• 0 Posts
  • 9 Comments
Joined 2 years ago
cake
Cake day: July 17th, 2024

help-circle


  • Not sure about keyword generation exactly as OP intends to apply this, but short simple code that needs to be validated on output seems like the better application. Models are tiny and fast, so mmap and run in parallel and race to what passes validation first. Each agent gets a git branch. Losers generate the anti patterns. Human or cheap tokens integrates the results from corpus back into harness. Next generation.


  • It isn’t even the LLM. Its the data scraped from the internet wholesale and users contributing to corps by use. Not to mention the staggering amount of engineers turning that raw data into better output. The LLM just spits those billions of dollars of effort back out. I only pick this nit because too many people see local LLM and think that’s where the magic happens. It is just a flash in the pan. you need an internal combustion engine, transmission, wheels, chassis, cabin, and probably some seats if you want to get anywhere with it. That said, it is the concentration of effort. Individuals can’t really rival that. I wonder if there are any initiatives in the open source community with this mindset…