If anything, there's so much new stuff these days that I don't even get the FOMO to try it all.. building it myself is easy too.
A year ago I never felt my MacBook was short on resources, but these days I find myself thinking a lot about the performance of harnesses and agent tools.
I used to work on how my writing would come across and on writing what people would like. Now I spend a long time on how to cut the LLM voice out of what AI wrote for me (I keep repeating "this sounds too AI, fix it"…
Claude Code and Codex still fall short in a lot of ways compared to how smart the models are.. even though models above Opus 5.5 or Astra are probably involved.
Some days I get far more done than my capacity and end up in an addicted state where I just can't stop working, and some days I don't even want to look at an AI chat window.. I also don't really get the feeling that I'm…
If anything, there's so much new stuff these days that I don't even get the FOMO to try it all.. building it myself is easy too.
A year ago I never felt my MacBook was short on resources, but these days I find myself thinking a lot about the performance of harnesses and agent tools.
I used to work on how my writing would come across and on writing what people would like. Now I spend a long time on how to cut the LLM voice out of what AI wrote for me (I keep repeating "this sounds too AI, fix it"…
Claude Code and Codex still fall short in a lot of ways compared to how smart the models are.. even though models above Opus 5.5 or Astra are probably involved.
Some days I get far more done than my capacity and end up in an addicted state where I just can't stop working, and some days I don't even want to look at an AI chat window.. I also don't really get the feeling that I'm…