Chief AI Officer, @cohere Professor of Computer Science, @mcgillu Core academic member, @Mila_Quebec Ex-Meta (FAIR team)
- Tweets 128
- Following 478
- Followers 15,525
- Likes 467
I’m thrilled to be joining @cohere in the role of Chief AI Officer, helping advance cutting-edge research and product development.
Cohere has an incredible team and mission. Exciting new chapter for me!
We’re excited to announce $500M in new funding to accelerate our global expansion and build the next generation of enterprise AI technology!
We are also welcoming two additions to our leadership team: Joelle Pineau as Chief AI Officer and Francois Chadwick as Chief Financial Officer.
cohere.com/blog/august-2025-…
123
71
1,686
180,650
This seems worthy of a first tweet! BB3 is the first 175B param, publicly available chatbot, with model weights, code, datasets, model cards, and focus on safety & explainability. See the Meta AI blog for more ai.facebook.com/blog/blender…. And try out the demo blenderbot.ai
26
141
943
Introducing LLaMA: A foundational, 65-billion-parameter large language model ai.facebook.com/blog/large-l…
9
106
791
141,463
Llama 2 is out! ai.meta.com/llama
This new version has better language generation, more layers of safety, a broad set of partners – and a license that authorizes commercial use.
I continue to believe that an open approach is the right path to build better models!
9
88
480
85,136
How many of the signatories have used their expertise, their influence, their resources to rigourously analyze and expose the risks of current AI systems, and to build solutions that address real harms? That is what we need. Not letters, headlines and inflammatory statements.
We’ve released a statement on the risk of extinction from AI.
Signatories include:
- Three Turing Award winners
- Authors of the standard textbooks on AI/DL/RL
- CEOs and Execs from OpenAI, Microsoft, Google, Google DeepMind, Anthropic
- Many more
safe.ai/statement-on-ai-risk
23
65
463
156,215
The notion of "object" (or "concept") is contextual and controllable. Segment Anything allows a greater degree of expressivity than anything we've done before, and is open-sourced for anyone to try. Exciting to see what this will enable, in terms of broader scientific discovery!
Today we're releasing the Segment Anything Model (SAM) — a step toward the first foundation model for image segmentation.
SAM is capable of one-click segmentation of any object from any photo or video + zero-shot transfer to other segmentation tasks ➡️ bit.ly/433YuBI
4
44
397
114,602
There are so many good use cases for large models. Great to see the Galactica project from FAIR coming out, with our usual high standards for open science: paper, model, demo, and more!
🪐 Introducing Galactica. A large language model for science.
Can summarize academic literature, solve math problems, generate Wiki articles, write scientific code, annotate molecules and proteins, and more.
Explore and get weights: galactica.org
2
29
390
We just released Meta Llama 3: the most capable openly available LLM available to date! The 8B & 70B models are out now, and we expect to release models with larger context windows, additional model sizes and more capabilities in the coming months.
Introducing Meta Llama 3: the most capable openly available LLM to date.
Today we’re releasing 8B & 70B models that deliver on new capabilities such as improved reasoning and set a new state-of-the-art for models of their sizes.
Today's release includes the first two Llama 3 models — in the coming months we expect to introduce new capabilities, longer context windows, additional model sizes and enhanced performance + the Llama 3 research paper for the community to learn from our work.
More details ➡️ go.fb.me/i2y41n
Download Llama 3 ➡️ go.fb.me/ct2xko
11
34
375
53,302
I’m excited to share a few things we’re releasing today at Meta FAIR. These new AI model and dataset releases are part of our longstanding commitment to open science and I look forward to sharing even more work like this from the brilliant minds at FAIR!
ai.meta.com/blog/meta-fair-r…
10
68
359
270,693
Honored to be elected to the Royal Society of Canada. My biggest gratitude to all my wonderful colleagues at McGill, Mila and Meta who have made this possible - and have been by my side throughout this journey in research!
@jpineau1 was recently elected a Fellow of the Royal Society of Canada! Prof. Pineau, a Canada @CIFAR_News AI Chair and VP AI Research at Meta, is recognized for her deep and prolific contributions to research in machine learning.
Learn more : pulse.ly/me3zt0d2e9
29
18
308
52,941
I'm excited to be joining the board of the Laude Institute! We need more support and incentives for university researchers who have great ideas and early results to accelerate their work, and build new real-world solutions that have a meaningful impact on people and society.
Today, I’m launching a deeply personal project. I’m betting $100M that we can help computer scientists create more upside impact for humanity.
Built for and by researchers, including @JeffDean & @jpineau1 on the board, @LaudeInstitute catalyzes research with real-world impact.
25
21
300
49,996
Another strong model released by our FAIR research team. This model is more general than Segment Anything (released earlier this month) and just as full of potential to power new applications of AI!
Announced by Mark Zuckerberg this morning — today we're releasing DINOv2, the first method for training computer vision models that uses self-supervised learning to achieve results matching or exceeding industry standards.
More on this new work ➡️ bit.ly/3GQnIKf
2
26
289
47,637
Open Source AI is the Path Forward
llama.meta.com/
Llama 3.1 release includes a new 405B model, improved 8B & 70B, 128K context length, multilingual support, new safety tools, full research paper.
9
38
278
24,449
Happy to share details about Stable Signature, an invisible watermarking technique to distinguish images created by open source generative AI models, from FAIR &INRIA. It's another step towards building tools for safety and responsibility in generative AI.
ai.meta.com/blog/stable-sign…
9
57
272
86,736
Glad to see the OpenAI team contributing to the OSS community! A lot of great work is enabled by having more and better open-weight models.
gpt-oss is out!
we made an open model that performs at the level of o4-mini and runs on a high-end laptop (WTF!!)
(and a smaller one that runs on a phone).
super proud of the team; big triumph of technology.
6
14
266
35,902
We're shipping the first models in the Llama 4 herd! Llama 4 Scout and Llama 4 Maverick, are our first MoE models, natively multimodal and are our most advanced yet — open-sourced as always. ai.meta.com/blog/llama-4-mul…
6
32
264
19,330
Exciting day today, as we're launching Code Llama, an extension of Llama 2 specialized for code generation, and natural language related to code.
As per our commitment to openness, it comes with paper, code, model weights, responsible use guide, and the same license as Llama 2!
Today we’re releasing Code Llama, a large language model built on top of Llama 2, fine-tuned for coding & state-of-the-art for publicly available coding tools.
Keeping with our open approach, Code Llama is publicly-available now for both research & commercial use.
More ⬇️
3
20
256
48,526
I’m excited to share a new AI Coding Competition from Meta and Microsoft Research building on Meta’s annual Hacker Cup! The most capable LLMs to date will be challenged to solve these questions. We invite major players in the code generation space to join.
Hacker Cup – one of the preeminent coding competitions started an AI track w/ Meta & Microsoft
problems are hardddd – only a handful of engineers reliably solve them – requires deep algorithmic knowledge, reasoning, planning and fast execution – to solve 5 problems in 30 minutes, multiple times over 5 rounds. As of today, AI fails miserably.
We hope this competition can be a big venue that rallies academic labs, industrial researchers and students to meaningfully advance AI coding capabilities.
Would love to see more major players in the code generation space join the Hacker Cup AI code competition – @openAI, @anthropicAI, @googledeepmind, @cognition_labs @cursor_ai @replit @sourcegraphCody @codeiumdev.
Register for the practice round Starting 20th: facebook.com/codingcompetiti…
The Discord community has grown to >2k folks. Partnering with Autogen, Langchain and SWE-agent – there are multiple starter kits showcasing baseline accuracy of modern models on Hacker Cup problems.
Join the Discord community for further rules and details discord.gg/gSupVz7bMc
4
37
238
60,322
Excited to announce #CICERObyMetaAI! Human-level performance in Diplomacy is a step towards AI systems that can collaborate and negotiate through natural language in communities of agents.
Meta AI presents CICERO — the first AI to achieve human-level performance in Diplomacy, a strategy game which requires building trust, negotiating and cooperating with multiple players.
Learn more about #CICERObyMetaAI: bit.ly/3GBwLzx
3
27
225
We just shared a new larger and improved Code Llama 70B parameter model. With every release, we have better generation quality, but also a better sense of how to make these models safe.
CodeLlama-70B-Instruct achieves 67.8 on HumanEval, making it one of the highest performing open models available today.
CodeLlama-70B is the most performant base for fine-tuning code generation models and we’re excited for the community to build on this work.
2
35
207
40,268
Thrilled to have Kaggle's partnership to support the ML reproducibility challenge!
We’re excited to announce that Kaggle will be sponsoring $500k in awards for outstanding papers and reviewers in the Machine Learning Reproducibility Challenge 2022. kaggle.com/reproducibility-c…
2
14
199
74,991
Thrilled to share new AI research advances from the FAIR team at Meta. These releases highlight our continued commitment towards open research.
Read our blog post for details. ai.meta.com/blog/meta-fair-u…
2
28
186
16,649
More than 25 years in tech, I should be used to this. Yet the gender imbalance on Twitter still managed to take me completely by surprise!
9
1
156
Today we announced FAIR’s latest open-source AI research, including code model and paper, that expands speech technology by up 40x! Excited to see what this will enable in terms of language preservation and accessibility.
Today we're sharing new progress on our AI speech work. Our Massively Multilingual Speech (MMS) project has now scaled speech-to-text & text-to-speech to support over 1,100 languages — a 10x increase from previous work.
Details + access to new pretrained models ⬇️
1
17
153
27,847
It's been a big year for AI, and today I'm excited to share nine new open source releases from Meta FAIR to wrap up the year — all part of our mission to achieve advanced machine intelligence (AMI). So proud of this team and looking forward to 2025! ai.meta.com/blog/meta-fair-u…
1
17
149
13,793
Two years ago we released 'No Language Left Behind', a first-of-its-kind project for neural machine translation directly across 200 languages. This has fuelled countless new innovations.
Today the work was published in @Nature — congratulations @costajussamarta + the NLLB team!
Newly published today in @Nature: No Language Left Behind (NLLB) is an AI model created by researchers at Meta capable of delivering high-quality translations directly between 200 languages – including low-resource languages.
Read more in Nature ⬇️
go.fb.me/7j8o4p
13
15
150
19,947
For 10+ years Meta open sourced quality models - many of them by the FAIR team.
Today, European companies, academics, developers are calling on the EU to embrace the potential of AI and create a regulatory environment that fosters responsible innovation.
EUneedsAI.com
4
16
147
39,448
Very excited to share our latest open source advancements in embodied AI from Meta FAIR: new work across touch perception, dexterity, and human interaction with robots. Crucially important capabilities on the journey towards Advanced Machine Intelligence. ai.meta.com/blog/fair-roboti…
4
27
148
33,131
Enough with the hyperbole from all sides! I'm glad to have participated in this work taking a principled approach towards assessing the societal impact of Open Foundation Models. There is a lot more work to do of course, but this is a step in the right direction.
A recent MIT study claimed open models can help create bioweapons. But it didn’t test if they’re more useful than just having internet access (and later studies found they aren’t).
How can we assess the impact of open foundation models? New paper: crfm.stanford.edu/open-fms
ALT Screenshot of the first page of the paper: https://crfm.stanford.edu/open-fms/paper.pdf
1
22
126
54,452
Today we’re celebrating the 10-year anniversary of the FAIR team! In the past decade, the field of AI has undergone a profound transformation, and it’s been the adventure of a lifetime to grow and support this amazing team to where it is today. Looking forward to the next decade!
10 years of FAIR.
10 years of advancing the state of the art in AI through open research.
We're celebrating the 10th anniversary of Meta's Fundamental AI Research team and continuing that legacy by sharing our work on three exciting new research projects today.
Details below 🧵
4
10
131
21,488
Congratulations @mziizm! Can’t wait to start at Cohere and work closely with you and the Cohere Labs team.
I'm excited to share that I'll be stepping into the role of Head of Cohere Labs. It's an honor and a responsibility to lead such an extraordinary group of researchers pushing the boundaries of AI research.
4
2
136
19,214
Top 3 most cited AI papers of 2023, Llama, Llama2, SegmentAnything. All from @AIatMeta
3
13
125
12,197
Our AI infrastructure serves as the compute backbone for all the great models we train at Meta AI / FAIR ! ai.facebook.com/blog/meta-ai…
21
125
25,456
If being an outlier means building a diverse team, leaning into openness, and having a more collaborative vision for AGI, I'll take it!
Making open-source, free AI models a priority has made Meta an outlier in the AI race.
But there’s another contributing factor to Meta’s against-the-grain approach, which has gone mostly unnoticed, even inside the company, @ReedAlbergotti reports.
semafor.com/article/02/02/20…
2
12
109
43,976
This is one of the most comprehensive, nuanced and inclusive frameworks for responsibly building and deploying AI models through an open approach. PAI's leadership has been invaluable in bringing together many different opinions and offering clear guidance for AI model builders.
Foundation models have become a cornerstone of AI research and development. We need a dynamic framework to enable responsible innovation while ensuring transparency.
Today, we’re proud to announce PAI's Guidance for Safe Foundation Model Deployment: partnershiponai.org/modeldep…
4
26
105
84,104
A big step for PyTorch, and for opening up governance of key AI technology!
A big moment for a key piece of open infrastructure powering the AI revolution. Introducing the #PyTorchFoundation!
ai.facebook.com/blog/pytorch…
6
104
We dropped another awesome open model: SAM 2. This one comes with the data and an easy-to-use demo. It extends the original Segment Anything Model, to work on video. Enjoy!
Introducing Meta Segment Anything Model 2 (SAM 2) — the first unified model for real-time, promptable object segmentation in images & videos.
SAM 2 is available today under Apache 2.0 so that anyone can use it to build their own experiences
Details ➡️ go.fb.me/p749s5
4
18
107
8,635
Another step towards better understanding of multi-agent learning with complex interactions. So many interesting unexplored questions in this space!
After building on years of work from MILA, DeepMind, ourselves, and others, our AIs are now expert-human-level in no-press Diplomacy and Hanabi! Unlike Go and Dota, Diplomacy/Hanabi involve *cooperation*, which breaks naive RL. arxiv.org/abs/2210.05492 arxiv.org/abs/2210.05125 🧵👇
11
98
Our latest work on self-supervised learning, including release of full paper, code and models.
It is through open and transparent research that we will collectively make the most progress.
I-JEPA: Efficient method for Self-Supervised Learning of image features.
No need for data augmentation, just masking.
Joint embedding predictive architecture, not generative.
And it's open source, of course.
Blog: ai.facebook.com/blog/yann-le…
Paper: arxiv.org/abs/2301.08243
Code & models: github.com/facebookresearch/…
2
14
96
19,000
Sharing some of our latest work on generative AI!
The video editing features and sound generation are especially exciting. And it comes with a full research paper.
🎥 Today we’re premiering Meta Movie Gen: the most advanced media foundation models to-date.
Developed by AI research teams at Meta, Movie Gen delivers state-of-the-art results across a range of capabilities. We’re excited for the potential of this line of research to usher in entirely new possibilities for casual creators and creative professionals alike.
More details and examples of what Movie Gen can do ➡️ go.fb.me/kx1nqm
🛠️ Movie Gen models and capabilities
Movie Gen Video: 30B parameter transformer model that can generate high-quality and high-definition images and videos from a single text prompt.
Movie Gen Audio: A 13B parameter transformer model that can take a video input along with optional text prompts for controllability to generate high-fidelity audio synced to the video. It can generate ambient sound, instrumental background music and foley sound — delivering state-of-the-art results in audio quality, video-to-audio alignment and text-to-audio alignment.
Precise video editing: Using a generated or existing video and accompanying text instructions as an input it can perform localized edits such as adding, removing or replacing elements — or global changes like background or style changes.
Personalized videos: Using an image of a person and a text prompt, the model can generate a video with state-of-the-art results on character preservation and natural movement in video.
We’re continuing to work closely with creative professionals from across the field to integrate their feedback as we work towards a potential release. We look forward to sharing more on this work and the creative possibilities it will enable in the future.
3
13
90
9,277
Thrilled to share that I've been appointed Vice Chair of the Board of Directors at Partnership on AI along with Jatin Aythora. I'm excited to continue supporting PAI's work and mission, in particular its deep commitment to nurturing a safe, fair, and transparent AI future.
Building a future where AI empowers humanity isn't a one-person job. Fortunately, we have an incredible Board of Directors to help realize our vision. Today, we're excited to announce new changes to our Board: partnershiponai.org/partners…
2
3
89
14,530
Hard to imagine how we went from building AI systems that only a handful of researchers would consider interesting, to helping build our Llama models, which today are in the hands of over 1B people! This is the power of open source – and of amazingly talented teams across Meta.
Llama has surpassed 1 billion downloads 🚀 And we would not be where we are today without the contributions of the entire open source AI community – thanks to each of you who helped make Llama better. We threw together a short recap of some of the amazing things we have built together the past two years and can’t wait to see what is to come. So thank you, a billion times, for the partnership, enthusiasm, and making Llama better every single day. 🦙 🦙 🦙
5
12
83
8,587
Cohere’s models and the fabulous @JayAlammar are hard at work, to help us explore all that NeurIPS 2025 has to offer!
The Illustrated NeurIPS 2025: A Visual Map of the AI Frontier
New blog post!
NeurIPS 2025 papers are out—and it’s a lot to take in. This visualization lets you explore the entire research landscape interactively, with clusters, summaries, and @cohere LLM-generated explanations that make the field easier to grasp.
Link in thread!
1
11
85
14,718
We just announced SeamlessM4T. This new model moves us closer to all-in-one translation, with ability to handle speech and text across many different types of translation tasks and languages.
Paper, code and demo available, as per Meta AI’s continued commitment to open research!
Introducing SeamlessM4T, the first all-in-one, multilingual multimodal translation model.
This single model can perform tasks across speech-to-text, speech-to-speech, text-to-text translation & speech recognition for up to 100 languages depending on the task.
Details ⬇️
12
80
14,242
So proud of @nikhilaravi for being named a 2024 Young Engineer of the Year! She has done amazing work at FAIR, including our Segment Anything Model, and there's so much more to come! Congratulations!!
raeng.org.uk/news/royal-acad…
3
11
80
8,355
Open sourcing AI models, with proper safety measures in place, makes sense for the science, for the business, and for society.
If you have questions about why Meta open-sources its AI, here's a clear answer in Meta's earnings call today from @finkd
3
6
75
30,991
PyTorch 2.0 announced today at the #PyTorchConference! Speed-up your training with out-of-the-box Compiled mode. The result of a strong partnership between Meta's #PyTorch team and the @linuxfoundation.
We just introduced PyTorch 2.0 at the #PyTorchConference, introducing torch.compile!
Available in the nightlies today, stable release Early March 2023.
Read the full post: bit.ly/3VNysOA
🧵below!
1/5
4
69
Excited to share updates about a new collaboration with the Bibliothèque Nationale de France, Fisheye Immersive, and the artist Ruben Fro, who used our SegmentAnything2 model, to produce a breathtaking new work called "Deep Diving".
2
10
74
8,775
As part of our work towards Advanced Machine Intelligence I’m excited to share a few new open source releases and milestones from our team at Meta FAIR!
ai.meta.com/blog/machine-int…
3
18
69
6,875
New results from the FAIR team on generative speech across multiple tasks. This model is based on Flow Matching techniques. Worth looking at the samples in the blog post!
Introducing Voicebox, a new breakthrough generative speech system based on Flow Matching, a new method proposed by Meta AI. It can synthesize speech across six languages, perform noise removal, edit content, transfer audio style & more.
More details on this work & examples ⬇️
6
67
11,780
Proud of this work coming out of a collaboration between FAIR researchers & Georgia Tech that was just published in @ACSCentSci! The dataset & AI models that we released together could help meaningfully accelerate research combating climate change. open-dac.github.io/
1
17
66
9,032
New work from FAIR exploring multimodal understanding and generation, all in a single model.
I’m excited to announce our latest paper, introducing a family of early-fusion token-in token-out (gpt4o….), models capable of interleaved text and image understanding and generation.
arxiv.org/abs/2405.09818
2
10
62
9,195
Love the thought-experiment! And love the work even more! Many interesting findings in this one, including on memory, and on the use of sensor-based regularization.
A thought-experiment to inspire scientists is to ask:
If you could write only 20 papers in your lifetime, would your current work be one of them?
This is one of my 20.
arxiv.org/abs/2301.13261
wijmans.xyz/publication/eom/
🧵👇
5
61
24,353
Congratulations @Ahmad_Al_Dahle + the entire GenAI team + all of the incredible XFN across @AIatMeta that came together to ship these new models today! 👏
It’s here! Meet Llama 3, our latest generation of models that is setting a new standard for state-of-the art performance and efficiency for openly available LLMs.
Key highlights
• 8B and 70B parameter openly available pre-trained and fine-tuned models.
• Trained on more than 15T tokens, 7x+ larger than Llama 2's dataset!
• Improved tokenizer with vocabulary of 128K tokens for better performance.
• State-of-the-art performance across industry benchmarks.
• New capabilities, including enhanced reasoning and coding.
• 3x more efficient training than Llama 2.
• New trust and safety tools with Llama Guard 2, Code Shield, and CyberSec Eval 2.
• Integrated into Meta AI, and available in more countries across our apps.
• And, just the beginning with more models and new capabilities coming soon!
Visit the Llama 3 website to read more and download the models. llama.meta.com/llama3
2
56
8,555
Exciting updates from our FAIR team in Paris! The progress in generative models opens up the ability to decode human perception directly from brain activity with remarkable expressivity!
Today we're sharing new research that brings us one step closer to real-time decoding of image perception from brain activity.
Using MEG, this AI system can decode the unfolding of visual representations in the brain with an unprecedented temporal resolution.
More details ⬇️
8
57
11,221
We've also updated SAM to make it more easily accessible to developers and artists everywhere. SAM 2.1 is now available on Amazon SageMaker JumpStart, making it easier to deploy and integrate into your applications and workflows. Check out updates here: ai.meta.com/blog/segment-any…
2
8
51
42,896
Exciting to see machine translation research going beyond written language, starting with Hokkien! The accompanying demo and open-source dataset should open the door to many more efforts in that direction.
(1/3) Until now, AI translation has focused mainly on written languages. Universal Speech Translator (UST) is the 1st AI-powered speech-to-speech translation system for a primarily oral language, translating Hokkien, one of many primarily spoken languages. bit.ly/3CJP3ew
10
51
I am eager to engage with the community through this new ML Commons AI safety initiative, as we strive for a safer AI ecosystem. This effort is highly complementary to the release earlier this week of PAI's Guidance for Safe Foundation Model Deployment.
Announcing the formation of the AI Safety working group. Making #AI safer for everyone. mlcommons.org/en/news/format… @MLCommons @AnthropicAI @CoactiveAI @Google @IntelAI @intelnews @inflectionAI @AIatMeta @Microsoft @OpenAI @NVIDIAA @joavanschoren @percyliang
4
33
41,735
Welcome to Meta / FAIR @TacoCohen!
Two weeks ago I joined Meta / FAIR, and I couldn't be more excited about this new chapter. Meta is indeed the only place left that supports highly ambitious long-term oriented & fundamental research projects and has a strong commitment to open science and open source. (and has hundreds of thousands of H100s lying around)
In my short tenure here I've already met tons of super talented, ambitious and kind folks, and I'm blown away by all the futuristic tech being developed, as well as the sheer scale of Meta's operations.
I will be joining @syhw & team to work on code generation. Codegen is obviously an important application domain, but it also provides a great setting for research on general methods for planning, search, reasoning, etc. Stay tuned!
1
3
43
9,610
It is absolutely crucial that we define a path for science- and evidence-based AI policy. I am happy to be working with this stellar group of AI experts to define the way forward.
Debates over AI Policy like CA SB-1047 highlight fragmentation in the AI community. How can we develop AI policies that help foster innovation while mitigating risks? We propose a path for science- and evidence-based AI policy: understanding-ai-safety.org
1
9
40
6,339
Quick update on BB3 release blenderbot.ai/ Over 70K chats so far. From feedback by participants, 0.11% of responses were tagged inappropriate, 1.36% nonsensical, 1% off-topic. Leaving the cherry-picking to others, lots of low-hanging fruits. But that's how we learn!
2
1
39
Don’t miss Imagine Flash, released as part of the Llama3 launch. It’s pretty awesome!
Check out the generative vision related release too
meta.ai/?icebreaker=imagine
Imagine Flash generates the image as you type
You can also "Animate" your images! (technique based on Emu Video emu-video.metademolab.com/)
Kudos to the team for putting this out :)
2
6
39
9,092
Glad to see Meta highlighting the benefits of AI, and the value of an open approach.
At Meta, we believe in the power of AI to advance the state of the art and to connect the world. We’re proud to sign this letter with @svangel today and are committed to our open approach to AI ⬇️
openletter.svangel.com/
5
36
5,867
Connecting the world, doing high quality research, and sharing it openly. I couldn't be more proud of this work, from the FAIR team!
We're proud to share that our work on SeamlessM4T — the first all-in-one, multilingual multimodal translation model — was recognized today as part of @TIME's Best Inventions of 2023 list!
More details and the full #TIMEBestInventions list ➡️ bit.ly/3Qd7Cyn
2
36
3,936
Should I wish for a moth avatar 😀
1
37
Couldn’t agree more with @rajiinio on this! I have seen many cases where smaller more narrow models are easier to weaponize than larger more general models. The corollary is that we may need to discuss regulation of all AI models, depending on potential for actual harm.
This is in my opinion the critical flaw of the EO and a lot of the current framing of AI regulation in general. Bigger models are not necessarily more capable or more dangerous. There's no definitive evidence of this. Gating regulatory control by model size is a serious mistake.
3
6
35
15,955
At last, a letter on AI which I am happy to co-sign!
The future of #AI governance is here. Help us shape AI for the benefit of people, not just corporations.
Join a global community of policymakers, engineers, and more in ensuring that AI remains open and accessible to all ➡️ open.mozilla.org/letter
2
29
8,776
I'm strongly supportive of this letter and its core message. We need a nuanced approach to the risks and benefits of AI, and more transparency is key to enable a wide group of stakeholders to join the conversation!
.@CenDemTech joins @mozilla & nearly 50 orgs + scholars in a letter to @SecRaimondo, urging her to protect openness & transparency in #AI in response to @NTIAgov public consultation to examine risks/benefits/potential policy approaches re: open AI models:
cdt.org/insights/cdt-joins-m…
7
33
6,964
End of an era. Thank you for everything you have done! PyTorch will go. And I'm sure your curiosity will lead you to new interesting places!
33
2,255
Interesting analysis on the impact of open LLMs.
The enterprise verdict on AI models: Why open source will win venturebeat.com/ai/the-enter…
6
31
5,033
Lots of new exciting updates to the Llama family!
📣 Introducing Llama 3.2: Lightweight models for edge devices, vision models and more!
What’s new?
• Llama 3.2 1B & 3B models deliver state-of-the-art capabilities for their class for several on-device use cases — with support for @Arm, @MediaTek & @Qualcomm on day one.
• Llama 3.2 11B & 90B vision models deliver performance competitive with leading closed models — and can be used as drop-in replacements for Llama 3.1 8B & 70B.
• New Llama Guard models to support multimodal use cases and edge deployments.
• The first official distro of Llama Stack simplifies and supercharges the way developers & enterprises can build around Llama to support agentic applications and more.
Details in the full announcement ➡️ go.fb.me/229ug4
Download Llama 3.2 models ➡️ go.fb.me/w63yfd
These models are available to download now directly from Meta and @HuggingFace — and will be available across offerings from 25+ partners that are rolling out starting today, including @accenture, @awscloud, @AMD, @azure, @Databricks, @Dell, @Deloitte, @FireworksAI_HQ, @GoogleCloud, @GroqInc, @IBMwatsonx, @Infosys, @Intel, @kaggle, @NVIDIA, @OracleCloud, @PwC, @scale_AI, @snowflakeDB, @togethercompute and more.
With Llama 3.2 we’re making it possible to run Llama in even more places, with even more flexible capabilities. We’ve said it before and we’ll say it again: open source AI is how we ensure that these innovations reflect the global community they’re built for and benefit everyone. We’re continuing our drive to make open source the standard with Llama 3.2.
4
29
3,189
We are releasing three more foundation models, for music and audio. It’s been fun to play with these models, and generate a variety of new outputs. There’s still so much more to do, and I’m glad we continue to embrace open research to foster innovation.
about.fb.com/news/2023/08/au…
5
22
1,725
A lot of amazing work being done by people around the world building on our open weight Llama models & tools!
Today we announced the recipients of the inaugural Llama Impact Grants and the Llama Impact Innovation Awards. It’s incredible to see the work people are doing with our open source Llama models to address major societal challenges, and I’m excited to see what kind of innovation this inspires moving forward. ai.meta.com/blog/llama-impac…
2
18
2,934
Indeed, there are people in the list doing really good work on this. This is what we should amplify!
3
12
2,568
As AI advances, a global standard will emerge, shaping technology, infrastructure, manufacturing, finance, and e-commerce. The transparency, openness, and accountability that Llama provides is essential for responsible AI development.
about.fb.com/news/2024/11/op…
1
3
9
2,177
These numbers should be taken with a (large) grain of salt. There is likely selection bias in what is flagged and by who. We'll need to see more and analyze to get reliable estimates.
1
11
Great to see the first batch of Slingshots projects from the Laude Institute! We often under-value the work on benchmarks and frameworks, but this is the very foundation of all good AI systems.
Meet Slingshots // One.
This inaugural batch includes leading-edge researchers advancing the science and practice of AI - with benchmarks, frameworks, and agents that ship real impact into the world.
We're honored to support research from:
@alexgshaw @Mike_A_Merrill @lschmidt3 @andykonwinski @lateinteraction @a1zhang @GregKamradt @mikeknoop @fchollet @jyangballin @KLieret @Diyi_Yang @tyler_griggs_ @profjoeyg @istoica05 @DachengLi177 @shiyi_c98 @matei_zaharia @sijun_tan @ralucaadapopa @ying11231 @lm_zheng @Tim_Dettmers @Muennighoff @AndrewYNg @YejinChoinka @atharva_sehgal @yisongyue @ChengleiSi @tatsu_hashimoto @KexinHuang5 @lisabdunlap @JacobSteinhardt @LakshyAAAgrawal
This first batch sets a VERY high bar, and we’re proud to call these researchers part of the Laude community.
More details here: laude.org/updates/slingshots…
14
4,521