@hirad@hirad.it
@hirad@hirad.it avatar

hirad

@hirad@hirad.it

Metalhead, #Socialist, #Linux user and #FreeSoftware advocate. Love #Selfhosting and #Aviation.

This profile is from a federated server and may be incomplete. Browse more on the original instance.

ianRobinson, to llm
@ianRobinson@mastodon.social avatar

The hype and utility about LLMs are overstated and will cause problems due to leadership teams in organisations (and Governments) buying into the hype.

However, they do have value as personal assistants, research assistants, and sounding boards as long as you treat all LLM output critically, especially on topics where you are not an expert.

I’m using Claude 3 Opus as a research assistant. It’s read more of the world’s info than I ever will. I am also trying out ChatGPT-4o.

hirad,
@hirad@hirad.it avatar

@ianRobinson Among the commercial ones that are available online, Claude is definitely the best one.
Though I personally run LLMs on my own PC 90% of the time.

metacosm, to random
@metacosm@mastodon.social avatar

Generative AI epitomizes the worse tendencies of our current society. It’s simply not worth the cost and only benefits a minority when the majority is paying the cost in terms of environmental, societal and mental wellness impacts.

The efforts spent on generative AI would probably be better served on solving actual problems instead of faking human competency but it’d be too easy if that was something techbros could fathom.

hirad,
@hirad@hirad.it avatar

@metacosm unfortunately you and many other people are only focusing on proprietary services and completely ignore what's happening in open source generative AI community.

I only wonder if it's intentional or comes from lack of knowledge.

hirad,
@hirad@hirad.it avatar

@metacosm what cost? Don't tell me "environment"!
Also what wrong problems exactly? It'll help if you are a bit more clear on these parts.

hirad, to Horizon
@hirad@hirad.it avatar

I've started playing #Horizon Zero Dawn in the last few days and I've gotta say, it's amazing. The incredible graphics and intriguing story that makes me wanna play more to learn about what is actually going on, but at the same time to spend my time enjoy and explore the environment as well.

#VideoGames #LinuxGaming

kellogh, to opensource
@kellogh@hachyderm.io avatar

#opensource WizardLM2 8x22B exceeds performance of GPT4 in some benchmarks

  • Apache2 👍
  • progressive learning instead of all-at-once means less power-hungry and more data efficient during training
  • Co-Teaching and Self-Teaching are intriguing, I want to hear more
  • from Microsoft #AI, I imagine GPT5 must be nigh, if they’re releasing competition for GPT4

https://wizardlm.github.io/WizardLM2/ #LLMs

hirad,
@hirad@hirad.it avatar

@me1000 @kellogh According to their own statement on twitter, they had mistakenly released it before the so called "toxicity test".

hirad, to socialism
@hirad@hirad.it avatar

Correct me if I’m wrong, but Russia is the only country that both men and women gained the right to vote at the same time.
#Equality #Socialism

thelinuxEXP, to random
@thelinuxEXP@mastodon.social avatar

All it would take for AI to completely collapse is a ruling in the US saying these companies have to licence the content they used to train these tools.

They simply would never reach a sustainable business model if they had to fairly compensate all the people who wrote, drew, edited, sang or just created the content they use.

Simply being forced to respect attribution and licenses would kill them. Will that ruling ever happen? Maybe not. Should it? I think so.

hirad,
@hirad@hirad.it avatar

@thelinuxEXP honestly, I don't think that's necessary. Training a LLM isn't the same as using copyright materials. That's like saying if I copy paste your this post into a text file on my computer requires me to pay you for it!
Instead, I'd argue to give incentives to companies to release their LLMs publicly, Like Meta and Mistral do.
Unless you are truly looking for killing generative AI, in which case, we can't have any discussion. But I can say throughout history, every new tech had faced people who thought it was their duty to destroy that technology no matter the cost.

CartyBoston, to mastodon
@CartyBoston@mastodon.roundpond.net avatar

You follow too many people, that's why you see posts you don't want to see.

If you disagree with someone about something important and their posts annoy you, don't follow them.

If someone you follow boosts posts you don't like, unfollow them too.

Be careful about following hash tags.

Be careful about following people simply because you share an instance.

Voilà, your #Mastodon network is fixed, "no algorithm" is your friend.

hirad,
@hirad@hirad.it avatar

@CartyBoston or...
If someone you follow boosts posts you don't like, select hide boosts from that user.

hirad, to KDE
@hirad@hirad.it avatar

WTF @kde? Who translated the name of the months in Persian calendar?! They're correct in English, but 6 out of 12 of them are wrong in Persian!

ErikUden, to random
@ErikUden@mastodon.de avatar

In around 13 hours Mastodon is projected to have 15.000.000 Users!

https://mastodon.social/@mastodonusercount/112110208856842868

Another million, steadily growing :mastodance:

hirad,
@hirad@hirad.it avatar

@ErikUden That's not what fedidb.org says. Probably this bot also counts all the accounts and servers that have been long deleted as well. Not number of current accounts.
Also, what matters the most is number of active users, which is just about 1m.

hirad, to llm
@hirad@hirad.it avatar

I downloaded and tried Dolphin 2.7 Mixtral 8x7b today. So far, it's performing well. It's not too slow, averaging around 7.5 tokens per second. Tomorrow, I plan to test its performance in different topics. Perhaps I should also compare it to the other models I've tried.
#LLM #LocalLLaMA #AI #GenerativeAI

hirad,
@hirad@hirad.it avatar

@nosherwan Intel i5-11400F with 32gb ram and Nvidia 3060.

hirad,
@hirad@hirad.it avatar

@nosherwan the one I got requires almost 30gb. You could try 2 or 3 bits ones, but they're not recommended.

https://huggingface.co/TheBloke/dolphin-2.7-mixtral-8x7b-GGUF

hirad,
@hirad@hirad.it avatar

@nosherwan
There's a lot of models you can use. You should just try some and see which one fits your needs best.
Beside this Dolphin Mistral, I've also used Wizardlm 13b, Mistral 7b and Nous Hermes 13b. They've all been decent. And you should be able to run most 7b or 13b models without issue.

igisho, to mastodon
@igisho@rockosbasilisk.com avatar

🧵 Thread 1/3. What about making Mastodon an invite-only platform. This would mean each member can invite others, but there's a twist: invitees' profiles link back to their inviter.

hirad,
@hirad@hirad.it avatar

@igisho Wow! Punishing someone for the crime of someone else! What a fantastic idea!

stefan, to mastodon
@stefan@stefanbohacek.online avatar
hirad,
@hirad@hirad.it avatar

@stefan that's a really bad idea. The only way I can think of that it can be done is by allowing the admin to choose such community members. And it'll only be shown in that instance.
Random admins, with random moderators, deciding what is right and what is wrong. What can possibly go wrong!

hirad,
@hirad@hirad.it avatar

@stefan but "some" level of trust isn't enough. Yes, one might believe his instance admin is a fair and reasonable person. But doesn't mean he trust him on every political, social, scientific, etc etc topic.

joel, to wordpress
@joel@fosstodon.org avatar

So, #WordPress (.com) is evil now, but is there anything stopping AI scraping of the #fediverse?

Wasn't #Tumblr going to federate? I am all for just blocking them now that these things are happening, but what's stopping others? Couldn't there be servers out there just silently taking away all of this?

We also have #RSS feeds for every user profile and hashtag, a feature I love, but a double edged sword.

I really want us to remain open and such, but they make it really hard...

hirad,
@hirad@hirad.it avatar

@joel This is why it’s important that people perceive the web as fish tank. Public for everyone to see, and be careful with what you share.

retiolus, to mastodon
@retiolus@mamot.fr avatar

I've already thought about setting up a server that only works between certain times of the day...

Do you know ? Well, the same thing. At a random time, for 10 minutes a day, you're allowed to post. And that's it.

People would receive a notification, through or whatever, to let them know when that moment is.

hirad,
@hirad@hirad.it avatar

@retiolus this is gonna put so much pressure on the network and other servers. Because servers will constantly try to push the federated contents to your instance and fail, cuz it's down all the time.

BeAware, to fediverse
@BeAware@social.beaware.live avatar

deleted_by_author

  • Loading...
  • hirad,
    @hirad@hirad.it avatar

    @BeAware have you joined any relays? Since if you have, considering how all posts are stored locally in the database, deleting your own posts is probably not gonna make any relevant difference.

    hirad, to macos
    @hirad@hirad.it avatar

    Plasma 6 has "borrowed" a feature from . the ability to search through menubar items, when using global menu.

    hirad,
    @hirad@hirad.it avatar

    @danialbehzadi interesting. I don't recall seeing this search function on Unity. But to be fair the last time I saw Unity was more than 15 years ago.

    hirad, to linux
    @hirad@hirad.it avatar
    hirad, to StableDiffusion
    @hirad@hirad.it avatar

    The idea of the #StableDiffusion of how a "beautiful #Iranian woman" looks like!
    #GenerativeAI

    hirad, to HeavyMetal
    @hirad@hirad.it avatar
  • All
  • Subscribed
  • Moderated
  • Favorites
  • normalnudes
  • rosin
  • ngwrru68w68
  • GTA5RPClips
  • DreamBathrooms
  • mdbf
  • magazineikmin
  • thenastyranch
  • Youngstown
  • Durango
  • slotface
  • everett
  • vwfavf
  • kavyap
  • megavids
  • Leos
  • khanakhh
  • cisconetworking
  • cubers
  • InstantRegret
  • ethstaker
  • osvaldo12
  • modclub
  • anitta
  • provamag3
  • tacticalgear
  • tester
  • JUstTest
  • All magazines