>You have six hours to get your point across before the thread dies
It used to take longer for a thread to die, but comments have always solidified rather quickly.
This was fine, anyways, because each thread was an archive of the state of mind of the people involved at the time. If the conversation continues, it should continue in an entirely different thread, and if you're new to the conversation, you can backtrack over the previous threads to get up to speed with everybody else. There was a real hivemind-like (in a non-derogatory sense) stream of consciousness that everybody could participate in.
This is impossible now, though, because of moderators preventing multiple threads subjectively "related" to one another being on the first 5-8 pages at once.
Worst of all is the "megathreads", which aggregate literally over a hundred links to external sites into a single post, where people can't even vote on them individually. The entire benefit of the format of Reddit is user-filtered content, so you don't have to read literally the entire internet to find your way to something useful.
With all the hype around AI, I feel like there is so much low-hanging fruit in the realm of "human algorithms" that everyone has failed to explore over the last 15ish years of stagnation (and decay) in social media. People need useful processes and structures, not magic question-answer machines, or recommendations based on "who they really are inside."
Just imagine, for instance, if HN allowed you to vote specifically on backlinks to previous threads for each individual post. As of now, there's no easy way to know how far a discussion on a particular topic has advanced in the past, and judge whether you're about to say something redundant. If users could easily post a related thread and get some signal back about whether it was helpful, it would occur to them more often to re-post relevant discussions. Thread-backlinks competing with fresh comments and other links is just too much noise.
If you look at the discussion around the "Clean Code, Horrible Performance" article, there is a followup which addresses some of the claims the commenters made about the original. Then, in the comments for that article (across all platforms), you find people claiming that nobody would ever actually say those things.
Eventually people are going to realize the UX provided by AI isn't actually that great, and rediscover that DAGs are really useful.
It used to take longer for a thread to die, but comments have always solidified rather quickly.
This was fine, anyways, because each thread was an archive of the state of mind of the people involved at the time. If the conversation continues, it should continue in an entirely different thread, and if you're new to the conversation, you can backtrack over the previous threads to get up to speed with everybody else. There was a real hivemind-like (in a non-derogatory sense) stream of consciousness that everybody could participate in.
This is impossible now, though, because of moderators preventing multiple threads subjectively "related" to one another being on the first 5-8 pages at once.
Worst of all is the "megathreads", which aggregate literally over a hundred links to external sites into a single post, where people can't even vote on them individually. The entire benefit of the format of Reddit is user-filtered content, so you don't have to read literally the entire internet to find your way to something useful.
With all the hype around AI, I feel like there is so much low-hanging fruit in the realm of "human algorithms" that everyone has failed to explore over the last 15ish years of stagnation (and decay) in social media. People need useful processes and structures, not magic question-answer machines, or recommendations based on "who they really are inside."
Just imagine, for instance, if HN allowed you to vote specifically on backlinks to previous threads for each individual post. As of now, there's no easy way to know how far a discussion on a particular topic has advanced in the past, and judge whether you're about to say something redundant. If users could easily post a related thread and get some signal back about whether it was helpful, it would occur to them more often to re-post relevant discussions. Thread-backlinks competing with fresh comments and other links is just too much noise.
If you look at the discussion around the "Clean Code, Horrible Performance" article, there is a followup which addresses some of the claims the commenters made about the original. Then, in the comments for that article (across all platforms), you find people claiming that nobody would ever actually say those things.
Eventually people are going to realize the UX provided by AI isn't actually that great, and rediscover that DAGs are really useful.
reply