Great write up, but substack is messing up the formulas displayed after clicks. It really messed up the equations as it appears that anything after "\" is escaped from latex rendering.
Beauty of current algorithm is its simple and easy to compute but most importantly anyone can submit the news and get a fair chance at it. Whether you are new user or old user, news is news and news trend based on votes. But your suggested algorithm (that requires 100 cpu and 128 gb ram btw but I digress) is very allowing the power to concentrate in a very few hands irrespective of the news value they submit. it very much incentivizing to create bots and voting rings and spam the site to harvest as much as votes as possible like stackoverflow.
Great write up, but substack is messing up the formulas displayed after clicks. It really messed up the equations as it appears that anything after "\" is escaped from latex rendering.
I noticed that Substack's LaTeX feature is still in beta.
Beauty of current algorithm is its simple and easy to compute but most importantly anyone can submit the news and get a fair chance at it. Whether you are new user or old user, news is news and news trend based on votes. But your suggested algorithm (that requires 100 cpu and 128 gb ram btw but I digress) is very allowing the power to concentrate in a very few hands irrespective of the news value they submit. it very much incentivizing to create bots and voting rings and spam the site to harvest as much as votes as possible like stackoverflow.
> But your suggested algorithm (that requires 100 cpu and 128 gb ram btw but I digress)
I don't think so, and HN does not have a billion users. I think it may have 15 million. You can compute PR with some okay hardware.
https://web.archive.org/web/20201109023658/https://michaelnielsen.org/blog/using-your-laptop-to-compute-pagerank-for-millions-of-webpages/