My reply to @spbeckman regarding his article titled "Brief idea about Steemit, monkeys on a keyboard, post005"

transhuman (60)in #blockchain • 6 years ago (edited)

In his article, @spbeckman stated:

How do we find content in the noise? Machine learning and data informatics. I don't know exactly how this is going to work yet, I think that we need algorithms that have a better understanding of language, but ultimately, we'll need some algorithm to extract the creative and original ideas from the noise.

You could invent "NeuSteem" where your ML algorithms sift out the creative and original ideas from the old Steemit to create a highly refined product.

Moreover, the underpinnings of the Steemit platform are catching on elsewhere; AKASHA, Yours, Synereo, MINDS and LBRY, to name a few. After 2 years on the platform, I'm getting the idea that Steemit isn't the penultimate social media on a blockchain. It may even be fading in popularity according to this recent Alexa analysis of Steemit traffic.

Nonetheless, there are some ideas that are Steemit concepts worth keeping. I'm particularly interested in how these ideas apply to science. For example, all bona fide science experimental data is worth saving on a blockchain, including those "failed experiments" not considered by major publishers. This saves duplication of work that only leads to blind allies. A blockchain consensus can insure validity of said experiments and save the data for all time instead of losing it in a lab drawer.

Furthermore, the cryptocurrency generated by blockchains can be used to advance research in a number of ways.

Bounties could be one method. Cryptocurrency rewards could be awarded to scientists who solve a particular problem or perhaps simply duplicate results.

Here is another way mining for crypto could advance science:

Blockchain technologies are based on crypto mining, which basically means that they combine the computational resources of millions of participants all over the world. Skychain will add social utility to crypto mining and make it more beneficial for all the parties by enabling crypto miners to provide a valuable service, using their computational resourcesfor neural network training and calculation. We have developed the proof-of-deep-learning-training and proof-of-inference algorithms that are at the heart of Skychain. [1]

This is a significant improvement over the Proof-of-Work algorithms which just grinds out the solution to an abstruse, useless asymetric cryptographic problem. Why not be using all that computer power to do useful work, like Folding at Home which solves protein folding problems for science. I believe they now pay crypto for use of that computer time.

I posted a blog about SkyChain here: https://steemit.com/crypto/@transhuman/skychain-saving-lives-with-artificial-intelligence-ai-on-the-blockchain.

There is much more, but this probably enough for now.

Looking forward to your feedback...