[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/biz/ - Business & Finance

Search:


View post   

>> No.6510212 [View]
File: 154 KB, 693x919, dbcshit.jpg [View same] [iqdb] [saucenao] [google]
6510212

>>6510009

> just a few short paragraphs of "AI is hard" on page 3
> only thing about actually using deepbrain chain fro the AI is a small blurb in section 2.3

>DeepBrain Chain realizes the decentralized supply of core computing capability of artificial >intelligence, but artificial intelligence, in addition to computing power, also needs algorithms and >data. Data are trained to be the model while the algorithm, with the model, generates artificial >intelligence applications. Therefore, DeepBrain Chain in the future will derive AI data trading >platform, AI algorithm trading platform, AI model trading platform, AI container trading platform, and >AI application trading platform. AI application trading platform has a certain scale: there are >hundreds of AI skills applications on DeepBrain, which can be sold to users. In addition, the >DeepBrain Chain ecosystem will help the artificial intelligence manufacturers to issue their own >virtual currency lobally, and the virtual currency of the manufacturer can be freely exchanged with >DBC


>>6510089
If you think this type of technology is useful for AI then you're an actual idiot. To properly train a neural network takes hundreds of thousands or millions or more iterations upon a sample set of hundreds of thousands or millions of samples. That isnt getting shoveled through any blockchain and CANT be split up because each iteration of training has to perform forward and back propogation on the previous samples results. You can't just divvy this up like you would a graphics card.

Navigation
View posts[+24][+48][+96]