A lot of people have questions about cryptocurrency, mining, and what this has to do with a rise in GPU prices.
You can find more information at various other subreddits dedicated to the subject:
What is cryptocurrency?
Currency, in general, is a medium for exchange that is based on promises for what that currency is worth. Commonly called "money."
Cryptocurrency is not centrally controlled or regulated and their value is based on the supply and demand; there are built-in limits for how much can exist (potentially curbs devaluation), public logs of the transactions (blockchain) and the cryptography algorithm make it difficult to counterfeit.
Maybe this old TechQuickie
can explain it better.
TL:dr an unregulated form of digital money
What is mining?
The most basic way to acquire a cryptocurrency is the same as acquiring money, in exchange for goods or services.
The other way to get cryptocurrency is by mining, or solving increasingly difficult math problems in exchange for the cryptocurrency.
What does this have to do with GPU prices going up?
Crypto mining started on CPUs, but it didn't take long for people to realize that GPUs, especially the 'heavy duty' ones intended for gaming, are really good at it.
The downside to GPU mining is heat and power consumption, this lead to mining systems designed for the task and eventually ASIC chips designed just for mining.
As a cryptocurrency matures the math problems become very difficult, leading to pools of miners that share resources - this has also lead to some malware using infected systems for mining.
An older currency like Bitcoin is well into that 'pooled specialty hardware' age, but newer options like Ethereum are aimed at GPUs; this increased demand means lower supply which means higher prices.
When will prices go back to normal?
Probably if or when the cost to mine via GPU exceeds expected returns.
There are some specialized cards set to hit the market which may ease the demand on enthusiast GPUs.
A word of warning, when this happens the market will be flooded with GPUs that were used for mining. The lower price may seem attractive, but these cards have been used in harsh conditions, 24/7 for who knows how long. Mining cards probably won't run very well/very long or they could work fine. You have to decide if it's worth the risk.
Should I start mining?
That is entirely up to you, but please take some time to educate yourself on the risks and benefits before you decide.
Take a look at a profitability estimator
to get an idea of what you might expect. These cryptocurrencies can be very, very volatile so don't quit your job expecting to strike it big with mining. Consider the cost of taxes, cooling, hardware, replacement hardware, and power.
Mining is hard on hardware, the wear and tear means things like fans, the GPU(s), and other parts may die prematurely. Keeping your hardware cool (about 300W to cool 1000W) can lead to additional costs for hardware and power or reduced output. The cost of electricity may not seem like much, but it can be enough to make or break a mining setup.
Take the time to figure out your ROI.
Should I sell my card?
Again, this is entirely up to you. There may be situations where selling a card and upgrading with the money can work out, but there are a lot of factors at play there, so do your research.
Check selling sites: eBay, Craigslist, hardwareswap
, LetGo, Facebook, and other (similar) selling sites for pricing. If shipping, package it how you bought it: clean, inside an anti-static bag, in a cardboard box with some padding. If meeting someone IRL, be careful, meet in a well-lit public place - some areas have exchange locations at places like police stations. I'd like to take a quick moment to thank Linux_PCMR for some insight, Graphics_Nerd for gathering some links for review, and the kind users that have replied to the number of posts on this subject.
E-mail as it is, is horribly broken. Horrendously broken.
It wasn't that many years ago that you could be assured your e-mail reaches whoever you were mailing to. Today it is a mere suggestion, that perhaps this should be delivered to this person, at least for any automated e-mail. This seems to be creeping to manual, organic email as well. Hell, we are seeing even internal e-mails being flagged by spamassassin as spam, organic, human written conversations! In that instance, the spamassassin is also maintained by one of the largest hosting providers in the world...
Hotmail/MS services has been for years (atleast about 4 years now!) been silently dropping email, not all, but some. There's a bit of relief lately, as they have started to favor a bit more marking as spam, rather than silently dropping.
I know, most email users don't see this problem, but those who use email a lot to do their work, and those who need to send automated emails (say, welcome e-mails for a service) this is a big problem. (Disclaimer
, for us, our niche of hosting probably causes flagging as well. Our site is blocked by many corporate firewalls for example) Blockchain to the rescue?
This is an idea i've been toying around with a few years. What if any single e-mail would cost a faction of a cent, and who receives the e-mail, gets paid for it? Now that would solve a lot of problems. I realize there has been some half assed attempts on blockchain based e-mail, but they are about replacing email (never going to happen). Using blockchain to enhance the current experience, with least minimal friction should be the goal, not re-inventing the wheel.
Imagine a say 0.01 cent (0.0001 USD) cost per e-mail. This price would not be cost prohibitive even for free e-mail service providers (Ad revenue etc. should exceed this value), never mind any legit e-mail users. Especially considering you get paid for receiving. So all legit e-mail services would work rather well regardless of the cost. (never mind free email service could profit from this)
Spam however? To send 1 million emails you would need to pay 100$. How many spammers would continue doing so? At least it makes things much harder, not so easy to use a botnet to send your email when you need to include your private key(s) to the botnet, or make some kind of private key management system, makes more complicated.
Small business newsletters? Say you need to send 100k e-mails to legit customers, 10$ is nothing
. To human time crafting that newsletter is order (possibly orders) of magnitude greater than that.
Price would also fluctuate as per the market. The most difficult thing would probably be setting the self balancing mechanisms to keep per mail cost sensible. As such, the biggest hurdle in this might not be technical at all. Technically, how could this work?
Sender sends a TX for e-mail they are sending for recipient. This TX contains message with mail ID, and a segment which can be used with the email contents to unlock the private key for the payment. This way it is verified that recipient mail servers receives and reads the email. Once the recipient server has calculated the private key, they can either TX the received sum to their wallet, or this needs to be formatted so that once the sender has sent it, they cannot recover the private key and double spend (technical hurdle A. For someone who knows their stuff unlikely to be an major hurdle)
Step by step repeat: * Sender checks if recipient has "MailCoin" capability * Sender sends TX to recipient * Sender sends the email to recipient * Recipient notices on mail header (say x-mailcoin-tx: TXID_HERE) that this is a "mailcoin" mail * Recipient checks TX if it has been received * Recipient puts the mail on delivery queue, antispam is instructed of heavy negative score (MTA admin configurable) * Recipient claims the value of the TX (this is the hurdle A). Recipient can only claim the TX value in case they have received the full e-mail. (Question, can this step be pushed even further down the delivery chain, but still remain MTA only level without mail client support?). Most likely solution is that the header contains the encrypted private key, and chain TX contains the key to decrypt that private key to claim the coins, or vice-versa?
Once recipient has the email & payment, they simply mark on their Antispam a automatic lower score and deliver it normally.
E-mail server side we have several components:
- Recipient server needs component to check for TXs
- Sender server needs component to send a TX, and check if the recipient has the capability
- Recipient server OR dns zone has indicator the recipient can utilize "MailCoins" (DNS IN TXT field most likely, ie. mailcoin.domain.com. IN TXT publicaddress)
Most typical scenario would be the Recipient server works as outgoing as well, with single wallet. So depending on your mail volume, do you send or receive more on that wallet you might never need to worry about the coins (except for value going skyhigh and having like 10k $ worth of "MailCoins").
So perhaps additional components on per use case are needed, or more likely rudimentary scripting capability (ie. "MailCoin" daemon api) to keep the balances in check. Technical hurdle B
: This needs to be super super simple
to setup. Or sufficient financial incentive. One would need to develop standard components & configs for exim, postfix, and other MTAs. Infact, make it autogenerate wallet ID etc. and easy to replace or import private keys etc. to put in coins for sending if you need to. Privacy
: On the blockchain you would not see the e-mail contents, only that e-mail likely took place (TX with mail UUID) to recipient. If sender can be deciphered it depends on them if it can be traced who they were. Automatic mixers? :) Recipient can also keep cycling the receive addresses to keep things private if they want to.
The biggest problem i see here, is that if an attacker can deduce the sender and/or recipient, it might to lead to some issues out of the scope of technical solutions. If attacker could read the emails, they would already have accomplished MitM and could just grab all e-mails.
Default implementation should be so, that from recipient address outsider cannot deduce the recipient server nor hostname.
Also, if attacker gains access to your mail with full headers, they could see the TXs in blockchain. MTA might need to scrub mailcoin related headers (yuck, scrubbing headers ....) for paranoid users, but most likely solution is that recipient retransmits those mailcoins as soon as they got the private key for the balance. Blockchain:
Blocks needs to be done every 10seconds or so, it needs to be fast. Preferrably even every 5 seconds, as not to cause any undue delay. Then again, if
your application is reliant on receiving email within seconds, one should consider another means for communicating. Imho
, email should be considered a little bit like snail mail, but on internet pace: Couple minutes delay is just OK.
Block size given the e-mail volume needs to be fairly large as well, considering the time between blocks. This is technical hurdle C: Hosting the full blockchain. I can easily foresee that this would grow to be terabytes in size. However, any large email operator would have vested interest in ensuring smooth operation of the blockchain, and for them, running a full node would have neglible cost.
(Technical hurdle C) Single email sent using the system could easily have TX contents of 100 bytes + TX headers + block headers etc. Say 100 bytes, and 100 million emails per day: 9.31GiB per day, 3 399GiB per year, 5 years later: 16.60 TiB just for the mail TXs.
Some estimate there is 200+ billion emails per day, but we all know large portion of this is spam. But even at 50 billion emails a day, 100 bytes per mail TX would add to 4.55TiB per day! So optimizing the blockchain size is obviously going to be important. The volume will be obviously much smaller as semi-spam (those daily half opt-in spamvertising from companies you know) will be lower as well. So probs 100+ billion emails per day at 100% adoption.
Blockchain should then be compressed, the whole block. Algorithm probably should favor speed over compression rate, and should be task specifically optimized (needs a simple reference release, where you can just stream the block contents into it and get output as compressed or uncompressed). The more compression there is, the more full nodes will be hosted by smaller operators :)
For large e-mail server clusters there should be central store for the blockchain, but this can be accessed on the system administratoconfig level already. The MTA components will just remotely talk to single full node daemon (so not really different from many implementations in existence right now), instead of each one running locally a full node.
At today's cheapest hosting rates 16.60TiB is roughly around 85-100€ a month. Purchase cost per 8TB drive is around 230€ mark right now, externals are cheaper. Not an issue for any even semi serious mail provider. Not even issue for datahoarder individuals. However
at 100 billion mails per day: 9.09TiB per day added, which is prohibitively large! We should be targeting something like 20bytes per mail final storage spent, or even less.
If it looks like it is going to grow really large, full node needs to have configurable multiple storages, so they can store parts of the blockchain on multiple different devices (ie. individual might choose to have it on 4 different external drives).
Filesystem side optimizations are needed as well, but these are fairly simple, just split into multiple subdirectories by the 10 thousand blocks or so, ie. 1 for blocks 1-10k, 2 for blocks 10 001 to 20k etc. Filesystems get exponentially slower the more files there is per directory. 10k might start to show slowing down, but is not significant yet.
Nodes could also implement secondary compression (compress multiple blocks together), if the blockchain starts to become stupid large. If it starts to become impossible to maintain, we could possibly implement a scrubbing methodology, where very old blocks get the TX contents wiped as they are not necessary anymore. Should not be an issue
Blocks with 10second target generated per annum: 3 153 600 Mails per 10second: 115 740 e-mails per 10second block. Final compressed size (say 20 bytes per mail): 2.20MiB + headers etc. per block Let's start small and allow linear growth to this, say 0.1% per day (36.5% annual) and start from 20k / 512KiB. After 3 years: 41.9k / 1072.64KiB per block, After 10 years: 93k / 2380.8KiB. (2027 we should have HDDs in the size of 30TB and daily max size for chain growth is 19.61TiB) On the positive side
every problem is an opportunity in disguise. If the blockchain is large, once again botnets will have a hard hard time to spamming, they can't host the full blockchain on infected machines. They will need to develop centralized mechanisms on this regard as well. One method i can see is by having TOR client built in, and via .onion domain to anonymize, but this is two way street, security researchers could exploit this (see above about the private keys) as well. Even without botnets, spammers will need to dedicate significant resources to host the full blockchain.
On the flip side, if spammer has also mining operation on the same local area network, they have both the income for mailcoins + full blockchain, and could leverage economies of scale, but this too would increase cost. And after all: This is all about increasing cost for spamming, while having the price in vicinity where real e-mail users, real businesses it is not a significant impact, or may even be an income source Client side
Zero, Nada changes. No changes to outlook, thunderbird etc. Everything works under the hood at the MTA level. Very easy adoption for the end user. Everything is in the backend, server side. Economics for users
Cost of operation has above been shown to increase wildly for spammers. But how about normal use cases? Joe Average
: They receive e-mail a lot more than they send, all kinds of order confirmations, invoices, newsletters and other automated e-mail. They will actually earn (however tiny amounts) from using this system. So for the masses, this is a good thing, they will see the earning potentials! which brings us to .... New business opportunities
! I could foresee a business setting up spam traps, the more e-mail you receive the more you earn! So it pays to get your receiver into spam lists. You don't ever need to read these, just confirm receive of them. All of sudden we could see even greater numbers of invalid e-mail addresses in spam lists, making spamming ever more expensive! Free email services
might proof to be extremely profitable, to the point of potential revenue sharing with Joe Averages (and above spamtraps). Because free email is mostly joe averages, they will have greater influx than outgoing. On the caveat, free email needs to have limits, but due to the low cost and potential of earnings, they could implement "mail credits" system, base is like 20 emails a day, but each received email could increase this credit limit. As such, it makes actually sense for free email services to implement this at the very least on the receiving side. Business mass emailings
. A business which has 100k valid e-mails on their database will not have a problem with paying few dozen bucks to have their mass mailing delivered. BUT they will make extra sure the content is good and targeted, something the recipient wants to receive. These will be the biggest spenders on email, apart from spammers. ISPs
, hell they get paid to provide e-mail. And they are on the same spot as free email service providers, they stand to earn more than spend! Blockchain economics
This is where things might get interesting, there is so much potential.
However, there are several things definitively should not be done:
- Initial Coin Offerings
- Non-consensus chain (ie. whatever devs say, goes!), always has to be consensus
- Infinite & preferential supply
1 & 2 are easy, just do not mine outside of testnet prior to launch. (If devs get paid by companies, there is conflict of interest as well, but let's not get into that right now)
3: Miners and/or full node maintainers decide what goes on. Probably miners like bitcoin is supposed to.
4: Infinite & preferential supply: No after the launch "contracts" etc. to give coins to preferential parties, it should remain as on the launch unless majority consensus says there will be a change. Proof of stake is gray area imho
, but then again also proof of work is the rich gets richer. Mining
: Storage requirement is a blessing in disguise
, the massive storages required for this to function means that there will be no central hardware developer who sells all the shovels, without significant other markets. Ie. WD, Seagate, Toshiba the main players.
This means algo needs to be based on the full blockchain being hosted. The hashing needs to be so that GPUs are the king most likely, since almost anything good for CPUs is also doable in GPUs. Eventually someone will likely come with ASIC alternative, but due to masses of data it WILL require high bandwidth, high memory. Nothing like bitcoin currently, where low bandwidth, no memory requirement for the ASIC. There needs to be some expensive commodity components in there (RAM, Storage), and as such GPUs are the most likely candidate, and the bottleneck will not likely be computation, but I/O bandwidth.
Quickly thinking, previous block could include number of blocks to be included on the next for verification, in a highly compressible format. Let's say difficulty is number of blocks to be hashed, or from difficulty you can calculate number of blocks to be included. Previous blocks miner just chooses on random blocks to be included on the next one. Listing 10 series of blocks to be included, which can include series instructions. It could request block #5729375+100, or #357492+500 stepping 5 (every 5th block). Hell the random generator could use last block as seed for the next one to make it deterministic YET random as the emails and TXs change. (WTF, Did i just solve how the algo needs to work?!?) Only blocks which would differentiate is the first few, and obviously Genesis, for which an "empty" block would be what is to be hashed.
Hashing algo could be SHA256 because of the high requirement of streaming data, and most ASIC miners lacking in bandwidth (infact, it could be made compatible with bitcoin, but only those ASICS with higher I/O bandwidth than storage/ram I/O bandwidth is could actually boost the perf)
Different hashable list operations could be (on the block list what to be hashed on the next one): * Single block * Block # + number of blocks * Block # + (number of blocks with stepping) * Block # + number of blocks chosen by random using each hashed block as the seed for choosing next one (makes prefetch, preread, caching not work efficiently) * Number of previous blocks mined (ie. 50 last blocks) * Above but with stepping operator * Above but with choose random next X blocks, with variations based on the last hashed, sum of the hashed * All random pickers would have operation modes for the seed to be used: From hashed sum, the whole block, block contents, block header
These modes would ensure the blocks are there and makes it a lot dependable on variable factors, RAM speed, I/O seek time, I/O bandwidth.
This way we have proof that the miner has access to those blocks in efficient manner and the full blockchain is stored there, even if it is not practically retrievable from him / her over the internet for others to obtain a copy. HOWEVER, due to the data volumes, i think it is given they have fast access, but a miner would probably prefer not to share their blockchain contents to have bandwidth free for their mining, as the deadlines are tight. It could be built into the full node spec that they do not accept new blocks from sources which are not ready to supply any given block, and perhaps even periodic test of this. However, this would be unenforceable if people start running custom coded nodes which disables this, as it is not part of the blockchain calculation. It is not miner's benefit to "waste" precious bandwidth to serve others the vast blockchain, meanwhile it is end users benefit those running full nodes without mining to get them fast. So an equilibrium might be reached, if miners start loosing out because other miners will not share their blocks, they will start offering them, even if prioritized.
At 2MiB blocks, 10 second deadline, a miner would preferentially want the new block within 500ms, which would be barely sufficient time for a round trip across the globe. 500ms for 2MiB is 4MiB/s transfer rate inbound, and when block found you want it out even faster, say 250ms you'll need 8MiB/s burst which very very few have at a home. At more usual 1MiB/s it would take 2secs to submit your new block. On the other hand, if you found the block, you'd have immediate access to begin calcing the next one.
Block verification needs to be fast, and as such the above difficulty setting alone is not sufficient, there needs to be nonce. Just picking the right block is not guarantee there will be match, so traditional !???? nonce needs to be set as well most likely. As such
, a lot of maths needs to be done to ensure this algorithm does not have dead ends, yet ensures certain blocks needs to be read as full and stored fully by the miners, just plain hashes of the blocks is not sufficient.
Perhaps it should be block data + nonce, then all the blocks hashes (with nonce, or pre-chosen salt) and to be generated block combined hash with nonce needs to have certain number of zeroes. Needs testing and maths :)
So there are many ways to accomplish proof of storage, we'd need just to figure out the which is the best. Sidenote
, this same algo could potentially be used with different settings for immutable, forever storage of data. Since there is no continuing cost to store data, TX Fee for every message (data) byte should be very high in such a coin. Supply
. Needs to be predictable and easy to understand. It would be preferential the standard mailing out is always 1x MailCoin, albeit coin itself should be practically infinitively divisable, and as such supply needs to be in the trillions eventually. But these things get complicated really fast, so we need to set a schedule.
Current email use is very large, so we should have something in the same magnitude. 8640 blocks per day - so maybe 10 000 coins per block == 86 400 000 new coins per day == 31 536 000 000 new coins per year, halving every 2 years. First halving: 63 072 000 000, Second halving: 94 608 000 000, Third (6 years): 110 376 000 000, but only halving 4 or 5 times to keep some new supply for ever increasing adoption and lost coins. Got all the way here? :D
Thanks for reading up. Let me know what you think, and let's start a discussion on the feasibility of such a system!
I cannot develop this myself, but i would definitively back an effort up in the ways i can if anyone attempts to do something like this :) And i know i got probably many of the details incorrect
The main point of the methods described above is ease of adoption. Without adoption any system is worthless, and with email, you just cannot replace it like that (see the attempts trying to replace IPv4 with IPv6 ...), but you can enhance it. adoption is very critical in communications systems. (No one would have a phone if no one else had a phone) Addendum 1:
Forgot to add about pricing and markets, read comment here Addendun 2: Bad actors and voting
ASIC Miner profitability ranking. Power cost $/kWh. Model Release Date Hashrate Power Algo Revenue 24h Profit 24h Top Coin; Bitmain Antminer Z15: Jun 2020: 420.00 kh/s: 1510W: Equihash $22.95 $19.32 Hush(HUSH) Equihash Goldshell HS3: Jul 2020: 2.00 Th/s: 2000W: Handshake $22.74 $17.94 Handshake(HNS) Handshake Innosilicon A10 ETHMaster 500Mh ... Siacoin Mining Profitability Calculator. Mining Siacoin can be very profitable, however, it isn't always clear how much you will make. This calculator is designed to help make this process a little easier. It works for both mining rigs at home as well as cloud mining services. Accurate Bitcoin mining calculator trusted by millions of cryptocurrency miners since May 2013 - developed by an OG Bitcoin miner looking to maximize on mining profits and calculate ROI for new ASIC miners. Updated in 2020, the newest version of the Bitcoin mining calculator makes it simple and easy to quickly calculate mining profitability for your Bitcoin mining hardware. Electricity cost is probably the factor that has the most impact on mining profitability. After all, Bitcoin's SHA-256 mining algorithm is classified as Proof-of-Work (PoW) because work must be done to validate the network. The 'work' is computational power – therefore electricity is required to validate the network. Most profitable ASIC miners in 2019 ASIC miner manufacturers wants to make each new device more profitable by increasing the hashrate, reducing power consumption, cost of the device and choosing the most optimal algorithm for mining cryptocurrency.
In todays video we breakdown all of the information we know about the Innosilicon GRIN miners as well as what we think the mining profitability of these GRIN miners will be once they all hit their ... What are the TOP 5 coins to mine for best profitability in 2019 with ASIC miners?! Today we will review Bitcoin, Ethereum, Dash, Litecoin, Zcash, and Grin mi... Bcause LLC a Bitcoin Cash mining farm "data center" is launching a crypto trading and futures service. Scott Offord has released the most extensive and best Bitcoin mining hardware profit calculator. This quite literally cuts mining profitability in half for Bitcoin mining farms. Currently it costs about ~$3,000 to $4,000 dollars to create a Bitcoin with an industrial Bitcoin mining farm ... 03:22 How to calculate mining profitability 04:28 Dash mining profitability in 2019 + best miners 07:52 SiaPrime Blake2b mining - A3/S11 08:38 Bitcoin mining profitability in 2019 + best miners 11 ...