A recent blog post, but the model came out a bit ago. And I'm glad that there is work on smaller models, but just so we're clear, it's still 7 billion parameters. So it's still large, but I guess not as large as some of the crazy large things that we're seeing. Yeah. It depends on what your trade-offs are in terms of the model parameters being large or having such a large retrieval database of those two trillion words.