Metas Llama 3.2 Is Much Bigger Than You Think!

[@Jayc5001]

Bro really started using an AI voice and thought we wouldn't notice. Weirdly enough it seems like a lot of people genuinely are not noticing.

[@MrMisturr]

People are here to watch a video on AI development, and you think they're worried about the narrator being an AI voice or not? I just don't think people care

[@r34ct4]

@@MrMisturr It's just low effort. i can get my AI news anywhere, I really liked his long rambles and his character. This is soul-less.

[@r34ct4]

I'm noticing... but yeah many people are not. He's writing the script with ChatGPT and using ElevenLabs for voice

[@wildstorm74]

I think over half those people just don't care. Im one of them. As long it's not robotic, like shit Ai voice overs. I really don't care less. Then again, thats my opinion. So you could be right, and people just haven't noticed.

[@coolcool2901]

​@@r34ct4😂😂

[@thedannybseries8857]

I am rooting for Meta now. No more OpenAI.

[@aeskape]

Yeah, me too, it feels amazing tbh to do things on device without any internet required and sending my personal data with this level of accuracy.

[@samson3523]

I heard we only got this because Llama got hacked & leaked early then decided to Opensource and pretended it was always the plan

[@MillionaireMindset548]

@@samson3523 Yes thats always the case

[@blackswann9555]

Sam Altman loves you money 💰

[@ALFTHADRADDAD]

Yama

[@zasta7]

Yamaha I guess lol

[@LeandroMessi]

He went correct Spanish pronunciation in this video. In honor of Tres.Dos

[@CommentGuard717]

The ai voice is sooo close

[@anubisai]

What the hell is Yama 😂

[@MemesnShet]

Llama 405B with Chain of Thought when?

[@ricpb]

And with reinforced learning to optimize the chain of thought and agentic frame, and with Tree of Thoughts using groq hardware (they achieve hundreds of tokens per second), thinking for 10 minutes would yield over 200k tokens of reasoning, resulting in 20k tokens of good-quality output tokens for each task. Groq: June 2024, 1253 tokens per second Imagine how far this could go with reinforced learning?

[@Unkn0wn1133]

They allready have it, they just agree to let open ai take the lead

[@Unkn0wn1133]

The competition is an illusion

[@bahroum69]

FIY, you can use it on Poe for the UE

[@MaxGrossenbacher]

With our 1950s minded EU politicians Orion will NEVER be sold in the EU. Same as Meta RayBan which are castrated in the EU.

[@alertbri]

I just watched an amazing interview with Gary Marcus on Machine Learning Street Talk - puts all of this into sharp perspective. Have a look, it's OUTRAGEOUS and will CHANGE EVERYTHING. 😁

[@farzadimanpoursardroudi45]

People could use Poe in Europe .

[@Matlockization]

I believe your Mistral results are old. They are doing a lot better than Meta.

[@JasonIchinco-cp7uh]

How do you actually get this?

[@Spaz42]

He couldn't even be bothered to correct the transcript errors. Eh seems lazy, hope his research isn't as lazy.

[@TheNewOriginals450]

I love the irony of your "...isn't at lazy" 🤭. Aww, you've gone and changed it now 😁.

[@user-tx9zg5mz5p]

Humans need to unionize against ai and robots...

[@alexgcuzweguz]

So we can continue spending most of our lives doing things we don't want to do? AI can free us from undesired work and allow us to live better lives.

[@TheSuperColonel]

If if we turn dysptopian it will be hell and likely a revolution. If we are turning into a Utopia we will likely need to work much less and spend more time with family etc. 80% of the ppl likely only do their jobs cause they need the paycheck to feed their families.

[@DarkSpice84]

Unionize against robots? Corporations have already made us robots. Working 50 hours a week to survive? We are already slaves.

[@DailyTuna]

Somebody needs to start up a virtual machine service that’s positioned outside the EU to run these things

[@ElaraArale]

Simply outstanding.

[@CustomComputing]

“Like a friend that can read graphs better than you can” Common man… stop sensationalizing everything

[@N1ghtR1der666]

we can somewhat run 70b models, I dont see how we can run 90b models so what does it matter if they are slightly better than lama 3

[@DailyTuna]

I always wanted to look like Woody Allen😂

[@narottamzakheim5051]

Why is UK getting lumped in with EU? What about BREXIT ffs?

[@firstlast8085]

I thought Llama 3.2 was 7 big, but it turns out its actually 8 big. Wow. Much hype. So AI.

[@pigeon_official]

nice

[@DailyTuna]

Most people are not nerds that they want to wear this on their face the whole point of AI, it to be a tool to do things for you so you can go out and do tai chi and pottery remember?. So we can spend that UBI that Altman so compassionately enables us to have😂

[@matt.stevick]

Wut??

[@Ginto_O]

The word is its bad

[@picadosinferno]

Chama!

[@will4us]

Lama3.2 90B failed on some basic and complex tasks and tests involving data analysis of diagrams and graph tables.

[@supaduppi]

Yanmar?

[@JohnsonNong]

OMG love your daddy

[@orangehatmusic225]

He's trying to get his AI slaves to call functions properly... wake up sheeple.

[@bauch16]

Still nothing compared to o1

[@emeraldmoonx1504]

It's meant to be on the level of smaller models like gpt4o-mini.

[@anubisai]

Lol how😂😂😂 can u even compare 🤣 😅 😂😂😂

[@AIEnthusiast-ht4c]

Yoooommmmaaaaa 😂

[@nono-yh2vi]

Unsubscribed I don’t want to hear a ai generated script with a ai voice

[@JaredFarrer]

Yama? You mean llama. Pronounce those L

[@Heisenberg2097]

Funky stuff... to entertain the toddlers on the next children's birthday party. A cheaper way to show the world who you are instead of wearing these ugly goggles is to take a sharpie and write the economic short five letters 'I', 'D', 'I', 'O', 'T' on your fore head so that the rest of the world is warned.

[@honkytonk4465]

?

[@nono-yh2vi]

Unsubscribed I don’t want to hear a ai generated script with a ai voice