Trends
Meta unveils Llama 3 to catch up with OpenAI
Meta Platforms unveiled early versions of its most recent large language model, Llama 3 in an attempt to overtake OpenAI, the leader in the generative AI market, on Thursday, 18 April. About Llama 3 According to Meta Chief Product Officer Chris Cox in an interview, the social media behemoth gave Lla…
Headline
Meta Platforms unveiled early versions of its most recent large language model, Llama 3 in an attempt to overtake OpenAI, the leader in the generative AI market, on Thursday, 18 April. About Llama 3 According to Meta Chief Product Officer Chris Cox in an interview, the social…
Context
Meta Platforms unveiled early versions of its most recent large language model, Llama 3 in an attempt to overtake OpenAI, the leader in the generative AI market, on Thursday, 18 April. According to Meta Chief Product Officer Chris Cox in an interview, the social media behemoth gave Llama 3 new computer coding skills and fed it images in addition to text this time. However, for the time being, the model will only produce text and subsequent versions will include more sophisticated reasoning, such as the capacity to create longer multi-step plans. According to blog posts from Meta, versions that are scheduled for release in the upcoming months will also possess “multimodality,” or the ability to produce both text and images.
Evidence
Pending intelligence enrichment.
Analysis
Cox said using photos to train Llama 3 will improve an update that was released this year for the Ray-Ban Meta smart glasses, which is a collaboration with the eyewear manufacturer Essilor Luxoticca. This update will allow Meta AI to recognize objects seen by the wearer and respond to inquiries about them. Meta CEO Mark Zuckerberg said Llama 3’s largest version is presently undergoing training with 400 billion parameters and has achieved a score of 85 MMLU, according to metrics that indicate the potency and calibre of AI models. Eight billion parameters and seventy billion parameters make up the two smaller versions that are currently being released; the latter scored approximately 82 MMLU or Massive Multitask Language Understanding. Also read: Artificial intelligence (AI) in everyday life Also read: WhatsApp trials Meta AI in India and other markets
Key Points
- Meta Platforms introduces Llama 3, the early version of its most recent large language model on Thursday, 18 April.
- To race to catch up to generative AI market leader OpenAI, Meta integrates Llama 3 into its virtual assistant Meta AI.
- Meta has equipped Llama 3 with new computer coding capabilities and fed it images as well as text this time.
Actions
Pending intelligence enrichment.
