this post was submitted on 20 May 2025
1483 points (98.2% liked)

Microblog Memes

7658 readers
3229 users here now

A place to share screenshots of Microblog posts, whether from Mastodon, tumblr, ~~Twitter~~ X, KBin, Threads or elsewhere.

Created as an evolution of White People Twitter and other tweet-capture subreddits.

Rules:

  1. Please put at least one word relevant to the post in the post title.
  2. Be nice.
  3. No advertising, brand promotion or guerilla marketing.
  4. Posters are encouraged to link to the toot or tweet etc in the description of posts.

Related communities:

founded 2 years ago
MODERATORS
(page 2) 50 comments
sorted by: hot top controversial new old
[–] d00ery@lemmy.world 41 points 16 hours ago (1 children)

Somewhat ironic that the avatar looks AI generated.

[–] AngryCommieKender@lemmy.world 6 points 13 hours ago

It looks like the avatars in the mobile game Hogwarts Mystery.

[–] huppakee@lemm.ee 59 points 18 hours ago (4 children)

I wish functions baked in to browsers could be disabled like an extension, this adding ai to everything is getting as bad as all the bloatware you get on a new PC

load more comments (4 replies)
[–] kshade@lemmy.world 8 points 14 hours ago (6 children)

Generally agree, I do appreciate Firefox' built-in translation tool though, that also falls under "AI" I guess.

load more comments (6 replies)
[–] Sivecano@lemmy.dbzer0.com 12 points 15 hours ago (1 children)
[–] HellsBelle@sh.itjust.works 8 points 14 hours ago (1 children)

DDG started with this bs yesterday and it drove me nuts.

[–] nilaus@lemmy.world 5 points 14 hours ago (1 children)

At least it is easy to disable in ddg. Read your comment opened browser. 10seconds later all ai features disabled.

[–] HellsBelle@sh.itjust.works 7 points 13 hours ago (1 children)

Yeah. Still pissed me off I had to do it at all.

If I want AI, I will search and dl. It shouldn't be added to any browser without permission.

load more comments (1 replies)
[–] cm0002@lemmy.world 25 points 18 hours ago (3 children)

I wouldn't mind a decent LOCAL open source AI helping

[–] TheTechnician27@lemmy.world 12 points 17 hours ago* (last edited 16 hours ago) (2 children)

Large X models lack a crucial component of "open-source". Freely redistributable and modifiable for any purpose, sure, but there's no chance in hell of auditing one, let alone if the training data is kept a secret. It's literally impossible; human beings cannot look at a trillion weights and biases representing a single highly chaotic, unfathomably complex nonlinear function whose input and output space are the totality of human language/images/etc. and say "yup, looks good to me." Deep learning models – contrasted with traditional machine learning models – learn their own features which almost 100% of the time would be nonsense to a human. You just have a blob of shareware when you run DeepSeek.

(They also just outright steal from billions of copyright-protected sources to create it, so calling it "open-source" is pretty funny.)

[–] cm0002@lemmy.world 7 points 16 hours ago

Auditing for bias purposes, yea true. But my primary concern is it having the capability to "phone home" which you don't really need to audit the model itself to be able to detect or prevent

load more comments (1 replies)
[–] Areldyb@lemmy.world 6 points 15 hours ago (1 children)

Firefox can use a local llamafile model, but you have to enable it in about:config first.

[–] gamermanh@lemmy.dbzer0.com 1 points 10 hours ago

Honestly it's easier to find an addon that'll hook to ollama instead, fire fox's inbuilt support is shit

[–] thatKamGuy@sh.itjust.works 7 points 17 hours ago (2 children)

DeepSeek’s model is open-sourced and can be run locally; though I think there some bits related to its training data they have been kept obscured (if I remember correctly) - likely due to the dubious nature of how it was acquired.

[–] rImITywR@lemmy.world 9 points 17 hours ago

Unless training data is made available, a model is not open source. DeepSeek is better described as "open weight".

[–] brucethemoose@lemmy.world 2 points 13 hours ago* (last edited 13 hours ago)

some bits related to its training data

AKA ANY details about its training data, and its training hyperparameters, and literally any other details about its training. An 'open' secret among LLM tinkerers is that the Chinese companies seem to have particularly strong English/Chinese training data (not so much other languages though), and I'll give you one guess on how.

Deepseek is unusal in that they are open sourcing the general techniques they used and even some (not all) of the software frameworks they use.

Don't get me wrong, I think any level of openness should be encouraged (unlike OpenAI being as closed as physically possible), but they are still very closed. Unlike, say, IBM Granite models which should be reproducible.

[–] MudMan@fedia.io 19 points 18 hours ago (2 children)

I'm far from an AI hater, but I fully agree with this.

I think there's a distinct business oppotunity coming up for two things: Hassle-free self-hosting and back-to-basics apps and services.

Nobody is tapping into those correctly (you're going to want to give me examples of self-hosted things, and you're wrong), and it's extremely hard to do either right, but if you can figure it out and are ballsy enough to build a proper business around it I may be interested in your pitch deck.

[–] Turret3857@infosec.pub 11 points 17 hours ago (1 children)

Can you elaborate on "Hassle-free self-hosting" & "and you're wrong"

genuinely curious to see what your argument is here.

[–] MudMan@fedia.io 7 points 17 hours ago (6 children)

Kinda not the point, but at the risk of starting a huge tangent: yes, there are a bunch of self-hosted applications that are reasonably practical and easy to install, but there's still the layer of having to understand how to access a thing in your LAN from each device, and ideally you'd want some sort of dedicated server running at all times and a bunch of this stuff is provided in multiple formats, including containerized versions or versions for virtual machines, all of which is way over the heads of normie users.

The closest to a fire-and-forget self-hosting platform is maybe Home Assistant or perhaps some of the commercial NAS sellers, like the Synology suite of apps that will mooostly set themselves up. Maybe Plex. But even those don't work in quite the way mainstream users think about applications working. You really need something you plug in and it goes. Maybe the branded Home Assistant hardware is closest to that, but HA itself is so overengineered and customizable it's not so much the start of a commercial self-hosting revolution as a relatively accessible hobby project rabbit hole.

load more comments (6 replies)
load more comments (1 replies)
[–] HappySkullsplitter@lemmy.world 4 points 13 hours ago

This is how I felt when Windows 3.1 dropped

[–] dadarobot@lemmy.sdf.org 11 points 18 hours ago (2 children)

yeah i noticed yesterday duckduckgo browser has ai now

[–] mj_marathon@programming.dev 15 points 18 hours ago (1 children)

DDG lets you turn it off completely fwiw

load more comments (1 replies)
[–] joyjoy@lemm.ee 10 points 18 hours ago (1 children)

Its not surprising. Duckduckgo search has ai.

[–] Gullible@sh.itjust.works 9 points 17 hours ago (1 children)

Which- why? Who’s using ddg without understanding how to use a search engine or recognizing the constant AI hallucinations?

[–] null_dot@lemmy.dbzer0.com 5 points 16 hours ago

The short answer is employees and family members.

Someone who manages tech for other users might configure ddg as default search. I guess people at ddg are concerned that this type of user might be resistant to using ddg unless it has zero-click results.

[–] dream_weasel@sh.itjust.works 6 points 16 hours ago (2 children)

I don't mind seeing an AI summary of search results as much as I mind sponsored links fucking up page rank. Sometimes it is even nice to see "hey your search doesn't make sense because you've conflated two terms". But I guess I'm in the minority.

Reminds me of early wikipedia when there was a deep trustworthiness problem. Seeing a wikipedia link on a presentation stole your credibility, but it was still a hell of a lot better starting point than grabbing an encyclopedia and asking jeeves until you found a thread to pull.

[–] Whats_your_reasoning@lemmy.world 13 points 15 hours ago (9 children)

AI summaries put another layer of interpretation between the reader and the source material. When having accurate and properly-sourced information matters, it's just not trustworthy enough. At least with Wikipedia, it tells you when there is potentially biased or improperly sourced material. Search AI will confidently assert their summaries as though they are factual, regardless of how reliable or unreliable their own sources are.

load more comments (9 replies)
[–] RoidingOldMan@lemmy.world 9 points 16 hours ago (1 children)

I've never had a result that helpful. I've seen it make up sports results in advance though.

load more comments (1 replies)
[–] Psythik@lemm.ee 0 points 8 hours ago (1 children)

Well I do, so fuck him and fuck you too.

load more comments (1 replies)
[–] Ragdoll_X@sh.itjust.works 7 points 18 hours ago

Laughs in LibreWolf user

[–] TachyonTele@lemm.ee 8 points 18 hours ago (10 children)

What i hate about firefox is the fucking wall of links on the home page. It takes forever to remove them, and then they just updated and all that crap is back.

[–] Suck_on_my_Presence@lemmy.world 8 points 18 hours ago (2 children)

I use an extension called Tabliss and set that as my home page. I have it customized so the links to my most visited pages are set up with an icon so it's very clean and minimalist.

load more comments (2 replies)
load more comments (9 replies)
[–] mechoman444@lemmy.world 2 points 13 hours ago

Accept cookies?

load more comments
view more: ‹ prev next ›