Meta says that Llama 3 is the best model
Llama 3 is a marathon, not a sprint: What tomorrow will openAI teach us about AI and how to make a killer margarita?
Even if Meta regains its title as leader of the open-sourced leaderboard with Llama 3, it won’t be known for certain what tomorrow will bring. OpenAI is rumored to be readying GPT-5, which could leapfrog the rest of the industry again. When I ask him about this, he says that Meta is already thinking of Llama 4, 5. To him, it’s a marathon and not a sprint.
“I don’t think that anything at the level that what we or others in the field are working on in the next year is really in the ballpark of those type of risks,” he says. I think we will be able to open source it.
A key focus for Llama 3 had to do with decreasing its false refusals or the number of times model says it can’t answer a prompt that is actually harmless. An example Zuckerberg offers is asking it to make a “killer margarita.” I gave him another one during the interview that I gave him when the earliest version of Meta AI wouldn’t tell me how to break up with someone.
An example is llama 3, which is a model of artificial intelligence. The biggest version of Llama 2, released last year, had 70 billion parameters, whereas the coming large version of Llama 3 will have over 400 billion, Zuckerberg says. The big version of Llama 3 has over 15 trillion token while the smaller version has 2 trillion token. OpenAI has yet to publicly confirm the number of parameters.
Meta AI Assistant: Features and Prompt Suggestions for AI AI-Aided Chatbots in Popular Social Networks and Apps
The Meta artificial intelligence assistant is now being integrated into the search box of several popular social networks. It is also going to show up in the main Facebook feed. The Meta apps have messaging inboxes for you to talk to it in. For the first time it can be reached via a website called Meta.ai.
Stories and Reels were both pioneered by upstarts, and then tacked onto Meta’s apps in order to make them even more ubiquitous.
While it has only been available in the US, Meta Artificial Intelligence is now also available in several other countries, including Australia, Canada, Germany, Jamaica, New Zealand, Nigeria, Singapore, South Africa, Uganda, and Zimbabwe. It’s a far cry from Zuckerberg’s pitch of a truly global AI assistant, but this wider release gets Meta AI closer to eventually reaching the company’s more than 3 billion daily users.
The Meta AI assistant is the only chatbot I know of that now integrates real-time search results from both Bing and Google — Meta decides when either search engine is used to answer a prompt. Its image generation has also been upgraded to create animations (essentially GIFs), and high-res images now generate on the fly as you type. Meanwhile, a Perplexity-inspired panel of prompt suggestions when you first open a chat window is meant to “demystify what a general-purpose chatbot can do,” says Meta’s head of generative AI, Ahmad Al-Dahle.
“This evaluation set contains 1,800 prompts that cover 12 key use cases: asking for advice, brainstorming, classification, closed question answering, coding, creative writing, extraction, inhabiting a character/persona, open question answering, reasoning, rewriting, and summarization,” Meta says in its blog post.
It should be mentioned that benchmark testing models is not perfect and is helpful in understanding how powerful they are. The datasets used to benchmark models have been found to be part of a model’s training, meaning the model already knows the answers to the questions evaluators will ask it.
