Is #mastodon becoming an echo chamber?
-
OK, this is going even MORE sideways so I need to make a few things clear:
1. I took a complex point and made it poorly
2. My goal was to ask for more inclusiveness
3. I am sickened by what happend to BlackTwitter and I don't want it recur
4. But I can't speak for BlackTwitter nor should I
5. I apologize to black mastodon users for making such a poor comparison
6. I'm not endorsing "AI Slop" they were a foil to make my point
7. I'm certainly NOT trying to compare AI bros to Black twitter (but, as I said, I can see how people made that connection. I'm trying to correct that here)@scottjenson This is one of the most unintentionally (I think) hilarious lists I’ve ever read. I hope you’re well and taking a break from opining in public.
-
@Gargron That is a personal choice and one which I totally respect. But I do think Mastodon should be big enough, and open enough, to allow an "AI community" to form, even thrive.
Too many people in my replies don't seem to agree with that.
@scottjenson @Gargron I personally won't be following them, and most will find themselves muted or blocked by me. But they are welcome to try to form a community here. The problem is, the Fediverse is made up of people trying to escape corporate control in at least one form, namely social media. We are often looking for other ways to separate ourselves from big tech as well.
This worldview isn't exactly compatible with AI/LLM use.
AI users simply aren't likely to fit in here. In my opinion, they are welcome to try. But I strongly suspect they will find themselves in their own echo chamber, with only other fans of AI listening. The end result will be nothing of value, much like AI itself.
-
As this conversation is spiraling a bit I want to make a few things clear:
1. I'd like Mastodon to be MORE inclusive and bring in more voices
2. Some people don't seem to want that
3. This is core problem to solve: How do we let more in, but not "pollute" your feed?
4. The solution is NOT "gatekeeping", revelling in the fact that AI journalists aren't welcome
5. This is the same reason we lost "Black Twitter" when it came over in 2022Yes, a lot of you don't want AI posts in your feed (or pick any other topic) but the solution isn't to keep "AI People" from joining Mastodon, any more than it is keeping marginalized communities off of Mastodon.
@scottjenson "Ai journalists" are not a "marginalized group", and trying to use that term to describe people who take chatbots too seriously is, IMO, fucking offensive to members of actual marginalised groups.
I don't know if you're conscious of how you're using a tired right-wing tactic often used to invade and divide leftist spaces; co-opting their language, mis-applying it to their own gain, and then declaring any pushback is actually not leftist, in the hope of leading to in-fighting among people with different levels of familiarity about the nuances of that language.
Here, you're taking "inclusivity" (which is about making sure people who are traditionally left out for factors like race, gender, sexual orientation, disability, appearance, age...), mis-applying it to tech bro's opinions LLMs (opinions aren't people, my dude), and declaring that if we aren't willing to listen to pro-LLM people waffle on, they "aren't inclusive". And that's bullshit.
Mate, let me be clear; when I say I do not want to hear your excuses for how chatbots are definitely very good, that OK they're actually bad but they're going to be good very soon, honest, that every kid who killed themselves because ChatGPT agreed with them that's it's for the best is just misuse on their part and doesn't mean there's a problem, that it was OK to suck up an entire generation's collective conversations and artworks and songs and videos because they put them up somewhere publically before LLM training was even a thing, that all the energy and water and time and effort and lives spent on this is definitely going to be worth it any day now... I do that in a completely inclusive way. *Everyone* earnestly foisting this tech despite being in a position to know all the harm it's causing can get fucked, every one of them.
I have space for people who are (somehow) using it in 2026 without realising the issues, or who are at workplaces that are forcing it on them and aren't in a position to push back. But so called "AI Journalists"? Get in the bin.
-
> See? Anybody can pull that out.
Out of where?-)
I am curious how you came to your opinion. Your opinion is open to being challenged, and AFAICT you opinion about how @scottjenson was glossing over ethica4issues with AI.
This I think is the point, and what Scott and I are pushing back against (me anyway, Scott speaks for himself). In 400 characters a lot of nuance is omitted and it is (IMO) unhelpful to draw conclusions about peoples motivations if not clearly stated
@worik @unlofl @scottjenson shame that the 500 character limit is considered a core Mastodon feature, try using a different instance
-
As this conversation is spiraling a bit I want to make a few things clear:
1. I'd like Mastodon to be MORE inclusive and bring in more voices
2. Some people don't seem to want that
3. This is core problem to solve: How do we let more in, but not "pollute" your feed?
4. The solution is NOT "gatekeeping", revelling in the fact that AI journalists aren't welcome
5. This is the same reason we lost "Black Twitter" when it came over in 2022Yes, a lot of you don't want AI posts in your feed (or pick any other topic) but the solution isn't to keep "AI People" from joining Mastodon, any more than it is keeping marginalized communities off of Mastodon.
I want my normie friends and family to join the fediverse, because their socmed of choice is facebook and I haven't touched that in over two years.
-
R AodeRelay shared this topic
-
RE: https://social.growyourown.services/@FediFollows/116274521264107379
Besides, there's no attitude against journalists here. There may be against PR shills, but some of my earliest follows from October 2022 are active and doing well. Fedi isn't their exclusive channel, but for many it is the insurance strategy. Hat tip to @molly0xfff on documenting that path and @mike and all of @Flipboard on holding the flag high. The journalists are too many to list here but @FediFollows has a "starter pack" for any niche.
@jdp23 @scottjenson @unchartedworlds@osma @molly0xfff @mike @Flipboard @FediFollows @jdp23 @scottjenson @unchartedworlds and there's a great, regularly updated list here: https://docs.google.com/spreadsheets/d/1uWj0j_AL6YQMK87U7_CFpvudK-Aygtx7Bea3fbjxgyo/edit?gid=876665327#gid=876665327
Today there's 1309 listed. Not all are still active, and I guess it's a "small" number in the grand scheme of things, but significant nonetheless....
-
@dzwiedziu @tiotasram @scottjenson @tael
Yes, it can.
If you're using SafeTensors, the model can absolutely include the tensors required for image analysis. The output gets converted to visual tokens in the language space. Basically the image analyzer tensors are an 'expert' in a MoE like system.
If you're using gguf with llama, then the mmproj- files are the separated visual tensors. This allows using bf16 image with a heavily reduced LLM, like a int4 model.
I'm also still learning a ton of this on the fly, by running and administering LLM operations on my network. As a systems engineer, this is the state of the art, and I absolutely will learn to run it for myself and any company I end up at.
@crankylinuxuser
> If you're using SafeTensors, the model can absolutely include the tensors required for image analysis. The output gets converted to visual tokens in the language space. Basically the image analyzer tensors are an 'expert' in a MoE like system.Sooo, the LLM for CV tasks requres an additional part to store tensors (SafeTensors) to be populated with visual tokens.
And as far as I've skimmed and understood »
-
@crankylinuxuser
> If you're using SafeTensors, the model can absolutely include the tensors required for image analysis. The output gets converted to visual tokens in the language space. Basically the image analyzer tensors are an 'expert' in a MoE like system.Sooo, the LLM for CV tasks requres an additional part to store tensors (SafeTensors) to be populated with visual tokens.
And as far as I've skimmed and understood »
« research papers those still need to be produced by an additional step, that suspiciously looks like CV.
…which in turn may be superseded by the existence of LVLMs (Large Visual Language Models), which makes this exercise a Ship of Theseus one.
…and looking at disk price graphs less pleasant than looking into the Ark of the Covenant…
-
@octothorpe @trisweb @cratermoon @scottjenson @Gargron I’ve personally seen a mix, certainly more heavily negative, on AI since I’ve been here. Time is precious. It’s a waste of my time to interact with an ephemeral LLM and I only want human interaction in a social context.
I do use AI at work and with local models to learn at home. My personal use runs on 3 GPUs that consume a max of 175, 125 and 35 watts. I find it interesting but I don’t expect anyone else to.
I would suggest the negativity is all on the LLM type of AI - I don’t personally see anyone being negative on the ML variants. Marketing LLMs as AI implies to me they’re pushing an AGI narrative which is simply false. These models lack cognition and are simply good enough at predicting the next token to fool many into believing they do.
@pwloftus @octothorpe @trisweb @cratermoon @scottjenson @Gargron not just the LLM side but I think especially with the big foundation models / platforms and their downstream chatbot peddlers.
That stuff is just so obviously for the exec friendly "disrupters" that never seem to disrupt in the direction of more money or power for people who have to work to survive.
-
@cratermoon @scottjenson @Gargron nobody has to earn their place on the Fediverse. Access to free and open social networking is a human right. You don't get to gatekeep this network.
@evan @cratermoon @scottjenson @Gargron You could gatekeep a part of the network if you are moderator for it, which often means you're one of the hosters.
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login