2023-12-01
Big milestone - 10 years since we founded FAIR - everything from PyTorch to Segment Anything to LLama created here and available for all to use!
Forbes
An interview with Yann LeCun and Joelle Pineau ahead of the 10th anniversary of Meta's Fundamental AI Research lab on ChatGPT, AI in the metaverse, and more
In an exclusive interview leading up to the 10th anniversary of Meta's AI lab, LeCun reflects on the artificial intelligence boom …
Big milestone - 10 years since we founded FAIR - everything from PyTorch to Segment Anything to LLama created here and available for all to use!
Bloomberg
Meta Chief AI Scientist Yann LeCun says “there is really no commercial downside” to sharing Meta's AI tech openly, even as many peers take the opposite approach
pytorch, detectron, segment-anything, llama, faiss, wav2vec, biggraph, fasttext, the Cake below the cherry, and so much more. Can't say we didn't change AI and to an extent the wor...
2023-11-21
I was an exec at @mozilla when @Google decided that the browser was too critical to be run by a nonprofit they don't control. Sounds a lot like @Microsoft and @OpenAI today. This is an unstable structure at its core. Hope all my friends over there are doing ok
Bloomberg
Ilya Sutskever and over 700 out of ~770 OpenAI staffers sign a letter saying they may quit and join Sam Altman unless the board resigns and reinstates Altman
- Majority of OpenAI employees sign letter seeking new board — Board member Ilya Sutskever is among the signatories
2023-09-18
What @kevin_scott did to vault MSFT forward in #ai is really impressive. A lesson for CTOs everywhere that the job is to make sure your company has the tech in needs *not* to try to invent everything yourself .... https://www.wsj.com/...
Wall Street Journal
A profile of Microsoft CTO Kevin Scott, whose OpenAI-or-bust strategy rankled some staff who found their AI projects pushed aside and their resources curtailed
2023-09-17
What @kevin_scott did to vault MSFT forward in #ai is really impressive. A lesson for CTOs everywhere that the job is to make sure your company has the tech in needs *not* to try to invent everything yourself .... https://www.wsj.com/...
Wall Street Journal
A profile of Microsoft CTO Kevin Scott, whose OpenAI-or-bust strategy rankled some staff who found their AI projects pushed aside and their resources curtailed
Chief Technology Officer Kevin Scott spearheaded the tech titan's unlikely partnership with OpenAI — In the tech industry's artificial-intelligence race,
2023-07-19
This is indeed huge news in the world of AI! So excited to see this out in the world!
Meta
Meta releases Llama 2, its open-source LLM with double the context length, for free for research and commercial use, and expands its Microsoft partnership
Recent breakthroughs in AI, and generative AI in particular, have captured the public's imagination and demonstrated what those developing …
2023-02-25
If you are paying attention to AI you should follow this - the core advance powering many AI powered experiences is a foundational model. These are very difficult and expensive to train - so access is limited. This brings state of the art models to the entire research community https://twitter.com/...
Meta AI
Meta releases Large Language Model Meta AI, or LLaMA, a foundational LLM designed to help AI researchers, available in sizes ranging from 7B to 65B parameters
As part of Meta's commitment to open science, today we are publicly releasing LLaMA (Large Language Model Meta AI) …
2 important details: a) Shows that smaller models trained with more data can outperform larger models (e.g. 13B outperforms GPT-3 175B) 2) The “larger” 65B model is competitive with best models - and is freely available to the research community! https://research.facebook.com/ ...
Meta AI
Meta releases Large Language Model Meta AI, or LLaMA, a foundational LLM designed to help AI researchers, available in sizes ranging from 7B to 65B parameters
As part of Meta's commitment to open science, today we are publicly releasing LLaMA (Large Language Model Meta AI) …
2022-11-16
Basic internet math for @elonmusk's claim load time in India is 20s because of 1200 RPC calls (v.s. 2s in US): Bangalore->SF is ~250ms 1200 *.25s = 5 minutes! Way off!! 20s load time = MAX of 80 sequential calls Reality is likely 5-10? 1200 is happing inside datacenter? https://twitter.com/...
Washington Post
In an email to Twitter staff, Elon Musk gives an ultimatum: commit to a new “hardcore” Twitter by 5pm ET on Thursday or leave with three months severance
It comes as Musk's first major feature, Twitter Blue Verified, was halted while the company probes issues that arose from its launch.
Basic internet math for @elonmusk's claim load time in India is 20s because of 1200 RPC calls (v.s. 2s in US): Bangalore->SF is ~250ms 1200 *.25s = 5 minutes! Way off!! 20s load time = MAX of 80 sequential calls Reality is likely 5-10? 1200 is happing inside datacenter? https://twitter.com/...
Bloomberg
Elon Musk fired at least two Twitter engineers who publicly criticized him on the social network, in one case announcing the firing in a now-deleted tweet
2022-11-02
Yet another example of stunning advances: our researchers built an AI model that predicts protein structures 60X faster than existing state of the art - it took two weeks for it to create the largest database of protein structures ever assembled https://ai.facebook.com/.... . ...
Nature
Meta researchers create ESMFold, an LLM to predict the structures of 617M+ metagenomic proteins that is less accurate than DeepMind's AlphaFold but ~60x faster
Ewen Callaway / Nature :
2022-07-26
I didn't even realize we had leap seconds but our teams have a convincing augment as to why they should be left behind! https://engineering.fb.com/...
CNET
Meta, Google, Microsoft, and Amazon join the NIST's push to scrap the leap second, which keeps clocks in sync with the Earth's rotation but can cause glitches
so Meta wants to get rid of them Mariella Moon / Engadget : Meta calls for the death of the leap second Ciaran / worthreading : Worth Reading Issue #12 Leigh Mc Gowran / Silicon Re...
2022-06-03
Some big news here: FAIR, Meta's amazing AI research lab, is joining forces with Reality Labs Research, which is inventing the technologies needed for the future of computing. An awesome combination of fundamental and applied research: https://ai.facebook.com/...
Axios
Meta announces a broad AI reorganization: VP of AI Jerome Pesenti will leave later in June, the Responsible AI team will join the Social Impact team, and more
Ina Fried / Axios :
AI is becoming the foundation of pretty much everything that matters at Meta - better recommendations, platform safety, revenue, and increasingly building for the Metaverse - which will be the way most of us interact with AI every day... https://www.axios.com/...
Axios
Meta announces a broad AI reorganization: VP of AI Jerome Pesenti will leave later in June, the Responsible AI team will join the Social Impact team, and more
Ina Fried / Axios :
2022-05-20
Nobody can predict what is going to happen over the next 12 months but we haven't had a real bad tech downturn since 2000 (2008 terrible for lots of other folks but nbd for tech) let me tell you what starting a company in 2000 was like.. https://twitter.com/...
TechCrunch
In a letter to portfolio founders, Y Combinator advises to “plan for the worst” by cutting expenses and focusing on extending runways amid the economic downturn
2022-05-04
Heard of advances in AI called Large Language models - e.g. GPT-2/3? Up till now only researchers at a handful of large companies could actually see what was behind the curtain. Now Meta AI has released a fully trained model for researchers... https://www.technologyreview.com/ ...
MIT Technology Review
Meta's AI lab creates Open Pretrained Transformer, a language model trained with 175B parameters to match GPT-3's size, and gives it to researchers for free
2022-04-13
IMHO we are going to see an explosion of applications of robotics outside of factories of the next decade. Indoor navigation used to be the stuff of research conferences and now it is mostly solved... https://www.mobihealthnews.com/ ...
MobiHealthNews
Diligent Robotics, which develops a robot called Moxi to help hospital staff with routine, not-patient-facing tasks, closes a $30M Series B led by Tiger Global
Emily Olsen / MobiHealthNews :
2022-02-25
~7-8 years ago we were just getting Supervised Learning to work and @ylecun told us to focus on Self Supervision as the current approach wouldn't scale. Took us a bit to catch-up so I'd listen to him this time:) https://ai.facebook.com/...
Vox
Meta says it has been developing CAIRaoke, a self-supervised learning AI neural model used to power a voice assistant for the company's AR/VR products
Universal translator from @StarTrek is within our reach if we focus on it. Access any information, speak to any person, regardless of your native language. So awesome #AI https://ai.facebook.com/...
The Verge
Mark Zuckerberg showcases a concept for voice-controlled VR space generated by AI, plans for a universal language translator, and more in a Meta livestream
internal models of how the world works — may be the key. Learn more: https://ow.ly/... https://twitter.com/... @metaai : We're excited to share details on Project CAIRaoke, a break...
~7-8 years ago we were just getting Supervised Learning to work and @ylecun told us to focus on Self Supervision as the current approach wouldn't scale. Took us a bit to catch-up so I'd listen to him this time:) https://ai.facebook.com/...
The Verge
Mark Zuckerberg showcases a concept for voice-controlled VR space generated by AI, plans for a universal language translator, and more in a Meta livestream
internal models of how the world works — may be the key. Learn more: https://ow.ly/... https://twitter.com/... @metaai : We're excited to share details on Project CAIRaoke, a break...