Sia - Siafund Redemption Deadline: June 1st, 2015

[Lordoftherigs]

well that's a bummer but nobody was expecting this will be an easy road. You already have some viable alternatives.

If you find a path with no obstacles, it probably doesn't lead anywhere.


by some smart guy

[1715126542]

1715126542

[1715126542]

1715126542

[Taek]

I think I've had a change in design philosophy. The original Sia was this gigantic system that was many levels of crypto and had many moving parts. It was too much to do all at once, and now we've paid the price for it. I think I would rather release a series of less ambitious decentralized cryptocurrencies that build on each other, and call the entire set of cryptocurrencies "Sia".

One of the original goals of Sia was to have a global price on storage, set by supply and demand. More available storage => price drop. More files being uploaded => price gain. This creates a place for people to manipulate the price though. That means the system needs a method for protecting storage against price manipulation. And the only way I know how to do that causes turbulence on the network, which means a greater amount of bandwidth used per file upload, which means a greater overall cost to storage. It also adds complexity.

I'm thinking about creating a first coin that's just a list of hosts and files, and some sort of method for you to confirm that they're storing your stuff, but allows hosts to get paid even if you aren't around to challenge them. Clients would be in charge of figuring out where to store their files and how to manage redundancy, and they'd need to negotiate on a per-host basis to set the prices of things. You end up with really free form decentralized storage, but a potentially powerful platform if the clients are smart enough. And we can put it on a single, heavyweight blockchain.

Then we can release another cryptocurrency that does something similar, but has a global price and methods for protecting that price from manipulation. And the two can coexist and we can see which one does better, all while building out a parallel-blockchain solution. And then some time in the distant future we can maybe release something with tree chains. If we get enough funding we can pursue centralized alternatives as well. (Centralized funding would be through typical VC stuff, for legal reasons it'd be a bad idea to try and crowd fund that).

I guess I'm feeling optimistic. Hopefully by Christmas we can have some sort of blockchain and a client with decent mechanisms for microchannel payments (used when downloading), atomic cross-chain trading (converting bitcoins to and from siacoins without a third party), and some sort of filesystem integration. Pretty much all of that except the filesystem already exists (at least theoretically) within Bitcoin.

Thanks for all the support, you guys are great.

[Tobo]

James is working on some storage stuff. I wonder if it is something interesting or not- https://nxtforum.org/unity/mofnfs-fully-encrypted-decentralized-m-of-n-file-storage/msg113973/#msg113973

[Taek]

I can't tell if it's interesting based on the post. I don't think a dht would be useful for the types of things that Sia is doing. I don't understand what the intention of the rest is, but things like erasure coding (M of N) and encryption (all the ciphers he listed) are relevant to us. I don't know what the rest of the architecture does though, I haven't been keeping up with James' designs.

Please make sure to inform us if/when these new ideas/paths change anything in respect to Sianote.

We have a responsibility to our funders to make the sianote valuable. This has to be done on decentralized storage. All decentralized storage systems that Nebulous Inc. releases in the future will somehow incorporate the sianote, filling the responsibilities stated on our website. If we end up releasing two or five or fifty-thousand decentralized file storage coins, sianote holders can profit from all of them.

Sianotes will not extend to any centralized solutions, should we develop any. Right now we're still prioritizing decentralized storage.

[Gaman]

The other, _other_ finally is that we could build a centralized system that employs all of the components we've talked about. Proof of upload would be really easy in the centralized case (you just send it to our servers, and then you've proven the upload!), and we'd just apply proof of storage, proof of capacity, and some form of transparent quorum-level consensus that gets validated through a centralized node. This is obviously not a decentralized file system, but it would operate in the same way that Sia originally was going to operate.

We'll be considering all four throughout the week and let you know what happens.

Sianotes will not extend to any centralized solutions, should we develop any. Right now we're still prioritizing decentralized storage.

Are you saying if you decide to build a centralized system like the one above, sianotes are worthless? and you are free from obligation to sianotes holders?

[Taek]

I'm saying Sianotes will not apply to anything centralized that we build. There are legal problems with doing that. Distributing money via a decentralized cryptocurrency isolates us from a lot of regulation. We aren't giving you the money, a protocol is. We have no control over the protocol (once it's been created) and that makes it a regulatory grey area.

But if we do control the streams, we're giving you direct access to money that's in our control. Suddenly there are regulatory problems. Are you workers? Stock holders? Why do you get access to this unique stream of revenue for the company? It's very different, and it will cost us more in legal hassle than we raised in the crowd fund

That said, I don't think building a centralized system makes sense right now. Where is the miner incentive? If we did something centralized, how would we be that much different from the competition? The original advantage to Sia was that you'd start out with a bunch of crappy storage rigs, but miners would get increasingly aggressive and innovative with building cheaper and better micro datacenters. If we centralized things, nobody would put up with the initial cheap and crappy service. We'd have not way to grow into the fully fledged and highly efficient mining system that 4 years of Bitcoin innovations has brought.

A lot of people don't realize it, but you are never going to make money from consumer grade hardware, at least not in the long run. It will be exactly like Bitcoin mining. At first mining on your spare hard drive space might make sense, but only if you have hundreds of gigabytes, and as soon as the system is popular people are going to be buying dedicated storage that's cheap and doesn't use any power. And then mining on a laptop for storage will be the same thing as mining on a laptop for Bitcoin: waste of effort. But it takes people who are willing to innovate to get that far. And I don't think people would be willing to innovate for a centralized service, especially when it's initially a lot more expensive than S3 or DropBox.

It's also worth talking about what our obligations are to the people who crowd funded us. The first thing to note is that when you crowd fund a project, it is possible for that project to fail. You take a risk on the project, and it's not a guaranteed return. That said, we still have a lot of the money that you guys gave us during the crowd sale. It would be many levels of dishonest to stop working on the things that you guys are paying us to do. Once your money runs out, we have other investors whom we have different obligations to, but we're not at that point yet. Even more, it seems like both your interests and our other investors are aligned at the moment; we want to see lots of growth in the cryptocurrency world and centralization is not the way to do that.

We'll see where the future goes. Right now we're still focused on releasing a decentralized platform for storage, and I think that legally we're locked into that because of the crowd sale. I don't know what will happen if the crowd sale money runs out. We've got around 90 Bitcoins left (additionally ~200k nxt), which may not sound like a lot compared to what we raised, (just over 200), but a lot of our expenses were one-time. (moving out of college was expensive, we've now settled in Boston and our monthly cost is a lot lower). Additionally, Bitcoin seems to be going up again (fingers crossed). If it goes up to the $1k or $2k mark in the next 6 months, those 90 btc are going to stretch a lot farther. Hopefully it will be enough for us to get a decentralized storage solution out the door, especially if we simplify things a lot.

I don't think it's worth a ton of conversation right now because we've got at least 6 months left on our current funds, more if the market starts doing better (I'm optimistic).

[Breasal]

Sorry, I must speak up here. I see many words but noting near what was promised when I funding Sia. You plan to wait out another six months in Boston (one of the most expensive cities in the US) and still have no product!

How about putting up a buy wall on those NXT sianotes for 2500? I think that's the right thing to do, don't you?

[grandpa_seth]

Sorry, I must speak up here. I see many words but noting near what was promised when I funding Sia. You plan to wait out another six months in Boston (one of the most expensive cities in the US) and still have no product!

How about putting up a buy wall on those NXT sianotes for 2500? I think that's the right thing to do, don't you?

I understand what Breasal is saying and how he feels. I would say putting a buy wall is appropriate if you have given up on sia. It does seem like despite your somewhat optimism on completing a decentralized sia, you do seem to be giving up on sia being potentially a huge cloud storage solution for the masses and guys like netflix. If you feel currently that sia no longer has the potential to make sianote holders a large return and you are just going through the motions then I would vote for buying them back with your remaining funds.

Clearly David, you are very intelligent and honest. So I would believe you if you tell us that Sia still has the potential to be a great investment for sianote holders. But if you are just fulfilling your crowdfunding obligation to release a storage coin that you don't truly believe in then I would rather have my sianotes bought back at a loss.

[dadingsda]

+1 grandpa_seth
Couldn´t say it better. Sia was one of my top 5.  But better stop it now than waste our money - PLEASE!

[Tobo]

+1 grandpa_seth
Couldn´t say it better. Sia was one of my top 5.  But better stop it now than waste our money - PLEASE!

I have to agree with this. It is better to stop doing it if the devs don't whole heartily believe in it.

[grandpa_seth]

Don't get me wrong. What I'm hoping will happen is Taek replying with "Hey guys, there is a misunderstanding. This is a big bump in the road and the road will take longer than expected, but sianotes will still be extremely valuable one day because I'm confident I will find the solution. I just need some time to sort it out."

[Breasal]

Don't get me wrong. What I'm hoping will happen is Taek replying with "Hey guys, there is a misunderstanding. This is a big bump in the road and the road will take longer than expected, but sianotes will still be extremely valuable one day because I'm confident I will find the solution. I just need some time to sort it out."

Yea, maybe this is the case but none the less sianote holders should see an effort made to place a buy wall imo. A buy wall speaks louder than what you have just written (with all due respect). If the sia team is so confident in producing a product, then they will have no problem placing this. Conversely, if they are "giving up" then they should put up a wall with remaining funds.

[Taek]

Boston was a logical choice given our proximity to other startups, our angel investor, and MIT + Harvard. The other choice would have been Silicon Valley, which is even more expensive. Also, the rent is not as bad as the pundits would have you believe, and beyond that cost of living is pretty tame. The public transportation is good and cheap, the food is decently cheap (grocery store food, restaurants are expensive but we don't eat out much), and that's about all a budding startup needs.

Don't get me wrong. What I'm hoping will happen is Taek replying with "Hey guys, there is a misunderstanding. This is a big bump in the road and the road will take longer than expected, but sianotes will still be extremely valuable one day because I'm confident I will find the solution. I just need some time to sort it out."

It would definitely be foolish to give up on the same week that you break your core technology. We broke our core technology 4 days ago and wanted to be upfront about it. The Sia that you've read about won't work. But that doesn't mean that there aren't other options available to us. It's also entirely possible that we'll find some alternative way to build Sia that gives it all of the same powers without being broken. It's clear that a large redesign is needed, and that any new product must have minimal complexity to have a decent chance, but there's no reason to give up hope yet.

Especially because I believe that decentralized storage is something that the world really needs. With the original Sia I was hoping to replace all cloud storage, but providing a reasonably cheap source of decentralized storage is still a huge value-add to the Internet. Even a simple blockchain scheme would potentially be enough to replace much of Bittorrent, which should provide an income that far exceeds the original investment in the Sianote. Decentralized storage is not dead, even if it seems as though it won't be replacing centralized storage outright.

How about putting up a buy wall on those NXT sianotes for 2500? I think that's the right thing to do, don't you?

Putting up a buy wall is something we could consider, but you need to realize that there are two sides to the equation. Not everyone will want that to happen, I'm sure that there are a decent percentage of sianote holders who would rather see this all the way through. I think a buy wall would be responsible if Luke and I as a team had given up on making the Sianotes viable. But we haven't, we still believe that we can make something valuable, even if our sights have been lowered a few notches. We certainly aren't just going through the motions. We want this to succeed more than any of you, so much that we've given up our previous career paths and left many of our friends.

but sianotes will still be extremely valuable one day because I'm confident I will find the solution."

Going to shoot as straight as possible with this one. Two weeks ago if you had asked me what I though was a reasonable end-game price of the Sianote, I would have said somewhere between $20,000 and $100,000 each, with potentials to hit $500k or $1m in the same way that Bitcoin has the potential to break $1m (eg very unlikely, but not impossible). Today, given our current set of potential solutions, I'm much more inclined to say that the end game value is in the $1000 to $5000 range. That assumes that Sia is the primary solution for decentralized storage, but is unable to compete in flexibility with centralized storage.

At the moment I'm not convinced that it's possible for decentralized storage to compete with centralized storage. But at the same time I'm not convinced that it's not possible either. I'm convinced that it's a very difficult problem and my previous approach to the problem is not a secure approach, and that a new approach is needed.


==================


So what will things look like going forward:

1. No buy wall, not until Luke and I are both convinced that we can't make decentralized storage happen. And we're very far from being convinced of that.
2. I've narrowed down the previous four options I listed to the following two:
    a. We pursue a simple blockchain model for decentralized storage. It will be a lot different, and it will be unfriendly to small files in the same way that Bitcoin is unfriendly to small transactions. But it will be useful for large files, and it will potentially be competitive with centralized storage. For large files. I believe we can get Sia to a point that's still "enterprise friendly", which was among the original goals.
    b. We spend some more time exploring consensus models that would enable small files with rapid updates to also be useful.

And actually, being enterprise friendly may be all we need to pump the value of the Sianote up to $10k+. But it's a new system and I need a lot of time to think about the design and implications of such a thing.

I appreciate all your feedback and opinions. Me knowing what you think helps us make decisions moving forward.

[apenzl]

[Tobo]

Sounds good. Thanks!

[grandpa_seth]

I believed Sia was a good investment because of the potential "enterprise friendly" goals. So I am for that. $1,000-$5,000 as the potential best case scenario upside isn't really a lot considering the risk and how much we paid. But hopefully, Sia will be bigger than even that estimate. Well, my unofficial vote is to continue the project. I bought mostly at post ipo prices which is a lot of $ but I believe as long as you guys still live and breathe Sia its a good investment. Thanks guys.

[gs02xzz]

Interesting article - http://blogs.wsj.com/accelerators/2014/10/10/weekend-read-the-imminent-decentralized-computing-revolution/

[Taek]

Got enough ideas down to put together a new whitepaper. We've opted for a dumb protocol that has smart clients. Our v1 is going to look a lot like Bitcoin, except that it's also going to have a list of hosts and a list of files. You upload files to one host at a time, and the host must agree to store the file, you can't just upload arbitrarily. This allows the hosts to protect themselves against illegal files.

Clients will contact a bunch of hosts at a a time, and store things redundantly. The blockchain will force hosts to perform proofs of storage. Clients and hosts will have a bunch of tools that allow them to negotiate price, but I don't think that the price will be able to change throughout the lifetime of the file. Hosts can agree to pay penalties if they lose the file or fail to perform storage proofs.

That's about all of the details I have right now but I'm optimistic. Our long term plan is as follows:

1. Create new whitepaper
2. Have experienced people look at it, such as the core devs
3. Launch an alpha and test things.
4. Launch a _beta_ product that is fully functioning, and has sianotes. We hope this is done in Q2 2015.
5. Grow and support the _beta_ product.

6. Create new whitepaper, incorporating more dramatic changes. There are a lot of things that I think could be improved, but they are all ambitious, and I think it would be foolish to try and complete any of them in the next year. This whitepaper will probably not be done until 2016.
7. Have it reviewed, hype it.
8. Create alpha, test
9. Create 1.0 for Sia. This step probably won't be complete until mid 2017.

====

So the idea is to do something easy and very similar to Bitcoin for our first product, but being transparent about the idea that it's not the final iteration. We want to get something out because we believe it will take a long time to have something more sophisticated that is also properly secure. I think we will be able to return to quorum-style scalability, which will once again enable many many small files.

Having a file storage beta that's making us money will give us experience, we'll know the types of features that are missing that are most useful after we have a cryptocurrency. We'll know where the most severe bugs are.

There is a possibility that we simply upgrade the beta to the 1.0 after a few years. This is hugely dependent on the miners being in agreement though. Altcoins like Nxt have had some success doing regular hard forks. If it looks like we can safely hardfork our beta product, we'll probably prefer that approach.

I'm very much into the idea of smaller iterations where we try new things each iteration. A cryptocurrency that does that is essentially centralized around the developers however. We'll start with this first whitepaper, have a big launch and start supporting actual decentralized storage apps. And then we'll figure out where to go from there based on what people like and don't like, and what they want or don't want.

Thanks for sticking with us. Working on a new whitepaper, no eta.

[grandpa_seth]

I'm kind of liking this. It shows you are committed to this for the long term which I think was the concern for some recently. As long as you guys are still trying to make Sia a huge revenue making machine for Nebulous then Sianote holders will mutually benefit obviously.

[Taek]

I had a sort of zen minimalist moment yesterday. We have a new specification for Sia, and it's really really simple. Even simpler than Bitcoin, because we simplified the way the scripting system works.

It's pretty much the most minimal possible protocol that allows decentralized storage to happen. There are hosts that announce they have storage, and there are files that hosts agree to store, and the hosts submit proof-of-storage on the files to get paid. If they don't submit proof-of-storage, then the uploader gets refunded automatically.

And that's pretty much it, happening on a blockchain. It's so simple that we should have an alpha ready to go within a few weeks, maybe 4 to 6. I'm actually really excited about this new protocol, because it's simple enough to implement quickly and powerful enough to allow a full structure of decentralized storage.

We do sacrifice a few things:
1. There is no global network price. Hosts set their own price.
2. There is no automatic file repair. Clients will need to manage file repairs themselves. This means they need to be online to make the repairs, but that means running 'repair' every few months instead of needing to be online all the time.
3. There is no 'sublinear consensus'. Sia is a single blockchain and will need to deal with blockchain bloat in all of the same ways that apply to other altcoins. This shouldn't be a problem until we have multiple millions of users.

That's basically it. We slimmed things down to the very core requirements: decentralized storage that doesn't require you to be online all the time. Turns out you can do this with only a few simple changes to the Bitcoin protocol. We should still be completely compatible with enterprise customers, which I think is where the majority of our users will eventually come from. Without sublinear consensus, I'm having troubles imagining that we'll be useful for files below 10GB in size, but that's a problem for the future. Enterprise customers are unlikely to need files less than 1TB in size (replace file with 'cloud disk' --> 'file' is a misleading term). A 2.0 down the line will probably address this (requiring either a hardfork or a different cryptocurrency). We want to delay sublinear consensus because right now there are no good solutions to decentralized storage. We can change this without decentralized storage and we think it's more valuable to release something simple now than something complex and potentially insecure after we spend 3-6 more months planning it out.

We're planning on releasing a whitepaper alongside a beta implementation. The beta implementation will be fully usable, but we're going to premine 99% of the coins on the beta, as sort of a "NOT FINISHED YET" red flag. After a few months of beta, we'll decide how much needs to change for a 1.0, and we'll release a 1.0 that has sianotes and no premining.

Luke and I are ecstatic. Wanted to share my enthusiasm. I know it's been a roller coaster, but that seems to be the expectation for startups and early stage investing.