someone-build

A browser extension or plugin for reading apps that acts like a "syntax highlighter" for written language. Feed the page you're reading to an LLM and let it rank each token based on some user defined filters and then highlighting phrases according to the filter rankings. 

The obvious example is to highlight phrases based on political leanings. but, I think the more interesting filters would be things like, is this phrase a technical explanation, a call for violence, is it supported by evidence provided in the text, is it logically consistent, is it attempting to shill something to the reader or trying to extract value, how much fluff does it contain

I think a tool like this would be extremely valuable for parsing the increasingly hazey information landscape

Interested in ZK, FHE, PUFs, hardware, system design, ai safety, and smart contact security.

Also would be a great addition in a text editor to evaluate your own writing