I wonder how this compares to my M4 air with 10 GPU cores and 32 MB of RAM. My system can only run ~14B sized models at any reasonable speed. The accuracy of these sized models can be underwhelming. I am looking forward to a time when it would be nice to run models locally at a reasonable price, at a reasonable speed and with reasonable accuracy. I don't think we are there just yet.

I am researching go-string-concat-benchmark [1]:

    A performance comparison of four common Go string building methods.

I recently updated my go-stats-calculator to include many more stats [2]:

    CLI tool for computing statistics (mean, median, variance, std-dev, skewness, etc.) from files or standard input.

I also created claude-image-renamer [3]:

    AI-powered image renaming script that generates descriptive filenames for screenshots. 

[1] https://github.com/jftuga/go-string-concat-benchmark

[2] https://github.com/jftuga/go-stats-calculator

[3] https://github.com/jftuga/claude-image-renamer

Cool project. Do you have a repo for it?