Last week a hot new language model entered the AI villa. Meet LLaMA (🦙), Meta’s new large language model. LLaMA joins Sydney (Microsoft’s ChatGPT-fueled Bing) and Google’s Bard in the competition to win the public’s love on AI island.

Mark Zuckerberg announces Meta's new large language model as A.I. race heats up

Last week a hot new language model entered the AI villa. Meet LLaMA (🦙), Meta’s new large language model. LLaMA joins Sydney (Microsoft’s ChatGPT-fueled Bing) and Google’s Bard in the competition to win the public’s love on AI island. Mark Zuckerberg announces Meta's new large language model as A.I. race heats up The model, called LLaMA, is intended to help scientists and engineers explore applications for AI such as answering questions and summarizing documents. https://www.cnbc.com/2023/02/24/mark-zuckerberg-announces-meta-llama-large-language-model.html

Jerome Cody

Meta Platforms has delayed the largest version of its flagship large language model Llama 4 to as late as this fall—or even later—over performance problems, according to a person familiar with the matter.

Llama 4 has encountered other delays in the past. Meta previously pushed back Llama 4’s release date because, during its development, the model didn’t perform as well on technical benchmarks as the company had hoped. The company has also delayed a reasoning version of Llama 4, though the reason for that delay couldn’t be learned.

Meta released smaller versions of Llama 4, known as Scout and Maverick, in April. Those models had a rocky debut, in part because Meta uploaded an experimental version of Llama 4 Maverick to the popular LLM leaderboard LMArena, which performed better than the version actually released to the public.

A spokesperson for Meta did not immediately respond to a request for comment. The Wall Street Journal first reported on a delay for the largest version of Llama 4, known as Behemoth.

https://www.wsj.com/tech/ai/meta-is-delaying-the-rollout-of-its-flagship-ai-model-f4b105f7?utm_campaign=%5BREBRAND%5D+%5BTI-AM%5D+Th&utm_content=1095&utm_medium=email&utm_source=cio&utm_term=124

Meta Platforms has delayed the largest version of its flagship large language model Llama 4 to as late as this fall—or even later—over performance problems, according to a person familiar with the matter.  Llama 4 has encountered other delays in the past. Meta previously pushed back Llama 4’s release date because, during its development, the model didn’t perform as well on technical benchmarks as the company had hoped. The company has also delayed a reasoning version of Llama 4, though the reason for that delay couldn’t be learned.  Meta released smaller versions of Llama 4, known as Scout and Maverick, in April. Those models had a rocky debut, in part because Meta uploaded an experimental version of Llama 4 Maverick to the popular LLM leaderboard LMArena, which performed better than the version actually released to the public.  A spokesperson for Meta did not immediately respond to a request for comment. The Wall Street Journal first reported on a delay for the largest version of Llama 4, known as Behemoth.  https://www.wsj.com/tech/ai/meta-is-delaying-the-rollout-of-its-flagship-ai-model-f4b105f7?utm_campaign=%5BREBRAND%5D+%5BTI-AM%5D+Th&utm_content=1095&utm_medium=email&utm_source=cio&utm_term=124

The company’s struggle to improve the capabilities of latest AI model mirrors issues at some top AI companies.

Exclusive | Meta Is Delaying the Rollout of Its Flagship AI Model

Nate

We often think of these big tech companies as having enormously deep pockets, but AI development is tearing a hole in those pockets. Meta generated $91.3 billion in cash from operations last year, out of which it spent $39 billion on capital expenditures and $35 billion in stock buybacks and dividends. This year Meta has projected capex to rise 60% to between $60 billion and $65 billion. We shouldn’t assume its operational cash flow will rise 60%, so you can see where a problem might arise. Cutting buybacks wouldn’t thrill shareholders, which are surely already wondering about the value of the AI investment.

Meta last year asked Microsoft, Amazon and others to help fund the development of its Llama AI, signaling the costs of AI are mounting.

https://www.theinformation.com/articles/meta-asked-amazon-microsoft-help-fund-llama?offer=rtsu-engagement-25&cio_link_id=CIO--LINKID&utm_campaign=RTSU%3A+Microsoft+fund&utm_content=7903&utm_medium=email&utm_source=cio&utm_term=5405

Meta last year asked Microsoft, Amazon and others to help fund the development of its Llama AI, signaling the costs of AI are mounting.  https://www.theinformation.com/articles/meta-asked-amazon-microsoft-help-fund-llama?offer=rtsu-engagement-25&cio_link_id=CIO--LINKID&utm_campaign=RTSU%3A+Microsoft+fund&utm_content=7903&utm_medium=email&utm_source=cio&utm_term=5405

Meta Platforms over the past year asked Microsoft, Amazon and others to help pay the costs of training Meta’s flagship large language model, Llama, according to four people briefed on the discussions. Meta’s overtures reflected worries about the growing costs of its artificial intelligence ...

Meta Asked Amazon, Microsoft to Help Fund Llama

Sam Mendoza

Meta Platforms on Saturday released two versions of Llama 4, the newest generation of its flagship open-source large language model. Both versions, Scout and Maverick, are mixture of experts models, which are trained so that only specialized parts of a model activate at a time. at least one version of Llama 4 would be a mixture of experts model.

Meta is working on a version of Llama 4 that is a reasoning model, which takes more time to “think” about its response to a query, CEO Mark Zuckerberg said in a post on Instagram. It is also working on a much larger version, Behemoth, that will also be a mixture of experts model.

Meta said in a blog post that its goal for Llama 4 was to remove bias from the model. “It’s well-known that all leading LLMs have had issues with bias—specifically, they historically have leaned left when it comes to debated political and social topics,” it said.

Llama 4 now powers the company’s Meta AI assistant in its social media apps. Meta AI has more than 700 million monthly active users, as of January. Some of that usage has been thought to be unintentional, though.

https://ai.meta.com/blog/llama-4-multimodal-intelligence/?utm_campaign=%5BREBRAND%5D+%5BTI-AM%5D+Th&utm_content=1095&utm_medium=email&utm_source=cio&utm_term=124

Meta Platforms on Saturday released two versions of Llama 4, the newest generation of its flagship open-source large language model. Both versions, Scout and Maverick, are mixture of experts models, which are trained so that only specialized parts of a model activate at a time. at least one version of Llama 4 would be a mixture of experts model.  Meta is working on a version of Llama 4 that is a reasoning model, which takes more time to “think” about its response to a query, CEO Mark Zuckerberg said in a post on Instagram. It is also working on a much larger version, Behemoth, that will also be a mixture of experts model.  Meta said in a blog post that its goal for Llama 4 was to remove bias from the model. “It’s well-known that all leading LLMs have had issues with bias—specifically, they historically have leaned left when it comes to debated political and social topics,” it said.  Llama 4 now powers the company’s Meta AI assistant in its social media apps. Meta AI has more than 700 million monthly active users, as of January. Some of that usage has been thought to be unintentional, though.  https://ai.meta.com/blog/llama-4-multimodal-intelligence/?utm_campaign=%5BREBRAND%5D+%5BTI-AM%5D+Th&utm_content=1095&utm_medium=email&utm_source=cio&utm_term=124

We’re introducing Llama 4 Scout and Llama 4 Maverick, the first open-weight natively multimodal models with unprecedented context support and our first...

The Llama 4 herd: The beginning of a new era of natively multimodal AI innovation

OpenAI found evidence that Meta’s Llama model had been used by the Chinese government to create an AI system to find anti-China posts on social media. Looking forward to the session about this one at LlamaCon!

 #🇨🇳 ChinA.I. 🤖🧠🦾🤖 
 #A Glance of China 行摄中国 

https://www.nytimes.com/2025/02/21/technology/openai-chinese-surveillance.html?unlocked_article_code=1.zE4.VE1v.QxVmvFuBeQtX&smid=url-share&ref=platformer.news

🇨🇳 ChinA.I. 🤖🧠🦾🤖

A Glance of China 行摄中国

OpenAI found evidence that Meta’s Llama model had been used by the Chinese government to create an AI system to find anti-China posts on social media. Looking forward to the session about this one at LlamaCon!   #🇨🇳 ChinA.I. 🤖🧠🦾🤖   #A Glance of China 行摄中国   https://www.nytimes.com/2025/02/21/technology/openai-chinese-surveillance.html?unlocked_article_code=1.zE4.VE1v.QxVmvFuBeQtX&smid=url-share&ref=platformer.news

The company said a Chinese operation had built the tool to identify anti-Chinese posts on social media services in Western countries.

Jerome Cody

A group of likeminded people in Artificial Intelligence are talking about this.