how they see me.”
“I think of myself over time less as a producer of content and more as people person, connector, developing networks”... internet writing his immortality
“As far as I can tell no one else is doing this - writing and recording for the AIs... you should be writing for the AIs...”
(Link below to that part of the interview)
With all of that in mind, this blog in 2025 will fall on these two extremes. Shorter-form “idea kernels” and longer-posts that can be consumed with o1.
Until next time
