DataDAO Spotlight: Human-SuperText DLP - Datasets that bring human intelligence to AI

The future of decentralised intelligence will thrive, not from a clash between humans and machines but from a powerful symbiosis that accelerates AI performance. 

SocialTensor, creators of Subnet 23 (SN23) on the Bittensor network, have launched VanaTensor on the Vana Network.  Vana Tensor has built the Human-SuperText Data Liquidity Pool (DLP) to improve the “ground truth” accuracy of AI real-time human data, feeding this data directly into their Bittensor miner and validator nodes.

By combining human ratings with AI-generated content, the Human-SuperText DLP builds powerful datasets that reflect real-world human interpretation. This is particularly important as humans adapt and become more attuned to AI-generated content, resulting in an ever-increasing bar for authenticity.  

The process begins with synthetic text generation from SN23 models, followed by human evaluations. Users rate and rank AI-generated text through an intuitive web interface, creating a structured feedback loop that captures human “truth” data and guides reinforcement learning for AI models. 

Through this recursive improvement cycle, the Human-SuperText DLP generates consistently high-volume and high-quality datasets on Vana. Developers seeking premium data can use these datasets to fine-tune large language models, including those used by Bittensor miners and validators in SN23. 

The result is a future-ready bridge between human intelligence and machine intelligence with co-created datasets built on Vana. 

Vana and SocialTensor demonstrate that the future of AI is not an “either-or” approach but a complementary relationship, where robust AI stems from both human insight and machine computation. 

Join the Movement

The Human-SuperText pre-mine launches next week, marking your chance to join one of the most exciting projects in Vana’s ecosystem. 

Follow Human-SuperText DLP on Discord and Twitter to be the first to know when it drops.

‍