Honestly though, I hope that the google notbooklm https://notebooklm.google/ doesn't go to the google graveyard. It is great for feeding a decent amount of information and helping you process it. I've found great success at it.
Totally agree. Great product. I wish I could feed in internal company meeting recordings though.
It's part of Google Workspace now so it's much less likely to go away.
Anyone have recommendations for chip datasheets? Ive explored a couple options so far, but getting some bitfields wrong is super annoying.
I see plenty of examples like the one here that are on easier extractions. A PDF to HTML or Markdown converter will probably get it right with OCR.
No direct recommendation for that use case, but one strategy I've heard being used and that works with complex documents (or where hallucinations are Very Bad™ - like invoice processing) is using multiple techniques and models at once in a quorum approach. For example, direct ingestion of PDFs into Gemini, OCR and ingestion of text, plus perhaps using another model like GPT. If they all agree on a fact, you're (probably) good. If not, it can be bumped up to human correction.
Interesting, although would be great to see some comparative results, e.g., with and without the html alt tag approach.
I haven't ever considered the self grading approach. Going to try that out!