Not a newt

O Fortuna (Carmina Burana).

Electric Callboy. I’m generally not fond of metalcore, but their songs are very catchy.

The politicians debating online abuse claim to mean well.

That’s not killing half the voting system to solve half of the vote manipulation. Downvotes do not even get used at the same ratio as upvotes. I’m sure someone can pull numbers, but I’d roughly estimate that in most communities no more than 10% of votes are downvotes. And even if they were, I’m not sure you quite parsed my full comment.

• I stated very early that I don’t specifically like disabling downvotes.
• I stated why I think that post-hoc remediations will not work.
• I proposed a potential compromise which can be used to mitigate abuse without a blanket downvote ban.

Blocking voting on fresh accounts is not a novel idea. As another commenter said, it’s the system used on Stack Overflow. Blocking all downvotes is not even the goal. The goal is to make brigading not worth the effort. The worst case scenario is that all downvotes get disabled (which still works, despite its unpopularity - it’s been implemented by instances like beehaw). But in the end, that’s just a baseline. It can be improved, and I like to believe that I was quite clear on that in my first comment.

I know? I didn’t say it didn’t happen, I said that positive vote manipulation can more easily be addressed with spam prevention measures.

Corollary: Don’t assume you’ll have access to your accounts forever.

Yep. One can definitely have a heated debate about the merits of one process over another, when the outcome is constructive. That’s a rational argument.

Trolls on the other hand get “you’re not arguing in good faith, this conversation is over” and a block.

As much as it pains me, I think the only solution to vote manipulation is to disable downvotes. Mind you, I don’t like it - I think downvotes are useful in a healthy self-governing community - but here’s my rationale as to why it’s the only solution:

• The goal of negative vote manipulation is to remove visibility from content. For that, the first few hours of the post’s or comment’s lifetime are critical. Sure, a mod can remove the downvotes, but it would likely be done after the content’s attention window is over, so the damage would be done. [1]
• Positive brigading (artificial boosting of content) is another problem, but out of scope of this post. I consider it to be in the “dealing with spam” category.

1. As I’m writing this, it comes to mind that perhaps we can selectively disable downvotes? Just like some instances don’t allow fresh accounts to post, perhaps something similar can be done for downvoting. Maybe it can also be extended to accounts below a certain up- to downvote ratio, to avoid mass downvoters.

By cults I refer to the CEOs and influencers proclaiming things like “if you’re not using AI for everything right now then you’re making yourself obsolete.” Then they go and show a few personal anecdotes to support their claim, while at the same time dismissing any and all criticism of LLMs and their products.

Before the AI craze, LLMs were a neat niche academic tool. Then the cults started forming…

At some point I’d have thought so too, but after Rebel Moon I’m convinced that just large scale space battles is not enough.