@alex.nex.tech: Colibri just ran a 744 billion parameter AI model on a laptop with no GPU, proving you don’t need expensive hardware. This smart streaming technique uses just 10 GB of RAM at a time, making offline AI accessible and affordable for developers and enthusiasts. #AI #AITools #TechInnovation #MachineLearning #OpenSourceAI

capybara :

dawg Ur ai😭🙏

hungryLittleMonster :

So if they don't need these massive data centres for AI, what are they actually for? 😳

Jakeyyx :

You get 1 token per hour. Good luck

Brian 🦁 :

I remember when Colibri was the smallest, lightest Linux opsys with the hummingbird motif. Is this the same group?

🎤❄️Holger❄️🎤 :

Have this guy even tried colibrí??? It's Slow-Motion.

Gloryfields :

370gb....ooof

nick owi :

Do your study! It takes forever to respond when you asked “How are you!”

Ravindu :

This is practically true but as a TikToker, a content creator, it's better if you can tell your viewers the reality of using this as well. Nobody wants to sit somewhere doing nothing, letting it work for more than 5 hours just to create a few words. Please research the downsides of this and tell it to your viewers as well

Void :

yeah but what about speed?

Hr. Morningstar :

ya amazing get ready to pay subscription fee for a computer u bought...

Athe :

you are crawling it, not running it. cant beat physics

AISweeties :

Sure at 1 token per 3 to 5 business days 😅

Vladislav :

0.01 token per centuries

حنو٣١٣ :

like gaming texture process

James Barone :

A lot slower

GDPlayer1035 :

0.3 tokens per minute

تدحم :

no link in the comment

Houseofwade :

Kevin Walstrom :

This is inference, not the same

MaximeQuoi :

32+24vram possibel ?

Colviop🍩 :

sure, about 5 tokens per day.

Rock Lobster🇨🇦 :

StudioVerg :

but must to be very slow