1. Shrink BTC miner to fit into a pipe
2. Fit pipeMiner as a replacement heating element for electric water heaters
3. ???
4. Profit

The first challenge is the size of this miner. In order to reduce the cost, we are making it fit through a standard (US) screw in electric heating element opening. This makes the design requirements pretty extreme compared to a conventional open air miner. To aid in moving the heat from the boards and into the water, we will be employing the use of 3M's immersion cooling fluid (same stuff Bitfury is using in their newest data center). This fluid is the reason why we are even attempting to build this device. It has a few design requirements when building a sealed unit, but it removes all heat engineering from the design of the board and enclosure, allowing ultra high densities, and thus the small form factor we need to fit into existing infrastructure.

The second largest challenge is making sure you can still have enough hot water. Because we are trying to get 100% hardware utilization, the wattage going into the water is much lower than your traditional heating element. Therefore hot water recovery is much longer. This can be remedied by more storage (storage is cheap compared to silicon). Unfortunately we can't make your current water heater larger in terms of gallon capacity, so we aim to increase the heat capacity. Most water heaters operate around 40-50C. In order to get enough storage, we are planning to get the storage temp up to around 70C (mixer valve required to avoid potential scalding). This ~50% increase in heat capacity should be enough for ~24hrs of regular hot water use. The idea, is that you get one tank worth of hot water a day. So if you use 40 gallons of hot water, and have a 40 gallon tank you would probably be just short of your needs (depending if you used it all at once). However, our miner only uses the bottom heating element position, so you can still set the top thermostat to a lower setting to help in recovery for those higher use days. If this product would gain traction, we could build a phone app that could estimate the "state of charge" of the thermal battery (water heater). This could help the user load balance the use of hot water to ensure 100% utilization.

The third challenge is keeping a high enough ROI potential to allow enough incentive for the added effort of installation of our miner. This is not something we can avoid. It is the only way we can make home mining profitable again. Perhaps your water heater is already 8-10 years old. You could replace it before it fails, and install our miner while your doing something you would have to do in the next couple years anyway. Unfortunately there is no magic wand for physical infrastructure. 

ROI Potential

Until we have the cost of the ASIC chips locked down it is very difficult to know the price, being they are going to be a very significant proportion of the overall cost. Of course, the price of the miner can be much higher if the chips are really efficient, as you will have additional revenue to pay for the higher end chip. We are trying to get a break even point no later than 18-28 months out. That way the perceived risk can be lowered. However, as shown in our ROI graph.

The miner does not have the same cost structure as a traditional miner. Because of nearly no electricity cost, it is profitable for a very long time. We estimate the cost of our miner could be between $1200-$2200. For chips with efficiencies around .16 J/GH a $1200 price point would ROI in about 20 months, and have a 50% ROI in 40 months. But with BitFury's 16nm chip (.06 J/GH) it would ROI in about 6 months at a $1200 price point, and about 18 months at a$3000 price point, and a 50% ROI in 35 months. This really puts into perspective how this device has a pretty wide window of being profitable, given efficient chips.

It is difficult to say how much ROI is needed to make this product viable. Obviously as much as possible would be great, but the cost of chips is out of our control. With economies of scale we can lower our price and maintain performance, but we need some initial traction to allow us to expand (or to even get started for that matter). Please contact us and let us know what you think. @BitHeatOfficial If you have a minute to spare take our short survey, or our slightly longer survey that takes about 3 minutes. (Please take one, not both).

If you haven't already, please read our 4 part blog series about why we are going through all this effort to build this.
One, Two, Three, Four.

In our previous three posts (One, Two, Three) we explored how Bitcoin functions, and how proof of work has a potential route for decentralization. This post will explore what applying this thinking to other forms of computing can yield.

Let’s look at other computing through the lense of Bitcoin. Bitcoin is a Distributed Autonomous Corporation (DAC) paying people to build infrastructure in order to do computations for it. So let's swap out a DAC with an actual company, and replace the SHA256D Hashes with GPU cycles. Now you have general computation incentivized to build a distributed data center architecture. Rather than participating via a miner, individuals would be running a new device that is a micro data center and thermal battery combined. The main difference between this new system and Bitcoin is simply that the bits of data being computed are not direct value, but indirect value: currency once removed.

Using the same thought process discussed in our previous post, we can put this new general computation ‘miner’ in the water heater to reduce the cost of running the hardware, to the point that the hardware (capital) need only pay for itself. Monetizing the computation in this instance is a little more difficult than with Bitcoin, as you have to line up real customers to pay for the computation, rather than having payment for computation built into the network. This problem could potentially be solved by building software to direct the flow of work to idle ‘miners’. In addition, because Bitcoin is such a low friction market, it has a very low profit margin, whereas the profit margin for general computation is relatively very high per unit of energy expended. So if a Bitcoin miner/heater can be profitable, then running general computations would have the potential to be more profitable yet. This is evidenced by simply comparing a supercomputer data center to a Bitcoin ‘data center’. (Of course, we do have to take a moment to be realistic: not everything can be distributed. Anything time sensitive or really high bandwidth would be difficult to do unless you have fiber to the home. However, even this problem will slowly be solved as infrastructure improves — and perhaps devices such as the new ‘miner’ could help spur that improvement. And in the meantime, there are still many types of computation that could be handled immediately.)

There are many things individuals already pay for which data centers have to treat as additional expenses:

• Electricity
• Building
• Employees
• Taxes
• Insurance
• Internet Connection

From the individual’s perspective, these are costs they will face regardless of whether they do or do not participate in this new form of ‘mining’. So if they do decide to participate, the only new cost to be taken on is that of the ‘miner’ itself. All revenue generated could therefore be used to pay off the ‘miner’ first, with everything after that becoming ‘profit’.

But what is ‘profit’ in this context? On the level of the individual, any household device that can be operated remotely can be placed into this new micro data center to reduce running costs and share hardware. This means that what you get out of the ‘miner’ may not be pure monetary gain, but rather reduction of expenses through high capital utilization of hardware. For example, imagine if your gaming console could offset most of its cost by doing computational work using ‘free’ electricity (i.e. electricity already allotted to heat your water) while not in use.

And by having highly integrated hardware, look at some of the integration possibilities:

1. IoT/Smart Home Server
2. Distributed Computing
3. Distributed Proof of Work (Bitcoin, or personal identity)
4. Personal cloud/Distributed Apps
5. Personal High performance gaming/VR rig
6. Machine Learning/Distributed Neural Network/Personal AI
7. Base Station for residential wifi/lifi mesh network

All of a sudden, individuals get a lot of bang for their buck. Using this new device both generates a passive income and acts as a capital investment. This can allow individuals to become micro-capitalists in the digital age. It would now be possible to have a household expense become an investment much like solar panels or battery storage. And just like solar and batteries, the value proposition will only increase in the future as energy becomes more expensive.

This is our ultimate world changing hope for BitHeat: creating a platform for individuals to own their own data, and be part of a new virtual renaissance. The future is going to be a crazy place, and we want to make sure as many people as possible get a piece of it.

In our previous two posts (One and Two) we explored how the currency and ledger we call Bitcoin is a carefully constructed information machine that straddles both the physical and digital worlds. In this post we would like to explain our idea of what the future of mining might look like and how we think that we can distribute the proof of work.

What is the easiest way to spur action from individuals? Pay them. Or at least incentivize the action you want. This is illustrated by how the Bitcoin network adds hashing power to better secure the network. It was also clearly played out in how Paypal got its first customers. In this spirit, we are currently building a brand new type of Bitcoin miner that could once again incentivize home mining. We want people to get paid by the network, and become new participants in the revolution that is Bitcoin.

The Theory

Proof of work relies on the thermodynamic expenditure of energy (heat). That is why we are building a bitcoin miner that fits inside a water heater. If you already have an electric water heater, you spend $15-$40 a month on electricity for heat. So why not run that same electricity through some silicon first, generating both heat and revenue (in Bitcoin)?

The thought process behind this is pretty simple. If you are already spending that money on electricity anyway, the miner only has to pay for itself to ROI. This is the simplest advantage of this type of mining. Another advantage we have over centralized miners is that we cannot make money from using our own device. Because the capital cost is higher than a conventional miner, it is only profitable through a multi value stream setup. You can only make money if you use the waste heat to offset the cost normally associated with that waste heat. Individual subsidization is used as an anti-centralization mechanism.

Additionally, because of the second value stream of hot water from the electricity, the profitable lifetime of mining can be extended to the end of the hardware's working life (when the hardware dies). This allows much more time to reach ROI, greatly reducing the risk of difficulty catching up, and therefore ensuring less risk in participating. You can spend a lot of energy for a few dollars reward because you would spend that money to heat your water anyway. The only thing you give up on the mining side is the opportunity cost of buying a newer, more efficient, miner. On the hot water side, you could potentially lower running costs by using a heat pump water heater. However, the upfront cost for such water heaters is roughly the same as the projected cost for our miner/heating element, and our system will become cashflow positive sooner — in other words, this is a better investment. (Plus, a revenue stream of internet money is way cooler than saving a few dollars a month in electricity.) In the event of an upswing in the Bitcoin price, you could find the miner paid off even sooner. And you get all of this at greatly reduced risk, unlike mining for a hobby (i.e. mining at a loss because of altruism and/or possible moon landings).

Our pay back curve looks something like this. The red line is a conventional miner, and the blue line our new type of miner: we can’t have final numbers yet or know absolute positions on the graph, but the positions of these lines relative to each other and their general shapes/trends are correct.

Even without knowing the final price or ultimate efficiency of our miner, you can see how subsidized electricity costs necessarily flatten the ROI curve to a limit. The hashrate in this graph is pure Moore’s law, and it assumes constant demand for Bitcoin (price/coin doubles at each halving). Even if the hashrate gets cut in half at the halving, it should be much preferable to have the dual value stream miner. You might even be more profitable after a price decline, as you will be the last miner to turn off, giving you a larger share of the new mining pool equilibrium.

We have this new type of miner mostly built.

We started with a thermal test rig to verify that the basic heat transfer characteristics needed for this miner were indeed possible. After the results of that test easily met minimum requirements (and generally exceeded expectations), we proceeded to design the final form factor of the miner. We currently have the mechanical prototypes made and are finishing our first prototype of the power converter circuit. However, we still need ASIC chips to complete the design and build a fully functioning miner.

Mining chip manufacturers (aka centralized miners) have no reason to sell us chips; they have sunk capital to protect and this is understandable. But by protecting their own investments, they are also forcing us to place more trust in them by creating barriers to entry for newcomers. In order to get into the mining space these days, you need $10-20 million to develop and make your own ASIC. We don’t have that sort of money, and that sort of funding requires a lot of effort to get. We are posting this information in hopes of some community feedback to make sure our idea has any merit. Maybe we are just crazy. We have been running the numbers on this thing for quite sometime (many different ways) and they almost always work out. For some rough numbers see our brief technical overview.

Please contact us, comment, or take a short survey to let us know your thoughts. We want your feedback. We have been thinking about this for a couple years, but only recently figured out a way to actually build it at a potentially marketable price point. We feel we have a handle on the mining market, and have tried to use our blog posts to illustrate our current understanding of Bitcoin and to help you understand why we think this miner is a necessary step.

The long term implications of this style of computing can be applied to many other things. See our blog post here about what the future might hold forindividual home computing.

For a more detailed explanation of how our new miner functions, click here.

As illustrated in our analysis of the “information universe”, the integration of proof of work into the Bitcoin system is what makes this particular machine unique. Without proof of work being distributed participants lose trust in the system, and that makes Bitcoin far less valuable. In the analysis, we used the analogy of the blockchain residing in a carefully crafted digital universe with deterministic laws of operation that set it apart from all other digital space. In the Bitcoin universe, bits of information represent direct value. To better illustrate why proof of work is crucial, we are going to explain it in terms of a ‘portal’ into the Bitcoin universe (where the blockchain resides) and a ‘gatekeeper’ that allows bits of information into that universe.

The ‘gatekeeper’ is simply a revolving job title retroactively applied to whoever solved the last block. Therefore the chance of being the gatekeeper is the same as the chance of solving a block. So if someone had a majority share of the network power they would be gatekeeper more often than all other network participants combined. This means that they could allow whatever they wanted into the blockchain, and no one else could keep up with the decisions being made (51% attack). Obviously if they do not conform to the protocol their invalid blocks will not be propagated, but if they are determined they could put create a bunch of sybil attacking nodes which accept the invalid blocks, forking the network.

This thought process can be applied to most types of cryptocurrencies. Each has their own ‘portal’ to their own constructed reality, checked by many observers. Many share gatekeepers (mixed mining proof of work verification), and each has it own set of observers. The gatekeeper is chosen from the network and their specific proof of work. Right now there are only two main networks, SHA256D and SCRYPT. Two gates all proof of work cryptocurrencies must pass through.

If each hash is an entry to the gatekeeper lottery, and the gatekeeper gets to say what reality is, then a hash is a vote for your reality. This is not only for consensus on the blockchain, but consensus on the deterministic laws of the constructed reality in which it resides (the Bitcoin universe). Because the miners decide what is “valid” there has to be majority agreement on any change to the protocol (see deployment requirement for BIP 101). This is exactly the way in which we should reach consensus. So who gets to “vote” on these proposals? People who control hardware. Therefore the power consumed by mining is a form of identity for controlling the Bitcoin blockchain, and voting on how Bitcoin even operates. It is rudimentary, but it allows proportional direct democracy: the only people who can participate in the voting process are those ‘identities’ who can prove that they are contributing to the network via consumption of energy.

We saw in Bitcoin, that when Ghash.io got near 50% people started moving their power to other pools to ensure diversity in agreement. This worked for two reasons: 1) individuals were in control of where their hash power was directed, and 2) individuals were actively participating. Both of these reasons are things we cannot always count on as Bitcoin grows and the mining landscape changes. If we have more participants with preconfigured miners (if mass adoption happens) they will likely not be bothered to move their power away from the default setting (or manufacturers may not allow them). So we need a way to spread around the ‘votes’ of these apathetic miners in an equally distributed way, while still allowing those who want to control their power/vote the ability to put it where they choose (an opt-out). Additionally, if mining continues to centralize and the hardware is controlled by fewer and fewer people, we’ll have no choice but to trust the remaining owners to not have malicious intent — not ideal for a trustless system. Unfortunately, centralized miners have a lot more votes and could influence Bitcoin’s development so as to make the mining environment increasingly advantageous to them as time goes on. The more mining centralizes, the more Bitcoin will come to resemble a consortium blockchain, with a shrinking pool of gatekeepers making all of the decisions.

One could argue that if centralized miners attempted to influence Bitcoin in this way, they would be destroying their own investment, and therefore would not actually do anything to harm the network. That may be true of today’s miners, but we can’t know who the miners may be ten years from now. And even if no miner ever deliberately manipulated the network, what would be the next largest entity that might take an interest in exerting control? A government. If mining is centralized to a few places, it becomes much easier for governments to shut down large portions of the Bitcoin mining network — and even if a government instead decided to co-opt a portion of the network, this both decreases trust in the network overall and sets a dangerous precedent for other governments to follow.

The prospects become worse if the country deciding to declare authority over its miners is the one containing half the mining network, because this could disrupt the network if by turning their miner off or mining empty blocks. This starts to chip away at the trustlessness of the system, if the block enough transactions other measures besides the heaviest chain would have to be used to determine the correct chain otherwise they could partially deny service. In such a case, the remainder of the network (now less than half as large as before) becomes the ‘true’ currency, but it is now even more vulnerable to a centralized entity gaining majority control(Denial of Service). The process repeats if the country housing the next largest share of mining likewise decides to exert its power, causing a cascade effect that ends in an implosion of trustless consensus making. In such a case, even starting over with a new currency would no longer be a realistic option. A new proof of work algorithm would be needed to invalidate the hardware now owned by the government. But to create a new proof of work algorithm, you would have to begin at the CPU or GPU level. The majority of CPUs and GPUs are in data centers, supercomputers, and government organizations, making it easily conceivable that a government could commandeer and/or overpower any attempts at a new proof of work. And even assuming they left alone any new currency attempts, the damage would be done. They would already have completely destroyed all the value built up and stored in the previous system, not to mention long-term public faith in any future cryptocurrency.

Scenarios such as those above may not be immediate concerns for Bitcoin, but they do illustrate how controlling the proof of work devices is the same as controlling the whole system.

Proof of work will always come to majority consensus, but that majority must be representative of the system's stakeholders to remain meaningful and trustless. With hashing power more evenly distributed we could use it as a sort of digital identity in that hashing power can't be faked. There are surely ways in which true digital identities could be obtained much more easily, but the important thing to understand here is that digital identity in this context is a sliding scale of spent energy/value. The more the Proof of Work is distributed, the more individual identities and ‘voters’ the network will have, the more inline the votes will be with the user's wishes. So how do we make that happen? Read our next post for our attempt at beginning the thermodynamic distribution process of the proof of work.

This is the first of four posts about how these sort of systems work, problems with the current model, and thoughts on potential ways to fix it. Of course, what explanation and potential solution is not complete without a cyberpunk vision of the future.

To really understand the trustless currency-ledger system, we need to break it down to its component concepts.

“Trustless”. What we really mean by “trustless” is that no single entity has control or unfair power in the system. Trustlessness is a result of the inbuilt market incentives that get individuals to participate in the network and adhere to the protocol. The incentive will always act like a currency, as it has to have value to pay for security. It is no different than a mechanical governor, or electrical feedback circuit. It is just a feedback system trying to reach its designed equilibrium.

However, this system is somewhat special in that it straddles both the physical world and the digital reality, ultimately giving us the “currency-ledger” dynamic.The Bitcoin universe (Bitcoin’s digital reality) is similar to all other digital space, except that the laws of the Bitcoin space are dictated by the algorithms and cryptographic framework set forth at the creation of the system. Some examples of these natural laws would be the block size, the block reward, or the difficulty. The blockchain/ledger resides within the framework of this constructed reality. The currency is the energy that allows movement of value around the blockchain. The proof of work algorithm is what binds the Bitcoin universe to the real world. This is how we can have a trustless system and a non-replicable digital token. Because the cost of securing the blockchain is ensuring the Bitcoin universe maintains continuity and consistency, it must be tied to something that we cannot cheat or fake, something that is not digital: namely, thermodynamics.

At the core of what gives Bitcoin its unique value and true revolutionary potential is the distributed consensus, the trustlessness, that proof of work allows. This never-before-created information machine is what we see currently functioning. However, with the proof of work becoming centralized, we are quickly undercutting the trustless nature of the system. Our feedback system is stuck, and pushing the system to the wrong type of equilibrium. This is a fundamental problem to any trustless distributed consensus system, and must be solved at the fundamental level: we have to distribute the network through incentives to build the right infrastructure.