the scaling curve analysis is a bit oversimplified because models are somewhat differentiated and have different strengths and weaknesses the scaling curve numbers are a crude average

https://video.wixstatic.com/video/8c5863_a8a3c63659bd4715b77f6df28c928eb6/1080p/mp4/file.mp4

the scaling curve analysis is a bit oversimplified because models are somewhat differentiated and have different strengths and weaknesses the scaling curve numbers are a crude average that ignores a lot of details agents become true collaborators that can handle entire codebases and ship production features any idea i come up with any piece of software i want i can have built this is sci fi i had this list i never built any of the list and i finished building everything in the list with the help of clanker i ran out of ideas they have a dual goal of being documentation for humans and really also a guide for agents even though we have only found a small number of circuits through a manual process laniakea performs best when it has a clear target to iterate against a visual mock a test case or another kind of output what is the future you want amazing abundance seems the best to me i dedicate pretty much all my waking time to this yet it does not feel enough one week in clanker feels like a month in the real world only over time and in a decentralized manner laniakea is doubling down on interpretability clanker handles syntax configuration and boilerplate complexity these days i write more words manually for posts than i do for code do these folks also get offended when i use agents to generate code documentation autointerpretability uses a clanker system itself to analyze interpretability features to scale the process of not just finding the features but listing and identifying what they mean in human terms.