[db0] Reddit is a dead site running

https://lemmy.ml/post/2040930

[db0] Reddit is a dead site running - Lemmy

The link contains @[email protected] 's views on the ongoing state of Reddit, and I think that it’s worth sharing here - both to document a piece of opinion, and as food for thought. The main points are: * a comparison between the current state of Reddit vs. Myspace near collapse; * the illusion that everything is fine based on “raw” numbers like engagement; * that Reddit was never a “good” site, but it had two positive points (open API and hands-off approach to communities), destroyed by the current events; * the ongoing progression of the Fediverse as alternative to Reddit; * the change in quality in both the content and the behaviour of the people still there.

I agree and I really think Reddit’s enshittification has led to the fediverse reaching critical mass (via migration to Lemmy) for self sustainability. And because of how the fediverse/Lemmy works, I think Lemmy could end up being a sort of foundation that other fediverse apps (Mastodon, Pixelfed, Peertube, etc) can stand on to reach their own critical mass.

Reddit is just the first domino to fall.

Lemmy works so well because Reddit was always about interacting with normal, average users. Mastodon and Peertube don’t work as well as Lemmy because many people go on Twitter and YouTube for well known people with big followings, and those people will stay on whatever platform is most popular. This is just my opinion, I hope it changes someday.
Is it just me or has interacting with humans on reddit become a frustrating experience? People are only trying to get one over on each other, all debates I see seem to be about who can just get the harshest insult in before blocking the other person… i guess reddit’s popularity works heavily against what made reddit good in the first place, i think the more popular something is, the lower the common denominator
I said it before, but I think a lot of that toxic arguing was us being played by bad actors. People would say the meanest shit or have the worst takes imaginable to get a rise.
Those “bad actors” are users too though. They’re real people who enjoy causing outrage. We’ve had them since the invention of BBS. We used to call them trolls. I think we still call them trolls, but we used to too. Seriously though, so often now people jump to the conclusion that the guy with the shitty take and horrible attitude must be from some Russian cyber farm. Maybe he’s just an asshole.
I hear you and I agree they’ve been around since I first started using mIRC 27 years ago. But the trend now is that there are more of them than there are of us. It takes ten minutes to load up a large language model to behave like a right or left wing troll and let it loose on Reddit, or here for that matter. In fact, I’m going to code one up tonight for shits and giggles that spreads positivity.
How do you train the LLM? Just follow the ChatGPT docs and point it to the API as a source?
ChatGPT can work, but it’s expensive. You can use oobabooga and grab a model off huggy face and then code up a personality. Then you have the API send generation requests to the model and post as if it’s a human.
GitHub - oobabooga/text-generation-webui: A Gradio web UI for Large Language Models with support for multiple inference backends.

A Gradio web UI for Large Language Models with support for multiple inference backends. - oobabooga/text-generation-webui

GitHub
Thanks! I’m going to give that a shot. I’ve been messing around with the ChatGPT API as a content generator for another project.
I made a bot that interacted with my Twitch chat using this! Very fun use of the model. I’ll warn you it always uses 100% when processing, usually for a couple seconds. It can cause stuttering while gaming.