This strategy didn’t work well for Scott Alexander, and will fail for you in a similar situation—if someone decides to publish an article containing your name (so it will appear in search results for your name), linking the semi-anonymous article (where people coming from their article will be able to verify that it is you).
The name that would appear on such posts is “anonymous” but the name would be a link to the author’s user profile.
If your username is your real name, then the link to your profile contains your real name, and therefore the HTML code of the semi-anonymous article also contains your real name. Not sure if this is enough for search engines to show the article when people search for your name.
This isn’t meant to be perfectly secure. It’s meant to be a bit more secure than currently.
It’s also better than the Scott Alexander situation, since your articles can only be doxed one at a time, rather than all at once.
Finally there’s ways of doing the link such that it reveals nothing. I’m fact it will need to be the case if you only allow users with some minimum of Karma to follow the link.
Sorry, missed that part. If the link to author profile is not included in the article, but is only downloaded after the user (with sufficient karma) clicks on the button, my objections do not apply.
It’s also better than the Scott Alexander situation, since your articles can only be doxed one at a time, rather than all at once.
It would be simple to write a bot that would scrape names from all the semi-anonymous articles. Someone could even set up a LessWrong mirror that automatically de-anonymised everything.
As suggested, we could force them to have a lesswrong account with e.g. 100 karma.
It should be straightforward to detect such a bot (the same user account clicking on every single semi anonymous article) and blocking it, and gaining 100 kudos is annoying enough to make it not worth doing repeatedly.
This strategy didn’t work well for Scott Alexander, and will fail for you in a similar situation—if someone decides to publish an article containing your name (so it will appear in search results for your name), linking the semi-anonymous article (where people coming from their article will be able to verify that it is you).
If your username is your real name, then the link to your profile contains your real name, and therefore the HTML code of the semi-anonymous article also contains your real name. Not sure if this is enough for search engines to show the article when people search for your name.
This isn’t meant to be perfectly secure. It’s meant to be a bit more secure than currently.
It’s also better than the Scott Alexander situation, since your articles can only be doxed one at a time, rather than all at once.
Finally there’s ways of doing the link such that it reveals nothing. I’m fact it will need to be the case if you only allow users with some minimum of Karma to follow the link.
Sorry, missed that part. If the link to author profile is not included in the article, but is only downloaded after the user (with sufficient karma) clicks on the button, my objections do not apply.
It would be simple to write a bot that would scrape names from all the semi-anonymous articles. Someone could even set up a LessWrong mirror that automatically de-anonymised everything.
As suggested, we could force them to have a lesswrong account with e.g. 100 karma.
It should be straightforward to detect such a bot (the same user account clicking on every single semi anonymous article) and blocking it, and gaining 100 kudos is annoying enough to make it not worth doing repeatedly.