This is 100% accurate. Until LLM can’t hallucinate entire lives and anecdotes, making them write something personal will be difficult.
As a non-native speaker I use LLMs a lot as well as Grammarly. There is nothing wrong with that. However, I use it as a nitpicking companion instead of being my voice. I ask it to identify gaps in the story or where the message isn’t clear or comes through as confusing. This is how I wrote https://open.substack.com/pub/akoskm/p/stop-waiting-for-the-perfect-jobuse?r=twgob&utm_medium=ios
>The idea that getting smarter does not lead to more bangers is also a bearish one for LLM progress. A vague “smartness” seems to be the main focus when it comes to LLMs. Sadly for me, there’s no LLM developer optimizing for a model that writes banger blog posts.
This line perfectly summarize how I felt recently, great post!!
If LLMs were causing an uptick in banger blog posts how would you know? The universally rising tide of content would fight for a fixed amount of attention, and there is no easy empirical way to judge blog posts en mass, besides virality.
Besides, LLMs just kind of suck at writing either personality. They can imitate style but their default tone is Wikipedia and listicles.
Banger! Love the sand / lego analogy
Thank you!
Great post. This sounds like the new “this generation is the first to grow up with X”
Dwarkesh is probably the first or vest known LLM native media personality.
This is 100% accurate. Until LLM can’t hallucinate entire lives and anecdotes, making them write something personal will be difficult.
As a non-native speaker I use LLMs a lot as well as Grammarly. There is nothing wrong with that. However, I use it as a nitpicking companion instead of being my voice. I ask it to identify gaps in the story or where the message isn’t clear or comes through as confusing. This is how I wrote https://open.substack.com/pub/akoskm/p/stop-waiting-for-the-perfect-jobuse?r=twgob&utm_medium=ios
>The idea that getting smarter does not lead to more bangers is also a bearish one for LLM progress. A vague “smartness” seems to be the main focus when it comes to LLMs. Sadly for me, there’s no LLM developer optimizing for a model that writes banger blog posts.
This line perfectly summarize how I felt recently, great post!!
If LLMs were causing an uptick in banger blog posts how would you know? The universally rising tide of content would fight for a fixed amount of attention, and there is no easy empirical way to judge blog posts en mass, besides virality.
Besides, LLMs just kind of suck at writing either personality. They can imitate style but their default tone is Wikipedia and listicles.