Dang is going to have 65,535 karma points soon

Downweight posts and comments based on the frequency and positive sentiment. IE things that are posted often and with high positive comments should bubble to the top less often.

Other people want us to downweight negative sentiment. I wonder what happens if you downweight all the sentiments.

Well my intuition is topics with high negative sentiment won't saturate the front page for long periods of time, unlike topics with high positive sentiment. The latest thing is obviously AI and chatGPT. I'm interested in both, but if there was an increasing downweight based on how long it is popular, and how popular it is, they would still show up on the front page, but not quite so frequently.

The solution is simple: downweight all content which generates engagement of any kind

Do you have a simple AB testing system in place to test hypotheses like this? If you're the only sherif in town running experiments, it doesn't take much work to build a simple one, probably around ~100 LOC ("do things that don't scale").

We don't. I know it sounds simple but I'm too tired.

Edit: to expand a bit lest that seem snarky - what I mean is that maintaining the current system takes so much energy that there's precious little capacity left over for creative exploration. This is a problem.

The obvious solution is to ban commenting altogether. On balance, it's just a bad idea. And the commenters. Ban everything, delete the site, degauss the drives and throw the server out of the window. Then throw a bigger, bulkier server out of the same window so it lands on the first server. Then break both servers down until their components are indistinguishable from one another, and mix them together. Then separate them into a number of small piles and hide each pile inside of a complex puzzle box - a different puzzle for each - in different sacred locations around the world, protected by whatever spells or enchantments are culturally appropriate. Then kill anyone who knows the location of the puzzle boxes, dissolve their bodies in lye, mix the lye into concrete blocks and dispose of the blocks in international waters, following a random dispersal path. Then blow up the boat with an orbital laser. Methods for the procurement and operation of an orbital laser are left as an exercise for the reader.

It's the only way to be sure.

I agree with you that popularity is a diminishing curve and at some point becomes annoying. With the current AI tsunami, we're well past that point for some people, but not others.

From an admin point of view it's tricky because (unless I'm high on koolaid?) this is a major technical development, so genuine advances are happening all at once. Too many for me to keep up with.

I understand. In the typical corporate world this is solved with "hire interns to work on all the cool things we don't have bandwidth to explore, while being simultaneously jealous of them". Or I'm sure a lot of people such as myself would happily volunteer to help out.

Outside of the UI difficulties in keeping a clean look, would different ways of providing response to a post or topic not help to solve this? For instance if there was some way to mark something as repetitious that could be contrasted against e.g. upvotes to get dynamic and 'automatic' feedback on the perceived 'freshness' of a topic.

In general downvotes seem like a relatively poor feedback mechanism because there's no shared agreement on how they should be used. This [1], perhaps ironically flagged, post offered feedback on why people downvote, and it's just all over the place. Even if there are guidelines, people will be people. At least with something like clear adjectives, the percent of 'intended' feedback would be higher.

[1] - https://news.ycombinator.com/item?id=23997697

Isn't it simpler to either nuke it all from orbit, or bring in the sharks (with lasers) ?

Have you looked into expanding the team? (I’m guessing the answer is yes: if so, how has it gone?)

Yes but it's complicated. Mostly by me.

I feel that the curve needs to be adjusted to curb clickbait from trending to the top.

There are a lot of stores that trend to number 1 in under an hour that end up disappearing completely from the front page, moving near the bottom, or flagged to death, within a couple of hours because it’s just not quality content for this forum.

I don't know how to do that! Wouldn't it require software that could distinguish high-quality from low-quality content?

The same bait (sensationalism, indignation) that makes many users upvote those posts is what makes other users flag them. This rise and fall has always been the pattern on HN—it's one of the cycles of life here.

It's like throttle/pedal mapping in a car. It's about the effective curve along time/position.

Right now, it's super sensitive where a little signal does a lot right at the beginning.

Stories rocket to the top because they're new and got a number of quick votes. The curve needs to be smoothed out so that things don't lurch to the top.

Edit: I've been around here for a little while. It used to be awesome to see things jump to the top (used to have a lot lower volume of posts and users), but the audience has changed a lot, and I think things need to be adjusted a bit to slow the meteoric rise aspect. Good content will surface.

Just to provide an example of the problem. This post reached #1 with 16 upvotes and 0 comments in less than an hour.

https://news.ycombinator.com/item?id=35478120