Ella Irwin

AI Product and Safety. Host of The CryRoom Podcast, dog lover. Views expressed on X are my own.

[Mustafa Suleyman @mustafasuleyman]

What I call Seemingly Conscious AI has been keeping me up at night - so let's talk about it. What it is, why I'm worried, why it matters, and why thinking about this can lead to a better vision for AI. One thing is clear: doing nothing isn't an option. 1/

[Ella Irwin @ellagirwin]

@steventaughtme Hope to be able to create more episodes later this year. The world of AI has kept me too busy to be able to record this year!

[Aaron Levie @levie]

Wow. Meta just launched its new Llama 4 open-weight model family. Pretty incredible cost/performance. Up to 10M token context windows. We're reaching the point where you're not paying a penalty with open models, but you get all the benefits control. Pretty incredible.

[Yann LeCun @ylecun]

BOOM! The Llama-4 brood is out.

[Prabhu @technogoy]

Llama 4 is here - SOTA multimodal models that are accessible across all of Meta's apps, open - you can download it on https://t.co/RapwxNetW9 and offers tremendous value. If you havent hopped on #Llama4 yet, what are you waiting for?

[AI at Meta @AIatMeta]

Today is the start of a new era of natively multimodal AI innovation.

Today, we’re introducing the first Llama 4 models: Llama 4 Scout and Llama 4 Maverick — our most advanced models yet and the best in their class for multimodality.

Llama 4 Scout
• 17B-active-parameter model with 16 experts.
• Industry-leading context window of 10M tokens.
• Outperforms Gemma 3, Gemini 2.0 Flash-Lite and Mistral 3.1 across a broad range of widely accepted benchmarks.

Llama 4 Maverick
• 17B-active-parameter model with 128 experts.
• Best-in-class image grounding with the ability to align user prompts with relevant visual concepts and anchor model responses to regions in the image.
• Outperforms GPT-4o and Gemini 2.0 Flash across a broad range of widely accepted benchmarks.
• Achieves comparable results to DeepSeek v3 on reasoning and coding — at half the active parameters.
• Unparalleled performance-to-cost ratio with a chat version scoring ELO of 1417 on LMArena.

These models are our best yet thanks to distillation from Llama 4 Behemoth, our most powerful model yet. Llama 4 Behemoth is still in training and is currently seeing results that outperform GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro on STEM-focused benchmarks. We’re excited to share more details about it even while it’s still in flight.

Read more about the first Llama 4 models, including training and benchmarks ➡️ https://t.co/9G3QgVdCkB
Download Llama 4 ➡️ https://t.co/eVomRvEr0w

[Ahmad Al-Dahle @Ahmad_Al_Dahle]

As of today, Llama 4 Maverick offers a best-in-class performance to cost ratio with an experimental chat version scoring ELO of 1417 on LMArena.

It's wild to think Llama was a research project a couple of years ago & amazing to see how much progress we've made in the last two years 🚀 And this is just the first taste of the Llama 4 collection – get ready for a herd like you’ve never seen before.

Very proud of the GenAI team.

[Ahmad Al-Dahle @Ahmad_Al_Dahle]

Introducing our first set of Llama 4 models!

We’ve been hard at work doing a complete re-design of the Llama series. I’m so excited to share it with the world today and mark another major milestone for the Llama herd as we release the *first* open source models in the Llama 4 collection 🦙. Here are some highlights:

📌 The Llama series have been re-designed to use state of the art mixture-of-experts (MoE) architecture and natively trained with multimodality. We’re dropping Llama 4 Scout & Llama 4 Maverick, and previewing Llama 4 Behemoth.

📌 Llama 4 Scout is highest performing small model with 17B activated parameters with 16 experts. It’s crazy fast, natively multimodal, and very smart. It achieves an industry leading 10M+ token context window and can also run on a single GPU!

📌 Llama 4 Maverick is the best multimodal model in its class, beating GPT-4o and Gemini 2.0 Flash across a broad range of widely reported benchmarks, while achieving comparable results to the new DeepSeek v3 on reasoning and coding – at less than half the active parameters. It offers a best-in-class performance to cost ratio with an experimental chat version scoring ELO of 1417 on LMArena. It can also run on a single host!

📌 Previewing Llama 4 Behemoth, our most powerful model yet and among the world’s smartest LLMs. Llama 4 Behemoth outperforms GPT4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro on several STEM benchmarks. Llama 4 Behemoth is still training, and we’re excited to share more details about it even while it’s still in flight.

A big thanks to all of our launch partners (full list in blog) for helping us bring Llama 4 to developers everywhere including @huggingface, @togethercompute, @SnowflakeDB, @ollama, @databricks and many others👏 This is just the start, we have more models coming and the team is really cooking – look out for Llama 4 Reasoning 😉

A few weeks ago, we celebrated Llama being downloaded over 1 billion times. Llama 4 demonstrates our long-term commitment to open source AI, the entire open source AI community, and our unwavering belief that open systems will produce the best small, mid-size and soon frontier models. Llama would be nothing without the global open source AI community & we are so ready to begin this next chapter with you. 🦙

Read more about the release here: https://t.co/7mbK3uggjO, and try it in our products today.

[Ella Irwin @ellagirwin]

Really good essay on the potential positive impact of AI. Love this: “one of my main reasons for focusing on risks is that they’re the only thing standing between us and what I see as a fundamentally positive future”. Protecting people = delivering incredible AI outcomes.

[Dario Amodei @DarioAmodei]

Machines of Loving Grace: my essay on how AI could transform the world for the better

https://t.co/CFGd4Sbq86

[Ella Irwin @ellagirwin]

So excited about MovieGen!

[Ahmad Al-Dahle @Ahmad_Al_Dahle]

I couldn’t be more excited to share our latest AI research breakthrough. We call it Meta Movie Gen and it’s a collection of state-of-the-art models that combine to deliver the most advanced video generation capability ever created.

Check it out: https://t.co/K4IxAVPO3y

[Paul Couvert @itsPaulAi]

Meta has casually released the best AI image editor

You can upload your image to Meta AI and just write the edits you want to make.

Accessible for free in WhatsApp, Instagram, Messenger, Facebook, etc.

[Linus ✦ Ekenstam @LinusEkenstam]

This is clearly the future

Meta is trippling down on it, and IMO it will work. This is the perfect form factor.

[AshutoshShrivastava @ai_for_success]

Llama 3.2 can do selective image editing, and it’s really impressive!

[Ella Irwin @ellagirwin]

@steventaughtme All good! Just out of the country on vacation until a few days ago!

[Ella Irwin @ellagirwin]

Spotify link here. https://t.co/8EQxEOqreh

[Ella Irwin @ellagirwin]

Managing up is such a loaded term. What does it even mean? How do you do it well? Had a chance to record an episode of the CryRoom podcast with @EthanEvansVP who is great at breaking this topic down. Check it out if this is a muscle you want to build. https://t.co/fU0gbUEF4F

[Vinod Khosla @vkhosla]

I am awe struck at the rate of progress of AI on all fronts. Today's expectations of capability a year from now will look silly and yet most businesses have no clue what is about to hit them in the next ten years when most rules of engagement will change. It's time to rethink/transform every business in the next decade. Read "https://t.co/RtyUiAXgOF" by @leopoldasch. I buy his assertion only a few hundred people know what is happening.

[Ella Irwin @ellagirwin]

@MailePRMedia 😬

[Ella Irwin @ellagirwin]

@MailePRMedia Hopefully I don’t disappear after liking this post. 😂

[Dark Tide Book @DarkTideBook]

Check out Edna Cowell Martin's revealing interview with @piersmorgan. Discover the chilling moment she realized her cousin, Ted Bundy, was a serial killer. Watch the full clip here: https://t.co/akrbEWlNX3 #TedBundy #TrueCrime @MyFavMurder @TrueCrimeObsess @tru_crime_tweet