It's AI
@ai_detection
Followers
715
Following
0
Media
2
Statuses
23
Easily detect LLM generated content using decentralized AI.
SN32
Joined March 2024
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
#MAGA
• 501470 Tweets
Harris
• 489141 Tweets
Kamala
• 429571 Tweets
The Boys
• 294141 Tweets
Leeds
• 263925 Tweets
Hulk Hogan
• 154474 Tweets
Tucker
• 122103 Tweets
Bob Newhart
• 107329 Tweets
Lou Dobbs
• 85697 Tweets
HAPPY BIRTHDAY APASRA
• 84626 Tweets
#23rdyokoyosterday
• 74236 Tweets
Newsmax
• 66978 Tweets
Melania
• 50592 Tweets
Tel Aviv
• 49590 Tweets
#RNC2024
• 48273 Tweets
Cardi
• 41447 Tweets
Independiente
• 37621 Tweets
أبيب
• 37228 Tweets
Ridery
• 35964 Tweets
Magic Year with TayNew
• 35071 Tweets
ステラミラクル
• 28375 Tweets
Eric Trump
• 19775 Tweets
DominATE STRAY KIDS COMEBACK
• 19614 Tweets
Dana White
• 17665 Tweets
内P復活
• 14414 Tweets
バイデン撤退
• 11603 Tweets
#スターレイル予告番組
• 11282 Tweets
システム障害
• 10971 Tweets
#RNCConvention
• 10758 Tweets
Pinned Tweet
🚀 Introducing our Subnet 32 on
#Bittensor
Network for AI detection: "It's AI"
Our subnet incentivizes the development of distributed solutions aimed at identifying LLM-generated content.
Here's a thread on what you should know about us 🧵👇
4
14
52
🎉 Big announcement! 🎉
We've launched our MVP version of the website. You can try it out here:
Our website's backend operates entirely on our 32 SN in Bittensor Network, enabling you to test whether a piece of text was generated by a LLM.
7
6
35
API service & Website Release
Finally, we're ready to release our first application for users outside of bittensor, which is going to replace our previous MVP website and provide much more opportunities for end-users.
While on the landing page you can find information about
1
9
26
Subnet 32 – It’s AI - Community Call
Time: Wed 13 March 2024 at 8:00pm GMT+7
Duration: 1 hour
Channel: X:
2
1
12
Just shared our latest updates:
1. Got our first emissions on the subnet - we're secure from de-registration. 🛡️
2. All 256 subnet slots filled, signaling growing competition among miners. 🏁
3. We're working on miner improvements and validation mechanism upgrades to
2
2
13
New Release: API Code for Validators
We’re excited to share that we’ve made the code for querying miners available, so that validators with at least 1K stake can use it to get predictions right from the miners and explore the value they create on their own!
Earlier this week,
1
1
9
If you are interested with our idea and vision, please do support us via follow, like and repost this tweet.
2
2
8
Leaderboard Released
We've just released a leadebord, where you can check all miner's scores. We're showing there f1, fp and ap scores for the last validation iteration, so it can variate on several percents depending on the validation sampels. We'll also soon release
1
3
8
📷 Miners and Validators update - v2.2.0
Miners
We are happy to announce the third generation of miners, where we achieved a significant improvement to compare with the second generation.
To provide a high quality of solving LLM-detection problem for our subnet we needed to
0
1
8
We strive to improve overall understanding of our subnet and enhance transparency with our community.
To achieve this, we have released a document that details the objectives of our project and the ultimate goals of our subnet:
0
1
8
We welcome all types of feedback. So if you have any suggestions or ideas, feel free to join our sub-channel on Bittensor:
Or reach out to
@lagunast
.eth on Discord.
2
1
4
To address this problem, we're developing a subnet that defines incentive mechanisms, validation processes, and a baseline model for miners.
For more details, check out our GitHub:
1
0
3
We're hosting a community call in order to get to know the community and would love to hear your feedback on our Subnet 32 – It’s AI.
Our lead engineers will be on the call too. It would much be appreciated if you can join and shoot us feedback and questions.
See you all soon!
1
1
2
Also, we're creating a MVP website so everyone can assess the quality of the miners in our subnet at any time and provide feedback.
At the moment, it only supports English and the quality varies depending on the LLM. But this is just the beginning!
1
0
2
We're addressing a real and growing problem: ChatGPT generates only 100B words daily! Compared to the 100T words generated by humans, it's a significant number.
We believe that the ability to accurately identify AI-generated text will become increasingly important.
1
0
2
Additionally, we aim to commercialize our subnet. This includes:
- API (for integrations)
- Webservice (for regular users)
In doing so we will become less reliant on emissions and start gaining real usage. For now, we are figuring out how to implement this in best way possible.
1
0
2