The bots are doing nothing 'wrong' as such. There are certain feeds from where >90% of the material ends up on HN. If somebody writes a bot to auto-submit items from that feed, it's helping HN be faster in accessing new and (mostly) relevant posts. What is not relevant should (in principle) not reach the home page.
The problem is that submission is winner takes all, i.e. first submitter reaps all the upvotes. Also, HN does not reward finding new sources of material more than posting from the predictable sources.
The logical endpoint of these two facts is auto-submission bots rising to the top of the Karma tables. If Karma is used as a way of anti-gaming (need karma threshold to downvote, etc.) then this is a way to circumvent that for a sufficiently motivated manipulator. They can make a bot that auto-submits from the known sources, and use it to build up an arbitrary number of accounts, from which they can then boost the articles they want (or bury the comments they don't like). For all we know, this could already be the case.
I have two suggestions that can sidestep this problem to a certain extent:
1. Reward articles from predictable sources less than articles from rare sources.
2. Split the karma benefit between the first submitter and the people who upvoted the article early (of course, this needs to take into account how selectively and successfully the user upvotes to avoid blanket upvoting 'just in case').
This all is still gameable, but probably not as easily.
The problem is that submission is winner takes all, i.e. first submitter reaps all the upvotes. Also, HN does not reward finding new sources of material more than posting from the predictable sources.
The logical endpoint of these two facts is auto-submission bots rising to the top of the Karma tables. If Karma is used as a way of anti-gaming (need karma threshold to downvote, etc.) then this is a way to circumvent that for a sufficiently motivated manipulator. They can make a bot that auto-submits from the known sources, and use it to build up an arbitrary number of accounts, from which they can then boost the articles they want (or bury the comments they don't like). For all we know, this could already be the case.
I have two suggestions that can sidestep this problem to a certain extent:
1. Reward articles from predictable sources less than articles from rare sources.
2. Split the karma benefit between the first submitter and the people who upvoted the article early (of course, this needs to take into account how selectively and successfully the user upvotes to avoid blanket upvoting 'just in case').
This all is still gameable, but probably not as easily.