Skepticism is a necessary trait in this type of work, for sure. I will say that the performance has improved substantially in the past year, and there are still PDFs that require a lot of work.
We went with official precinct results for two main reasons: there are differences between election night and final results (some of them non-trivial) and to make the work more manageable. Agree that historical results are a real problem, and as a PA native I know only too well the errors that the state data contains, which is why we go county-by-county there.
Totally reasonable view, and one of our volunteers actually got the law in Kansas changed to mandate electronic publishing of statewide precinct results in a structured format! But finding legislative champions for this issue isn't easy.
Many jurisdictions do risk-limiting audits using the original ballots, so futzing with the results wouldn't necessarily make that easier. Also, cast vote records are public in many states - those are records of each ballot cast. So people can check.
Yeah, this is a very well-traveled road, but LLMs have made some big improvements. If you asked me (the guy who wrote the original piece linked above) what I'd use if accuracy alone was the goal, probably would be AWS Textract. But accuracy and structure? Gemini.