> Patent similarity search doesn't work that well at present.
Neither does automatic detection of fair use - that's the point. Just crank up the false-positive rate so anything vaguely related counts as prior art. If the submitter is not happy about it, they can argue their case in a proper court.
The status quo is close to what you describe. Very few patent applications are not rejected in the first action. I did a quick Google search that indicates that over the entire USPTO, 88% of patent applications are rejected by an examiner in the first action [0]. I myself never allowed an entire application on the first action!
However, it's rare that an attorney will argue that their patent should be issued in court. Almost always they argue with an examiner.
A lot of people here seem to believe that the USPTO issues patents like handing out candy to children, but the reality is that a large fraction of patent applications are never granted. Mistakes happen, yes, but that has more to do with how little time the USPTO gives examiners than anything else. It's really easy to miss something. If you want to improve patent quality, tell your congressmen to increase funding at the USPTO so examiners can get more time. Nothing else will work in my view. Right now the USPTO is funded solely by fees paid by patent applicants, which isn't enough and probably distorts the incentives.
ican confirm it can be more like pulling teeth than freely receiving candy. I’ve pulled a few teeth and by the time I got it, I wasn’t sure it was worth it.
The uspto seems to me like the fda in that they typically have a fraction of the resources and a sub-fraction of the compensation of those on the other side of the table, yet seem to get blamed for the downfall of society based on their decisions.
Neither does automatic detection of fair use - that's the point. Just crank up the false-positive rate so anything vaguely related counts as prior art. If the submitter is not happy about it, they can argue their case in a proper court.