Meta is making Llama 2 free for all
Meta introduced Llama 2, its first large language model that’s available for anyone to use, and for free. Yes, you read that right.
Compared to the previous model, or Llama 1, the newer one is way more sophisticated and is trained on 40% more data, with double the context length.
We support an open innovation approach to AI. Responsible and open innovation gives us all a stake in the AI development process, bringing visibility, scrutiny and trust to these technologies. Opening today’s Llama models will let everyone benefit from this technology.
Statement of support for Meta’s open approach to today’s AI
However, Llama 2 does share some faulties with its rivals. For example, it also has a tendency to generate false information and offensive content. Moreover, since the company didn’t share any information on the dataset used for training, it is still not known whether they have also been using copyrighted material or personal data.
On the other hand, considering that Meta has put a huge emphasis on researchers from the start when it comes to testing this model and giving feedback, and especially since it’s made open source now, it’s expected to learn important lessons to make Llama safer, less biased and more efficient. In addition, it’s already more transparent and customizable, allowing companies to create products and services faster than a big, sophisticated proprietary model.
With the … wisdom of crowds you actually make these systems safer and better and, crucially, you take them out of the … clammy hands of the big tech companies which currently are the only companies that have either the computing power or the vast reservoirs of data to build these models in the first place.
Meta’s president of global affairs Nick Clegg for BBC Radio 4
Why is this a good thing for users?
It’s no secret that Meta expects this decision to give them a bounce over their competitors, primarily ChatGPT.
Considering that Meta was a bit slower than its rivals OpenAI and Microsoft, and that it released a smaller version of Llama, restricted only to researchers, they had to come up with a big move. And nothing makes a fuss like giving away free stuff.
The question is, why would users even have an interest in testing out a new chatbot, having in mind there are already established, impressive alternatives out there?
Well, first of all, there have been noted some inconveniences with ChatGPT and Bard that users are not particularly fond of, and, over time, you can get bored of the same repetitive patterns if you’re using these models every day.
Moreover, users like to try out different options and decide for themselves which one suits them most. Even better, knowing what the available options are, users can better decide how to combine them and what are the strongest sides of each one.
This benefits the entire AI community and gives people options to go with closed-source approaches or open-source approaches for whatever suits their particular application. This is a really, really big moment for us.
Meta’s vice president Ahmad Al-Dahle for MIT Technology Review
If we compare this with browsers – yes, most users worldwide are turning to Google Chrome first, but alternatives like Edge, Brave, and Mozilla all have their benefits over Chrome.
In other words, one is not enough for different scenarios you may need AI assistance with; it’s good to be familiar with the competitors and then adjust them accordingly to your needs.