Before talking about OMG or the OMG coin price, let’s first talk about the blockchain trilemma, one of the central problems of our industry. The blockchain trilemma is a name coined by Vitalik Buterin that proposes a set of three main issues that developers encounter when building blockchains. More often than not, developers are forced to choose two out of three: Decentralization. Rather than being managed by a single entity, blockchains distribute control over the network equally to all participants. Security. Blockchain networks should have ironclad defenses that prevent malicious entities from taking over. Scalability. Blockchains should support an enormous number of transactions and users without faltering by increasing fees and transaction times. In a single sentence, the OMG network is a non-custodial scaling solution for Layer-2 which is developed over the Ethereum blockchain. What this means is that OMG is designed to help users transfer tokens faster and at signi...
How close are we in creating AI systems like Jarvis?
I find this question quite interesting to answer. I will try my best to give my view of current AI and Jarvis, and try to relate both.
AI or Artificial Intelligence is a machine made smarter like humans. That means if we can’t see what’s on the other side then we can’t differentiate whether on the other side is human or machine, idea is same as of Turing Test.
So to achieve artificial intelligence we work on various fields, i.e. Machine Learning, Computer Vision, Robotics, Image Processing, Natural Language Processing etc.
On the other hand, Jarvis is a fully developed Artificial Intelligence. Marvel comics have shown Jarvis to be what we actually desire to build. Jarvis have intelligence same (or more) than human with insane amount of computing power. What I feel impressive about Jarvis is that it can understand and also give sarcastic comment.
Right now, we are far far behind creating a AI system like Jarvis. Jarvis is fully developed AI whereas on the other hand we are still struggling to create an AI system as smart as a mouse. We are doing good in supervised (it can be seen from examples such as movie recommendations, personalized search results, predicting malignant cell etc) but we are nowhere near acceptable in unsupervised learning. We can outperform human in supervised learning (AlphaGO) using resources like GPU etc. We yet to have good results for unsupervised learning.
Think of unsupervised leaning as a common sense. If we train a machine to do a certain task and then we leave the machine to a new environment with different task it will perform horribly in that new task. This is what AI lack, ability to perform in new environment but humans can do this pretty well.
Tremendous amount of research is going on all over the world to tackle that problem. You can see research being published by scientists all over the world. Huge amount of money is also being invested by business and government for the same. Today’s most of the business and scientists understand that this is the next big thing and this could be worth billions and trillions of dollars.
So to summarize, we are doing wonderfully in supervised learning and trying to make it even stronger, but unsupervised leaning and other branches are far behind. On a positive note, 20–30 years back we didn’t even had machine learning but here we are now, solving some of the major problems in the world through machines. So, the time is not far away when we will be ruled by machines. (fingers crossed)
Comments
Post a Comment