[meta] v0.2 released, briefing on hold due to rate limiting

github.com/JonGerhardson/masshoLLM

I’ve updated the code for the script I’m using to create these briefings on github. See https://github.com/JonGerhardson/masshoLLM for details.

Some improvements include better excluding pages that are not new from being sent to LLM provider for processing, parsing of meeting announcements, and some extremely rudimentary fact checking that is not in anyway a substitute for actual fact checking except that if the fact check step fails in a certain way, it’s probably more likely to be wrong than not. . . or it’s wrong because the API call to LLM failed and the factually inaccurate text is “Processing Error” or something like that.

The new exclude rules helped me narrow down over 700 urls marked as updated yesterday to only about 200 likely to contain “new” information. But even then, I’m still hitting rate limits on Gemini’s free tier, and my GPU bit the dust recently so unless someone’s got a bunch of tokens or old hardware they want to give me, the actual briefing is on hold for this week.

xoxo

jon.gerhardson@proton.me