OptimusPrime

joined 1 year ago
MODERATOR OF
[–] OptimusPrime@lemmynsfw.com 3 points 1 year ago

No, I assume that the bots would only repost the most highly voted content. If a bot were to post everything, including low-quality content, the moderators could simply ban it.

[–] OptimusPrime@lemmynsfw.com 4 points 1 year ago

As for having the bots in another instance, people are going to go to the instance with the most content, and that would be the bot instance. Do we really want to drive the users away?

[–] OptimusPrime@lemmynsfw.com 4 points 1 year ago (2 children)

As for different communities, people who don't want to see bots can disable them in the user settings, so the communities wouldn't change for them. I don't see the point in having bot communities.

[–] OptimusPrime@lemmynsfw.com 0 points 1 year ago

I don't agree that a bot would steal content. Content creators need exposure and they post for free on Reddit, so why wouldn't they want the same here?

[–] OptimusPrime@lemmynsfw.com 10 points 1 year ago (1 children)

There is a list of federated and blocked instances on the bottom of the page. https://latte.isnot.coffee/instances

Also this script is to see which instances block or are blocked by the instance you choose.

[–] OptimusPrime@lemmynsfw.com 10 points 1 year ago* (last edited 1 year ago) (1 children)

I'm honestly more confused than I was before. With so many opinions, I don't know how this could ever be implemented in a way that satisfies people.

9
... (www.youtube.com)
submitted 1 year ago* (last edited 1 year ago) by OptimusPrime@lemmynsfw.com to c/localllama@sh.itjust.works
 

Microsoft Research has developed Orca, an open-source project that introduces a progressive learning model based on GPT4. This model achieves impressive performance comparable to GPT 3.5 and GPT4 while using minimal storage and operating offline. By incorporating complex explanation traces, Orca enhances interpretability and addresses challenges in complex machine learning models. It outperforms other models in terms of accuracy, performance, and interpretability. The training process involves tokenization, sequencing, and loss computation. Experiments demonstrate Orca's proficiency in various tasks and domains, showcasing its capabilities in writing, comprehension, and reasoning. The research paper offers detailed insights and comparisons with GPT 3.5 and GPT-4, highlighting the potential of Orca to empower smaller models in competing with larger counterparts.

[–] OptimusPrime@lemmynsfw.com 6 points 1 year ago (1 children)

This reminds me of the Black Mirror episode where everyone's social status was determined by their ratings.

[–] OptimusPrime@lemmynsfw.com 2 points 1 year ago* (last edited 1 year ago) (1 children)

Yeah you are right! I'm sure you can do better.

 

Over the past few days, I've witnessed a remarkable surge in the number of communities on browse.feddit.de. What started with 2k communities quickly grew to 4k, and now it has reached an astonishing 8k. While this exponential growth signifies a thriving platform, it also brings forth challenges such as increased fragmentation and the emergence of echo chambers. To tackle these issues, I propose the implementation of a Cross-Instance Automatic Multireddit feature within Lemmy. This feature aims to consolidate posts from communities with similar topics across all federated instances into a centralized location. By doing so, we can mitigate community fragmentation, counter the formation of echo chambers, and ultimately foster stronger community engagement. I welcome any insights or recommendations regarding the optimal implementation of this feature to ensure its effectiveness and success.

view more: next ›