Can the open-source community achieve AGI AIs with Llama 2?

Llama 2 democratizes the way we all interact with AI.

by Flavius Floare

Flavius is a writer and a media content producer with a particular interest in technology, gaming, media, film and storytelling. He enjoys spending time in nature and at… read more

Updated on July 20, 2023

Could Llama 2 somehow encourage the open-source community to work together toward achieving AGI? Well, one can dream, but the Microsoft-Meta AI partnership is definitely a big step for it.

If you don’t know, at the Microsoft Inspire Conference 2023, something really exciting for the AI community happened.  Microsoft joined forces with Meta to form an AI partnership with Llama 2.

Today, at Microsoft Inspire, Meta and Microsoft announced support for the Llama 2 family of large language models (LLMs) on Azure and Windows. Llama 2 is designed to enable developers and organizations to build generative AI-powered tools and experiences. Meta and Microsoft share a commitment to democratizing AI and its benefits and we are excited that Meta is taking an open approach with Llama 2. We offer developers choice in the types of models they build on, supporting open and frontier models and are thrilled to be Meta’s preferred partner as they release their new version of Llama 2 to commercial customers for the first time.

Microsoft

And so Llama 2 made its way into the world. It will be open-source and it makes it possible for developers and organizations to build generative AI tools and generative AI experiences. Plus, you’ll be able to fine-tune and deploy diverse Llama 2 models into your own AI tools. You’ll have at your disposal Llama 2 variants that can reach a 70B-parameter limit.

That means a lot of training and even more context length than ever before. And with Microsoft AI models, such as LongMem, one could wonder if we can truly achieve AGI now.

Achieving an open-source AGI with Llama 2

If you don’t know, AGI stands for Artificial General Intelligence, and it equals the human mind. In other words, an AI with AGI will be exactly like a human person. It will be able to think by itself, make conscious decisions, and rationalize events. Just like you do.

AGI and ASI (which stands for Artificial Super Intelligence) have been the subject of discussions when it comes to AI. For example, the recently-founded Superalignment project aims to achieve AGI four years from now. They say once that is done, ASI will soon follow, and a lot of humanity’s problems will quickly be solved.

So are there other ways? Well, Llama 2 is an incredible model. There is no doubt Meta and Microsoft will update it even further, and if it’s open-source, then people can actually build more performant AI models as time goes on.

With such an increased volume of parameters, Llama 2 can offer developers the tools and training to create AI models that can be trained to surpass the volume. They will also be able to come up with new ways to engage, train, and learn.

The open-source community will learn only if it will keep being open-source, and let other people improve on their AI models.

And if projects such as Superalignment collaborates with Meta and Microsoft, then AGI will be here even sooner.

Llama 2 is really a breakthrough in AI because it democratizes the way we all interact with it. While the future might seem scary with an AGI AI, there are possibilities to make this right.

What do you think? Let us know your thoughts and opinions in the comments section below.