Exploring the Top 12 BitTensor Subnets

Subnet Zero: Rout

First up, we have Subnet Zero, also known as Rout. This is where all the voting happens. The top 12 validators vote on governance proposals and decide how the daily emissions should be distributed among the other subnets. You can check out the live stats here, and remember, all links in this video will be in the description.

Subnet One: Open Tensor Foundation

Next, we have Subnet One, which is managed by the Open Tensor Foundation. They are working on text prompting, and I would wager that this is going to be used in many projects.

Subnet Two: Luk Cus Capital

Subnet Two is managed by Luk Cus Capital, and their subnet specializes in machine translation. Machine translation is all about using AI to translate text from one language to another. It's not just about swapping word for word; it captures the full essence of the original text into a new language. That is cool.

Subnet Three: Git Phantom

Moving on, Subnet Three by Git Phantom focuses on scraping, and I think it's going to be scraping a lot of Twitter content. We'll see. Anyway, I recommend checking out his Twitter for more updates.

Subnet Four: Manifold

Subnet Four is managed by Manifold, and they specialize in multimodality. Imagine if you're telling a story, but you can only use words; that would be one mode. But imagine using pictures, sound, and words all at the same time. That's what multimodality is all about—using multiple modes or ways to communicate a message. You can learn more about them on their website and Twitter.

Subnet Five: Creative Builds

Next, we have Subnet Five by Creative Builds, and the focus here is image generation. I think I'll be keeping an eye on this because I really like AI image generators, and they can be useful for so many different things.

Subnet Six: The Mystery Subnet

Subnet Six is a bit of a mystery. One user on Discord joked that it was his and that he was developing an AI model capable of translating the sounds of monkeys so he could ask them how it feels to have a tail. That was brilliant. Somebody even made a picture with his name on it; he really is a good and fun community member.

Subnet Seven: Salah Hawk

Subnet Seven is managed by Salah Hawk, and it specializes in storage. It has the potential to be used by many BitTensor projects. You can learn more about Salah Hawk on his personal webpage and on his GitHub.

Subnet Eight: Tshi

Next, we have Subnet Eight, managed by Tshi. The focus here is on time series prediction. Time series prediction is like having an AI crystal ball that uses historical data, like past sales numbers, to make predictions. It's like the AI is peeking into the future based on what it learned from the past. I like that.

Subnet Nine: Encon

Subnet Nine is managed by Encon, and he is working on AI pre-training. AI pre-training is like teaching it the basics before anything else. The AI is first trained on a large general dataset before it learns a specific task. This gives the AI a broad understanding of language, images, or whatever it's learning about. It's like giving the AI a head start so it can learn its specific task more effectively later on. It's a crucial step in making AI models smarter and more efficient.

Subnet Ten: Chain Dude

Chain Dude on Discord is behind Subnet Ten, and he is working on MapReduce. MapReduce is a computational model that processes large datasets by breaking them into smaller pieces, processing these pieces in parallel, and then combining the results. It's a key tool in big data and distributed computing. Unfortunately, I couldn't find any links regarding this subnet, so if anybody has any information, please leave it in the comments, and I will pin that comment.

Subnet Eleven: Encon

Subnet Eleven is another subnet managed by Encon. This is a secondary text generator/text prompt subnet that he uses mostly for testing, if I understand correctly. It's cool to see him working in several places.

Subnet Twelve: Unknown

Finally, we have Subnet Twelve, and I don't know its current purpose, at least not at the time of recording. So if anybody does know, please let me know in the comments.

  Source : @Eden.