Subscribe to Bankless or sign in
AI researchers have hit a "data wall," where publicly available internet data is no longer enough for large-scale model training.
In response, tech giants are increasingly buying private data, netting platforms like Reddit hundreds of millions annually from selling user-generated content as training material. Photobucket, Tumblr, and Stack Overflow profit by licensing user data to AI developers, with the individuals whose content drives these advancements rarely receiving compensation. Shutterstock has inked deals valued between $25M and $50M to license its stock media libraries to AI companies, while Meta even considered acquiring Simon & Schuster for access to its e-book catalog.
This growing economic divide reflects a broader trend in which access to data is increasingly controlled by a few wealthy tech companies. This highlights a deeper issue: User data holds immense value, yet most see no return for what they create.
Abonnez-vous gratuitement pour continuer à lire
- Soutenez le mouvement Bankless
- Accès à des milliers d’articles
- Archive complète des épisodes Bankless
- Lancez-vous dans des quêtes gratuites sur Airdrop Hunter
- De l’alpha quotidien dans votre boîte mail
Déjà abonné ? Se connecter
