TLDR:
Meta has launched an open-source AI app, Llama 3.1 405B, that they claim is competitive with closed rivals like OpenAI and Anthropic. This model allows developers to fully customize and train on new datasets without sharing data with Meta. Mark Zuckerberg believes that open-sourcing AI is necessary for a positive future, despite concerns about bad actors utilizing AI for harm. The model is currently available in the US and 21 other countries, with plans for wider accessibility in the future.
Meta launches open-source AI app ‘competitive’ with closed rivals
Meta has claimed that its new artificial intelligence model, Llama 3.1 405B, is the first open-source system that rivals products from competitors like OpenAI and Anthropic. The company states that Llama is competitive across a range of tasks and offers developers the ability to customize the model for their specific needs and applications, without sharing data with Meta.
Mark Zuckerberg, Meta’s co-founder, emphasized the importance of open-source AI for a positive future, despite concerns about potential misuse by bad actors. He believes that wider deployment of AI will enable larger actors to counter the power of smaller bad actors utilizing the technology for harmful purposes.
Currently, Llama 3.1 405B is only available to users in 22 countries, including the US. Meta reportedly avoided an EU launch due to regulatory concerns, but the open-source nature of the system suggests that it will likely be accessible to users outside the EU through other channels in the near future.
Despite Meta’s claims about the power and competitiveness of Llama 3.1 405B, third-party testing will be necessary to determine its effectiveness compared to current leaders in the field, such as GPT-4o. The open-source nature of the model allows for greater accessibility and customization, with the potential to democratize access to AI technology.