Anybody have experience with https://github.com/chopratejas/headroom? They seem to have similar goals in token reduction, but headroom appears to be broader in scope.
That's fair. There are even many dimensions to define 'quality' which include use case (coding? writing? multimedia?) and prompt. I suppose if you ask testers to provide benchmarks with their analysis, that might hamper their desire to share.
Wonderful idea to document, share and even have positive outcomes.
The author would probably love this YT channel which is all about helping others come to the same realization as he did: https://www.youtube.com/@socialanimal
TinyProgrammer is a Raspberry Pi on my desk that autonomously writes little Python programs forever. It types code at human speed, makes mistakes, fixes them, and has moods.
Both good and bad responses to successful terrorist attacks threaten government credibility, simply because they were successful. The key is for a government to prevent them from being successful in the first place.
Only bad disaster responses make administrations look pretty bad.
>There's a reason people end up in these ruts to begin with.
Perhaps there is the lack of self-awareness that dopamine comes into play. This is where the (effective) simplicity of this article comes into the picture.
I'd think that despite variations in environments in a home, there are either 1) enough accommodating tools to normalize (e.g., lights that have a variety of settings) or 2) the majority of the processes required to handle the variations are figured out (e.g., in consistently poor lighting, you need package B instead of package A).
I see some hardware packages for home studios going on sale, but nothing to the level of hardware/software integration necessary for the most effective online instruction/discussion.
[ my public key: https://keybase.io/alexkoo; my proof: https://keybase.io/alexkoo/sigs/MvYxxCbnvEPq3N5Nd0SsuepBrCswVZ7yJjirq7WyR5E ]