At 8:30, finish browsing the homepage—this one is a bit interesting: a search giant bought the data of a bankrupt airline from the auction house, claiming it was for training AI.
“AI supply bottleneck” has long stopped being about compute power; it’s about data. The auction prices for scarce training corpora have been rising steadily over the past two years. Big model companies are competing on who can get data that others can’t.
On the crypto side, this is when an old narrative is getting popular again. Move training data on-chain to enable decentralized distribution. On the $TAO side, they’ve been pushing this route—pricing, rights confirmation, and revenue sharing for the corpora are all embedded into the protocol. Whether it can actually run will be judged by institutional orders.
The $BTC limited benefits are there, but the on-chainization of the data element has come with an extra layer of credibility. The real question is whether orders can keep up this year.
#数据要素上链 #训练数据拍卖溢价 #technology
“AI supply bottleneck” has long stopped being about compute power; it’s about data. The auction prices for scarce training corpora have been rising steadily over the past two years. Big model companies are competing on who can get data that others can’t.
On the crypto side, this is when an old narrative is getting popular again. Move training data on-chain to enable decentralized distribution. On the $TAO side, they’ve been pushing this route—pricing, rights confirmation, and revenue sharing for the corpora are all embedded into the protocol. Whether it can actually run will be judged by institutional orders.
The $BTC limited benefits are there, but the on-chainization of the data element has come with an extra layer of credibility. The real question is whether orders can keep up this year.
#数据要素上链 #训练数据拍卖溢价 #technology