Damn it, turns out I posted it to technical support, which is prob not appropriate. I don't see how I can delete or move it. Moderator: feel free to delete this, and I'll put it to a better fitting place.

Hi everyone,

I want to collect a list of software libraries that people use with Bitcoin. Anything useful, really. Tools to integrate, tools to read, tools to analyze, tools to build web apps, anything. I need these for the work I do in sourcerer.io - we analyze source code to understand what programmers do, and libraries are a big part of it. An example sourcerer profile: https://sourcerer.io/adnanrahic

So if you have a moment, and use Bitcoin in your programs, could you respond here with a suggestion?

Thank you!

Yes, efficiency is a big plus here. What makes me excited most is that it distributes freshly minted crypto tokens into the hands of open source developers. Bitcoin rewards people who bought computers to mine bitcoin, so it's very self serving. This coin will ride on top of existing open source effort and will reward engineers who build new features and fix bugs in open source.

I continue brainstorming around whacky crypto ideas. Here is another one. Original post is here: https://blog.sourcerer.io/no-proof-of-work-required-cryptocurrency-for-open-source-8aa0b01371e

I pasted the text below as well. Any thoughts are welcome. Especially, if you see something that does not make sense.

****

No Proof of Work Required (Cryptocurrency for Open Source)

Last time around (https://medium.com/@sergey_surkov/random-a-marriage-between-crypto-and-ai-1a3f4aa752ca), I described a blockchain with a proof of useful work: a blockchain that uses processing power to train neural networks instead of hashing nearly the same data over and over. One of my readers said something interesting to me: “So it’s a proof of work by artificial intelligence? Is there a counterpart for humans?” And this got me thinking. Is there already a stream of verifiable work made by people that cryptocurrency can ride on top of?

Of course, there is. The work that the open source community does every day is exactly that. Every day, commits flow into myriads of open source projects all over the world representing human work that everybody can see and that is very difficult to fake, especially in more popular projects that have to follow strict code review procedures, require extensive unit testing, who are very careful about stability of their projects, and are interested in preserving their reputation. What would a crypto currency on top of open source activity could look like?

Let’s start with a commit in a respectable open source project serving as a proof work. The author of such commit then will have the right to add a block to a cryptographic ledger. Commits are usually not cryptographically signed although it’s possible for example for Git (https://git-scm.com/book/id/v2/Git-Tools-Signing-Your-Work). But normally it’s just email that gets saved in a repository. This means that we will need some sort of verification process but I will leave this out of the scope of this post.

Once the author forms the block of new transactions, and lists their new open source commit ID there, they will also add a transaction that rewards them for participation in the system - and for their work on open source. This is a counterpart of a coinbase transaction. To promote open source effort, we could try to issue higher reward for more complicated commits, for instance; commits that are code edits versus README edits, commits that change older code, orr commits that change files that are imported from many places, and so on.

A single open source commit may appear a flimsy foundation, but it’s easy to strengthen it by requiring multiple authors from unrelated projects to sign the new block as well. They will verify all previous commits, and pending transactions, and then will add their coinbase transactions into their block, and sign the whole thing. Block acceptance by nodes will work similarly: all network nodes will need to clone open source repositories from well known locations, and verify that transactions, signatures, and open source commits all valid. A block then is considered valid if a majority of its commits are valid (to account for rare cases when an open source project by some reason somehow destroys the signing commit).

What this does is a very interesting thing. It puts newly minted tokens into the hands of open source developers. If there is a demand side to this somehow, for example, freelancers or outsourcers asking to be paid in this open source coin, it could provide a universal way of monetizing all useful open source effort. How cool is that?

Not doing an ICO for this one either. But it could be a useful feature for https://sourcerer.io/.

There is a famous book by Benjamin Graham called "Intelligent Investor". It describes an investment philosophy based on underlying value of security. It brings up two basic points:

1. If you don't have a lot of time to spend on investment, just buy a mix of stock and bond indices and hold them forever.
2. If you do have a lot of time, look for undervalued assets. He also proposes some criteria for deciding if a stock is undervalued, e.g. a company's market cap is lower than the value of assets it owns, it's undervalued.

Does value thinking exist in cryptocurrency? Speculation aside, crypto should be driven by the same basic forces as fiat. People would want a crypto token if they want something that can be purchased with the token. Do crypto investors look for underlying value of a token? And if they do, do they look for undervalued assets? Is it even possible in the crypto world?

I certainly have some technical skills. I used to work in Google search ranking, I built some augment reality technology and AI for ecommerce, and I currently work on https://sourcerer.io where we want computers to understand what program code is about. I don't have crypto background though, at least not yet And yes, building ncoin would be a lot of work.

This ncoin idea started as a conversation inside sourcerer, because everyone is fascinated with deep learning, so somebody asked a question "can we merge crypto and ai"? Didn't make sense at first, but then it somehow did.

N is for neural, that's right. It was the first thing that came to my mind

I haven't heard of Gridcoin before; thank you. It looks like they use a proof of stake to run a blockchain and organize volunteers who donate their computing resources. This is admirable, and very useful. What was the name of the other platform?

(I need a name for my idea as well; let me call it NCoin).

NCoin is more commercial, and focused on neural networks.

It takes a lot of computing power to train a neural network; it takes a lot computing power to mine a block. NCoin is a way for a blockchain to have proof of work by training a neural network.

Here is why it makes sense:

1. Demand for deep learning goes up, so should the demand for computing resources for training deep learning networks. From this perspective, NCoin is more interesting than say Filecoin, because storage cost is negligible compared to computing cost in AI. This means there will be a lot of people willing to pay for training neural networks, and this allows to create liquidity for NCoin.

2. All those rigs people built for mining Bitcoin and other proof-of-work based currencies are perfectly usable for NCoin. So when Ethereum transitions to proof-of-stake, or Bitcoin reaches its limit, all these computing resources will become available.

Yup, I was going from the deep learning angle because it's something I know about. Neural networks are already built for distributed training, and improvement in network as training progresses is very difficult to fake.

Idea on how to train deep learning neural networks in blockchain and use that as a proof of work.

https://medium.com/@sergey_surkov/random-a-marriage-between-crypto-and-ai-1a3f4aa752ca

Brainstorming. If there is a better forum for this, please me know.

The gist is using gradient descent in deep learning as a proof of useful work. Somewhat technical read.