The first ‘Fairly Trained’ AI large language model is here (venturebeat.com)

submitted 7 months ago by kromem@lemmy.world to c/technology@lemmy.world

7 comments fedilink hide all child comments

top 7 comments

sorted by: hot top controversial new old

[-] GenderNeutralBro@lemmy.sdf.org 22 points 7 months ago

Weird that they act like the 1.7B model is too big for a laptop, in contrast to a...4060 with the same amount of memory as that laptop. A 1.7B model is well within range of what you can run on a MacBook Air.

I don't think a 170M model is even useful for the same class of applications. Could be good for real-time applications though.

Looking forward to testing these, if they are ever made publicly available.

[-] agressivelyPassive@feddit.de 4 points 7 months ago

They wrote "cheap m1 MacBook". That's 8gb ram in total. You can't reasonably compare a GPU with 8gb of dedicated RAM to 8gb of shared main memory.

[-] FaceDeer@fedia.io 8 points 7 months ago

I wonder where the goalposts will shift to next.

[-] TheOneCurly@lemm.ee 2 points 7 months ago

The goal posts of ... respecting basic copyright?

[-] FaceDeer@fedia.io 8 points 7 months ago* (last edited 7 months ago)

No, the goal posts of "AI is evil and should be fought until are resolved."

The question of whether training an AI even violates copyright in the first place is still unanswered, BTW, the various court cases addressing it are still in progress. This current target is about "ethics", which are vague enough that anyone can claim that they're being violated without having to go to the hassle of proving it.

[-] JackGreenEarth@lemm.ee 2 points 7 months ago

AGI is scary because it can't be aligned with human values, seemingly leading to a Universal Paperclips/Nick Bostrom style scenario. The AI we have today is a tool, that seems on the whole beneficial.

[-] A_Very_Big_Fan@lemmy.world 1 points 7 months ago

The only argument against it that I think is valid is that it shouldn't be outputting copyrighted content, which seems like a pretty easy problem to solve but I'm no expert. If YouTube can do it with the volume of data they work with, it can't be that hard to do it with text and pictures.

But the idea that an AI model just looking at a picture is somehow stealing is just absurd to me.

this post was submitted on 21 Mar 2024

79 points (90.7% liked)

Technology

59144 readers

2451 users here now

This is a most excellent place for technology news and articles.

Our Rules

Follow the lemmy.world rules.
Only tech related content.
Be excellent to each another!
Mod approved content bots can post up to 10 articles per day.
Threads asking for personal tech support may be deleted.
Politics threads may be removed.
No memes allowed as posts, OK to post as comments.
Only approved bots from the list below, to ask if your bot can be added please contact us.
Check for duplicates before posting, duplicates may be removed

Approved Bots

founded 1 year ago

MODERATORS