The Big Y #168
LLM business plans
Hi! 👋 Welcome to The Big Y!
Everyone has been talking about Llama 2, Meta’s updated version of their open-source LLM which you do not need to apply for access to anymore. With OpenAI tightly guarding their secret sauce for ChatGPT, it is an interesting (but positive) path for Meta to take to open up access to Llama 2 for both research and commercial use. This open source move from Meta reinforces my idea that the applications built on top of LLMs will be much more important than the models themselves, but also brings up questions about the long-term money making strategy for the companies behind the LLMs.
With the launch of Llama 2 as open source, it doesn’t give Meta a direct income, and we know that Facebook revenue is what is paying for the development and maintenance of the models their teams are working on, so what’s the long-term play here? Alongside the Llama 2 announcement, was an announcement that they’ve made a deal with Qualcomm to launch Llama on their devices in 2024, which moves us in the direction of AI applications running directly on devices rather than in the cloud. For Meta, enabling LLM powered apps that run on devices is better for them than relying on cloud providers to provide the infrastructure to run the models.
Speaking of cloud providers… It seems like Microsoft has their fingers in every LLM on the market, firing on all cylinders to ensure that all LLMs are run in Azure. They don’t care which model you’re using as long as it is powered by Azure infrastructure, again focused more on the application layer than on the specific model. With Google, we know they have a lot of development going on behind closed doors, but they’re really keeping it close to themselves so it will be interesting to see what their plan is outside the (not so great) Bard pieces they’ve slowly put out. And then lastly, it really seems like Amazon is totally lost. There was a small mention in the Meta announcement that you can access Llama through the AWS Marketplace, but it seems like they haven’t used their advantage as the biggest cloud provider to make a move within the generative AI space at all.
The Tidbit: A handful of leading tech companies in the US have committed to new voluntary standards for safety, security and trust at the White House. In theory, this is good, but it will be interesting to see if this has any measurable impact other than a positive marketing spin for these companies.
Know someone who might enjoy this newsletter? Share it with them and help spread the word!
Thanks for reading! Have a great week! 😁
🎙 The Big Y Podcast: Listen on Spotify, Apple Podcasts, Stitcher, Substack