Mark Zuckerberg and Meta have really made their play in artificial intelligence announcing this week that they're going to work with Microsoft to open source their Ilama 2 model now this is a very different strategy from a lot of the other companies that are not only building models but them Building Products on top of them but there's a reason that meta is doing this andt hink this is really a scorched Earth.
Strategy the company doesn't have the same incentives as a lot of its competitors and that allows it to take a very different strategy and potentially destroy a lot of the value that companies like openai have built, So want to get into exactly what was announced and what think it means for investors.
So let's start with the announcement from meta and this is the announcement with Microsoft they're introducing a Next Generation model called llama 2. this is building on their last large language model that's where the LL comes from in llama and that one was open source but only for researchers it was not open for commercial use this one is open for commercial use.
It is going to be available for a vast majority of companies and companies who are trying to build on top of a model so there's obviously a lot of value in building an artificial intelligence model so why is meta doing this,
The big reason is they have different incentives than a lot of their competitors meta doesn't need tomake money on the model itself it needs to make money on the product so the advertising that's on Facebook and Instagram as a result it can release these models as open source and hope that developers continue to add value to the model.
Something that happens typically in open source versus closed systems this release says by making Al models openly available they can benefit everyone giving businesses startups entrepreneurs and researchers access to tools developed at a scale that would be challenging to build themselves backed by computing power that may not otherwise access will open up a world of opportunities for them to experiment innovate in exciting ways and ultimately benefit from economically and socialy.
Microsoft is going to be basically an infrastructure partner but it's not just going to be Microsoft they're working with Amazon web services hugging face there are going to be a number of other providers who are working to basically Advance the infrastructure of this model that is now open source.
Now l want to cover the caveat here because think this is important this is the license agreement that made a released with this model and this section, number two additional commercial terms and what this says is that if on the Llama 2 version release date the monthly active users of products or Services made available by a license is greater than 700 million monthly active users basically meta can deny a model or come up with some sort of Licensing agreement.
What this basically means is companies like Apple Snap TikTok a lot of the other big tech companies are not just going to be able to use this open source model freely that's something that is explicitly called out now note that their trend for these models from meta has been towards more open Source so it's possible in the future that a llama 3 or a llama 2.5 is just available to everybody.
We don't know exactly where that's going but just wanted to note that Trend this is an exciting development fora number of different reasons one think artificial intelligence is going to move more and more to on-device inferences.
Basically when you call an artificial intelligence so if you're doing a search on Chatgbt that is an reference right now most of those inferences are happening in the cloud but these Al models are not actually that big and so if we open source them developers can cut them down make them more manageable a lot of them can actually end up on devices.
We know that Qualcomm is one of the companies that has talked about already doing this with lama as early as next year so think in the future we're going to have more of these models on device that means a lot of products are going to move to the device from the cloud as well this benefits a company like meta in a number of different ways like said developers are now aoina to be able to work on this model that's now open source so they're going to be able to make the model better.
But if we're moving inferences onto the device who can tap into that well Metacam one of the challenges for tech companies is that as we move to more artificial intelligence their margins have actually been coming down because a lot of that computing power is relatively expensive especially when on a per unit basis you're talking about pennies or fractions of a penny for any given ad that's shown by a company like meta on its social networks so any cost that you can cut out of the equation is really going to be a benefit think this is one of the Hidden benefits for meta is that they're going to move a lot of that Cost onto the device so we're not necessarily going to notice it as users because the cost is really just electricity so maybe you have to plug your phone in a little bit more maybe the battery needs to be a little bit bigger.
But there's not going to be a huge cost to moving these models onto devices but there's going to be a massive benefit for a company like Meta that operates scale so they I think that's going to be a big benefit but at the end of the day Meta makes its money on the products that it has, so the underlying infrastructure it has often open sourced a lot of these things, because it doesn't need to make money on the infrastructure it doesn't need to keep them closed it actually wants to open them up and allow other developers to make them even better so that the products The Meta is making that is sitting on top of these open source platforms are better than they would be if meta had kept all of these systems closed.
Think that's really the way to think about this but if you're an Alc ompany and you're building a model or you're even building a product this could be disruptive to you becausea company like open Al is using a relatively expensive API right now companies that are building on top of that APl are going to be upended by the fact that models are going to be living on devices in the very near future.
So this is something that's playing out in real time think Meta has a very different strategy than a lot of its competitors so worth noting for investors the takeaway think is that this is going to be great for users long term more open source development is good for the advancement of artificial intelligence and for the products building on top built on top of it and meta may be a bigger player than you realize what did you think about this news leave your comments in the comments section below don't forget to follow CHALTRENDS for more Economic and finance news
Thanks you, see you next time
.jpeg)
0 Comments