Top 25 AI chip companies: A macro step change inferred

The total computing power now dedicated to securing the bitcoin blockchain has set yet another record.

According to data from mining services operator, the average bitcoin mining hash rate over the last two weeks has reached 71.43 quintillion hashes per second (EH/s), up from 64.49EH/s on July 23. The threshold was breached as bitcoin adjusted its mining difficulty at block height 586,672 on Monday 2:52 UTC – that is a 6.94EH/s, or 10.78 percent jump since mid July.
Bitcoin mining difficulty is a measure of how hard it is to compete for mining rewards on bitcoin. Just how difficult the bitcoin software makes it to generate new blocks adjusts every 2,016 blocks – approximately every 14 days – to ensure the block production time remains about 10 minutes at the next cycle.
Assume this additional 6.9EH/s (or 6.9 million tera hashes per second, TH/s) computing power has all come from powerful ASIC miners, such as Bitmain’s AntMiner S17 or MicroBT’s WhatsMiner M20S, both of which boast a mining rate of around 55TH/s and recently hit the market.
That means more than 100,000 top-of-line ASIC miners could have been switched on within the past two weeks. Further, given these products have been sold for at least $2,000 each, this equates to some $200 million in revenue pocketed for major miner makers.
The continued interest in bitcoin mining comes at a time when the cryptocurrency’s price appears to be en route to challenging all-time highs, however distantly, and amid the arrival of the rainy season in China, which leads to cheaper hydropower electricity costs in the country’s southwest provinces – a region that is reported to account for 50 percent of the global mining activity,
Miners in China estimated earlier this year that bitcoin’s hash rate in the summer would break the level of 70EH/s. To be clear, at several single points of time, bitcoin’s hash rate had already crossed that level in June and even reached 80EH/s around Aug. 1.
However, today marks the first time that the two-week average computing power has been able to remain above the 70EH/s threshold. As such, bitcoin’s mining difficulty has also set a new record of nearly 10 trillion.

Market change

Amidst this uptick in mining interest, there have been notable changes in the mining market, where top manufacturers are racing to produce more powerful equipment.
For instance, in Bitmain’s 2018 initial public offering prospectus, the Beijing-based mining giant claimed it had a 70 percent market dominance. Now, it may be facing serious competition from rival players that some believe are capable of shipping more top-of-line products with better profitability.
Michael Zhong, a former mining analyst who now operates mining farms at a startup called Force Mine, told CoinDesk that based on his experience, the production capacity ranking among major Chinese miner makers for their flagship products have changed over the years.
Zhong explained that from 2017 to 2018, Bitmain had topped the list with its AntMiner S9 series miners, followed by Canaan’s Avalon 8 series machines. InnoSilicon, Ebang and former Bitmain design director’s MicroBT were all in the third position at the time.
But from January to June this year, the delivery capacity ranking has reshuffled, with now MicroBT’s WhatsMiner M20 series at the top, followed by Bitmain’s S17 series miners and then InnoSilicon, Canaan, and Ebang, Zhong added.
According to F2pool’s miner profit tracker, Bitmain’s flagship AntMiner S17 Pro ranks third in terms of mining profitability, following BitFury’s Tardis and MicroBT’s WhatsMiner M20S. The cost for WhatsMiner M20S is around $3,000, while that of AntMiner S17 Pro is around $4,000 each, based on the information advertised on the two firms’ websites.
Although orders for these flagship machines have queued up until November and December this year, MicroBT’s founder Zuoxing Yang told CoinDesk previously that the bottleneck of production capacity is the availability of chips from suppliers.
For example, MicroBT uses 10-nm chips for its M20 series, which are relatively more affordable with a higher level of availability compared to more advanced 7-nm chips used by Bitmain for its AntMiner S17 series equipment.
While Bitmain has always been relying on chips supplied by Taiwan Semiconductor Manufacturing Company (TSMC), MicroBT has switched from TSMC to Samsung earlier this year for its flagship products.
Both TSMC and Samsung have estimated in their most recent Q2 earnings calls that the demand for cryptocurrency mining chips will come back in the third and the fourth quarter this year.
Operating miners image courtesy to Hashage
submitted by Muxa84 to Bitcoin [link] [comments]

Vitalik's response to Tuur

I interlaced everything between Vitalik and Tuur to make it easier to read.
1/ People often ask me why I’m so “against” Ethereum. Why do I go out of my way to point out flaws or make analogies that put it in a bad light?
2/ First, ETH’s architecture & culture is opposite that of Bitcoin, and yet claims to offer same solutions: decentralization, immutability, SoV, asset issuance, smart contracts, …
Second, ETH is considered a crypto ‘blue chip’, thus colors perception of uninformed newcomers.
Agree! I personally find Ethereum culture far saner, though I am a bit biased :)
3/ I've followed Ethereum since 2014 & feel a responsibility to share my concerns. IMO contrary to its marketing, ETH is at best a science experiment. It’s now valued at $13B, which I think is still too high.
Not an argument
4/ I agree with Ethereum developer Vlad Zamfir that it’s not money, not safe, and not scalable.
@VladZamfir Eth isn't money, so there is no monetary policy. There is currently fixed block issuance with an exponential difficulty increase (the bomb).
I'm pretty sure Vlad would say the exact same thing about Bitcoin
5/ To me the first red flag came up when in our weekly hangout we asked the ETH founders about to how they were going to scale the network. (We’re now 4.5 years later, and sharding is still a pipe dream.)
Ethereum's Joe Lubin in June 2014: "anticipate blockchain bloat—working on various sharding ideas".
The core principles have been known for years, the core design for nearly a year, and details for months, with implementations on the way. So sharding is definitely not at the pipe dream stage at this point.
6/ Despite strong optimism that on-chain scaling of Ethereum was around the corner (just another engineering job), this promise hasn’t been delivered on to date.
Sure, sharding is not yet finished. Though more incremental stuff has been going well, eg. uncle rates are at near record lows despite very high chain usage.
7/ Recently, a team of reputable developers decided to peer review a widely anticipated Casper / sharding white paper, concluding that it does not live up to its own claims.
Unmerciful peer review of Vlad Zamfir & co's white paper to scale Ethereum: "the authors do NOT prove that the CBC Casper family of protocols is Byzantine fault tolerant in either practice or theory".
That review was off the mark in many ways, eg. see, and by the way CBC is not even a prerequisite for Serenity
8/ On the 2nd layer front, devs are now trying to scale Ethereum via scale via state channels (ETH’s version of Lightning), but it is unclear whether main-chain issued ERC20 type tokens will be portable to this environment.
Umm... you can definitely use Raiden with arbitrary ERC20s. That's why the interface currently uses WETH (the ERC20-fied version of ether) and not ETH
9/ Compare this to how the Bitcoin Lightning Network project evolved:
elizabeth stark @starkness: For lnd: First public code released: January 2016 Alpha: January 2017 Beta: March 2018…
10/ Bitcoin’s Lightning Network is now live, and is growing at rapid clip.
Jameson Lopp @lopp: Lightning Network: January 2018 vs December 2018
Sure, though as far as I understand there's still a low probability of finding routes for nontrivial amounts, and there's capital lockup griefing vectors, and privacy issues.... FWIW I personally never thought lightning is unworkable, it's just a design that inherently runs into ten thousand small issues that will likely take a very long time to get past.
11/ In 2017, more Ethereum scaling buzz was created, this time the panacea was “Plasma”.
@TuurDemeester Buterin & Poon just published a new scaling proposal for Ethereum, "strongly complementary to base-layer PoS and sharding":
Yay, Plasma!
12/ However, upon closer examination it was the recycling of some stale ideas, and the project went nowhere:
Peter Todd @peterktodd These ideas were all considered in the Treechains design process, and ultimately rejected as insecure.
Just because Peter Todd rejected something as "insecure" doesn't mean that it is. In general, the ethereum research community is quite convinced that the fundamental Plasma design is fine, and as far as I understand there are formal proofs on the way. The only insecurity that can't be avoided is mass exit vulns, and channel-based systems have those too.
13/ The elephant in the room is the transition to proof-of-stake, an “environmentally friendly” way to secure the chain. (If this was the plan all along, why create a proof-of-work chain first?)
@TuurDemeester "Changing from proof of work to proof of stake changes the economics of the system, all the rules change and it will impact everything."
Umm... we created a proof of work chain first because we did not have a satisfactory proof of stake algo initially?
14/ For the uninitiated, here’s a good write-up that highlights some of the fundamental design problems of proof-of-stake. Like I said, this is science experiment territory.
And here's a set of long arguments from me on why proof of stake is just fine: For a more philosophical piece, see
15/ Also check out this thread about how Proof of Stake blockchains require subjectivity (i.e. a trusted third party) to achieve consensus: … and this thread on Bitcoin:
Yes, we know about weak subjectivity, see It's really not that bad, especially given that users need to update their clients once in a while anyway, oh and by the way even if the weak subjectivity assumption is broken an attacker still needs to gather up that pile of old keys making up 51% of the stake. And also to defend against that there's Universal Hash Time.
16/ Keep in mind that Proof of Stake (PoS) is not a new concept at all. Proof-of-Work actually was one of the big innovations that made Bitcoin possible, after PoS was deemed impractical because of censorship vulnerability.
@TuurDemeester TIL Proof-of-stake based private currency designs date at least back to 1998.
Oh I definitely agree that proof of work was superior for bootstrap, and I liked it back then especially because it actually managed to be reasonably egalitarian around 2009-2012 before ASICs fully took over. But at the present time it doesn't really have that nice attribute.
17/ Over the years, this has become a pattern in Ethereum’s culture: recycling old ideas while not properly referring to past research and having poor peer review standards. This is not how science progresses.Tuur Demeester added,
[email protected] has been repeatedly accused of /criticised for not crediting prior art. Once again with plasma:
I try to credit people whenever I can; half my blog and posts have a "special thanks" section right at the top. Sometimes we end up re-inventing stuff, and sometimes we end up hearing about stuff, forgetting it, and later re-inventing it; that's life as an autodidact. And if you feel you've been unfairly not credited for something, always feel free to comment, people have done this and I've edited.
18/ One of my big concerns is that sophistry and marketing hype is a serious part of Ethereum’s success so far, and that overly inflated expectations have lead to an inflated market cap.
Ok, go on.
19/ Let’s illustrate with an example.
20/ A few days ago, I shared a critical tweet that made the argument that Ethereum’s value proposition is in essence utopian.
@TuurDemeester Ethereum-ism sounds a bit like Marxism to me:
  • What works today (PoW) is 'just a phase', the ideal & unproven future is to come: Proof-of-Stake.…
21/ I was very serious about my criticism. In fact, each one of the three points addressed what Vitalik Buterin has described as “unique value propositions of Ethereum proper”.
22/ My first point, about Ethereum developers rejecting Proof-of-Work, has been illustrated many times over By Vitalik and others. (See earlier in this tweetstorm for more about how PoS is unproven.)
Vitalik Non-giver of Ether @VitalikButerin: I don't believe in proof of work!
See above for links as to why I think proof of stake is great.
23/ My second point addresses Ethereum’s romance with the vague and dangerous notion of ‘social consensus’, where disruptive hard-forks are used to ‘upgrade’ or ‘optimize’ the system, which inevitably leads to increased centralization. More here:
See my rebuttal to Tuur's rebuttal :)
24/ My third point addresses PoS’ promise of perpetual income to ETHizens. Vitalik is no stranger to embracing free lunch ideas, e.g. during his 2014 ETH announcement speech, where he described a coin with a 20% inflation tax as having “no cost” to users.
Yeah, I haven't really emphasized perpetual income to stakers as a selling point in years. I actually favor rewards being as low as possible while still being high enough for security.
25/ In his response to my tweet, Vitalik adopted my format to “play the same game” in criticizing Bitcoin. My criticisms weren't addressed, and his response was riddled with errors. Yet his followers gave it +1,000 upvotes!
Vitalik Non-giver of Ether @VitalikButerin: - What works today (L1) is just a phase, ideal and unproven future (usable L2) is to come - Utopian concept of progress: we're already so confident we're finished we ain't needin no hard forks…
Ok, let's hear about what the errors are...
26/ Rebuttal: - BTC layer 1 is not “just a phase”, it always will be its definitive bedrock for transaction settlement. - Soft forking digital protocols has been the norm for over 3 decades—hard-forks are the deviation! - Satoshi never suggested hyperbitcoinization as a goal.
Sure, but (i) the use of layer 1 for consumer payments is definitely, in bitcoin ideology, "just a phase", (ii) I don't think you can make analogies between consensus protocols and other kinds of protocols, and between soft forking consensus protocols and protocol changes in other protocols, that easily, (iii) plenty of people do believe that hyperbitcoinization as a goal. Oh by the way:
27/ This kind of sophistry is exhausting and completely counter-productive, but it can be very convincing for an uninformed retail public.
Ok, go on.
28/ Let me share a few more inconvenient truths.
29/ In order to “guarantee” the transition to PoS’ utopia of perpetual income (staking coins earns interest), a “difficulty bomb” was embedded in the protocol, which supposedly would force miners to accept the transition.
The intended goal of the difficulty bomb was to prevent the protocol from ossifying, by ensuring that it has to hard fork eventually to reset the difficulty bomb, at which point the status quo bias in favor of not changing other protocol rules at the same time would be weaker. Though forcing a switch to PoS was definitely a key goal.
30/ Of course, nothing came of this, because anything in the ETH protocol can be hard-forked away. Another broken promise.
Tuur Demeester @TuurDemeester: Looks like another Ethereum hard-fork is going to remove the "Ice Age" (difficulty increase meant to incentivize transition to PoS).
How is that a broken promise? There was no social contract to only replace the difficulty-bombed protocol with a PoS chain.
31/ Another idea that was marketed heavily early on, was that with ETH you could program smart contract as easily as javascript applications.
Tuur Demeester @TuurDemeester: I forgot, but in 2014 Ethereum was quite literally described as "Javascript-on-the-blockchain"
Agree that was over-optimistic, though the part of the metaphor that's problematic is the "be done with complex apps in a couple hours" part, NOT the "general-purpose languages are great" part.
32/ This was criticized by P2P & OS developers as a reckless notion, given that every smart contracts is actually a “de novo cryptographic protocol”. In other words, it’s playing with fire.
See above
33/ The modular approach to Bitcoin seems to be much better at compartmentalizing risk, and thus reducing attack surfaces. I’ve written about modular scaling here...
To be fair, risk is reduced because Bitcoin does less.
34/ Another huge issue that Ethereum has is with scaling. By putting “everything on the blockchain” (which stores everything forever) and dubbing it “the world computer”, you are going to end up with a very slow and clogged up system.
Christopher Allen @ChristopherA: AWS cost: $0.000000066 for calc, Ethereum: $26.55. This is about 400 million times as expensive. World computer?
We never advocated "putting everything on the blockchain". The phrase "world computer" was never meant to be interpreted as "everyone's personal desktop", but rather as a common platform specifically for the parts of applications that require consensus on shared state. As evidence of this, notice how Whisper and Swarm were part of the vision as complements to Ethereum right from the start.
35/ By now the Ethereum bloat is so bad that cheaply running an individual node is practically impossible for a lay person. ETH developers are also imploring people to not deploy more smart contract apps on its blockchain.
Tuur Demeester @TuurDemeester: But... deploying d-apps on the "Ethereum Virtual Machine" is exactly what everyone was encouraged to do for the past 4 years. Looks like on-chain scaling wasn't such a great idea after all.
Umm.... I just spun up a node from scratch last week. On a consumer laptop.
36/ As a result, and despite the claims that running a node in “warp” mode is easy and as good as a full node, Ethereum is becoming increasingly centralized.
@TuurDemeester Finally a media article touching on the elephant in the room: Ethereum has become highly centralized. #infura
See above
37/ Another hollow claim: in 2016, Ethereum was promoted as being censorship resistant…
Tuur Demeester @TuurDemeester: Pre TheDAO #Ethereum presentation: "uncensorable, code is law, bottom up".
Yes, the DAO fork did violate the notion of absolute immutability. However, the "forking the DAO will lead to doom and gloom" crowd was very wrong in one key way: it did NOT work as a precedent justifying all sorts of further state interventions. The community clearly drew a line in the sand by firmly rejecting EIP 867, and EIP 999 seems to now also be going nowhere. So it seems like there's some evidence that the social contract of "moderately but not infinitely strong immutability" actually can be stable.
38/ Yet later that year, after only 6% of ETH holders had cast a vote, ETH core devs decided to endorse a hard-fork that clawed back the funds from a smart contract that held 4.5% of all ETH in circulation. More here: ...
See above
39/ Other potential signs of centralization: Vitalik Buterin signing a deal with a Russian government institution, and ETH core developers experimenting with semi-closed meetings: …,
Hudson Jameson @hudsonjameson: The "semi-closed" Ethereum 1.x meeting from last Friday was an experiment. The All Core Dev meeting this Friday will be recorded as usual.
Suppose I were to tomorrow sign up to work directly for Kim Jong Un. What concretely would happen to the Ethereum protocol? I suspect very little; I am mostly involved in the Serenity work, and the other researchers have proven very capable of both pushing the spec forward even without me and catching any mistakes with my work. So I don't think any argument involving me applies. And we ended up deciding not to do more semi-closed meetings.
40/ Another red flag to me is the apparent lack of relevant expertise in the ETH development community. (Check the responses…)
Tuur Demeester @TuurDemeester: Often heard: "but Ethereum also has world class engineers working on the protocol". Please name names and relevant pedigree so I can follow and learn.
I personally am confident in the talents of our core researchers, and our community of academic partners. Most recently the latter group includes people from Starkware, Stanford CBR, IC3, and other groups.
41/ For a while, Microsoft veteran Lucius Meredith was mentioned as playing an important role in ETH scaling, but now he is likely distracted by the failure of his ETH scaling company RChain.
I have no idea who described Lucius Meredith's work as being important for the Serenity roadmap.... oh and by the way, RChain is NOT an "Ethereum scaling company"
42/ Perhaps the recently added Gandalf of Ethereum, with his “Fellowship of Ethereum Magicians” [sic] can save the day, but imo that seems unlikely...
Honestly, I don't see why Ethereum Gandalf needs to save the day, because I don't see what is in danger and needs to be saved...
43/ This is becoming a long tweetstorm, so let’s wrap up with a few closing comments.
44/ Do I have a conflict of interest? ETH is a publicly available asset with no real barriers to entry, so I could easily get a stake. Also, having met Vitalik & other ETH founders several times in 2013-’14, it would have been doable for me to become part of the in-crowd.
Agree there. And BTW I generally think financial conflicts of interest are somewhat overrated; social conflicts/tribal biases are the bigger problem much of the time. Though those two kinds of misalignments do frequently overlap and reinforce each other so they're difficult to fully disentangle.
45/ Actually, I was initially excited about Ethereum’s smart contract work - this was before one of its many pivots.
Tuur Demeester @TuurDemeester: Ethereum is probably the first programming language I will teach myself - who wouldn't want the ability to program smart BTC contracts?
Ethereum was never about "smart BTC contracts"..... even "Ethereum as a Mastercoin-style meta-protocol" was intended to be built on top of Primecoin.
46/ Also, I have done my share of soul searching about whether I could be suffering from survivor’s bias.
@TuurDemeester I just published “I’m not worried about Bitcoin Unlimited, but I am losing sleep over Ethereum”
Ok, good.
47/ Here’s why Ethereum is dubious to me: rather than creating an open source project & testnet to work on these interesting computer science problems, its founders instead did a securities offering, involving many thousands of clueless retail investors.
What do you mean "instead of"? We did create an open source project and testnet! Whether or not ETH is a security is a legal question; seems like SEC people agree it's not:
48/ Investing in the Ethereum ICO was akin to buying shares in a startup that had “invent time travel” as part of its business plan. Imo it was a reckless security offering, and it set the tone for the terrible capital misallocation of the 2017 ICO boom.
Nothing in the ethereum roadmap requires time-travel-like technical advancements or anything remotely close to that. Proof: we basically have all the fundamental technical advancements we need at this point.
49/ In my view, Ethereum is the Yahoo of our day - an unscalable “blue chip” cryptocurrency:
Tuur Demeester @TuurDemeester: 1/ The DotCom bubble shows that the market isn't very good at valuing early stage technology. I'll use Google vs. Yahoo to illustrate.
Got it.
50/ I’ll close with a few words from Gregory Maxwell from 2016,:
See my rebuttal to Greg from 2 years ago:
submitted by shouldbdan to ethtrader [link] [comments]

Long live decentralized bitcoin(!) A reading list

Newbs might not know this, but bitcoin recently came out of an intense internal drama. Between July 2015 and August 2017 bitcoin was attacked by external forces who were hoping to destroy the very properties that made bitcoin valuable in the first place. This culminated in the creation of segwit and the UASF (user activated soft fork) movement. The UASF was successful, segwit was added to bitcoin and with that the anti-decentralization side left bitcoin altogether and created their own altcoin called bcash. Bitcoin's price was $2500, soon after segwit was activated the price doubled to $5000 and continued rising until a top of $20000 before correcting to where we are today.
During this drama, I took time away from writing open source code to help educate and argue on reddit, twitter and other social media. I came up with a reading list for quickly copypasting things. It may be interesting today for newbs or anyone who wants a history lesson on what exactly happened during those two years when bitcoin's very existence as a decentralized low-trust currency was questioned. Now the fight has essentially been won, I try not to comment on reddit that much anymore. There's nothing left to do except wait for Lightning and similar tech to become mature (or better yet, help code it and test it)
In this thread you can learn about block sizes, latency, decentralization, segwit, ASICBOOST, lightning network and all the other issues that were debated endlessly for over two years. So when someone tries to get you to invest in bcash, remind them of the time they supported Bitcoin Unlimited.
For more threads like this see UASF

Summary / The fundamental tradeoff

A trip to the moon requires a rocket with multiple stages by gmaxwell (must read)
Bram Cohen, creator of bittorrent, argues against a hard fork to a larger block size
gmaxwell's summary of the debate
Core devs please explain your vision (see luke's post which also argues that blocks are already too big)
Mod of btc speaking against a hard fork
It's becoming clear to me that a lot of people don't understand how fragile bitcoin is
Blockchain space must be costly, it can never be free
Charlie Lee with a nice analogy about the fundamental tradeoff
gmaxwell on the tradeoffs
jratcliff on the layering

Scaling on-chain will destroy bitcoin's decentralization

Peter Todd: How a floating blocksize limit inevitably leads towards centralization [Feb 2013] mailing list with discussion on reddit in Aug 2015
Nick Szabo's blog post on what makes bitcoin so special
There is academic research showing that even small (2MB) increases to the blocksize results in drastic node dropoff counts due to the non-linear increase of RAM needed.
Reddit summary of above link. In this table, you can see it estimates a 40% drop immediately in node count with a 2MB upgrade and a 50% over 6 months. At 4mb, it becomes 75% immediately and 80% over 6 months. At 8, it becomes 90% and 95%.
Larger block sizes make centralization pressures worse (mathematical)
Talk at scalingbitcoin montreal, initial blockchain synchronization puts serious constraints on any increase in the block size with transcript
Bitcoin's P2P Network: The Soft Underbelly of Bitcoin someone's notes: reddit discussion
In adversarial environments blockchains dont scale
Why miners will not voluntarily individually produce smaller blocks
Hal Finney: bitcoin's blockchain can only be a settlement layer (mostly interesting because it's hal finney and its in 2010)
petertodd's 2013 video explaining this
luke-jr's summary
Another jratcliff thread

Full blocks are not a disaster

Blocks must be always full, there must always be a backlog
Same as above, the mining gap means there must always be a backlog talk: transcript:
Backlogs arent that bad
Examples where scarce block space causes people to use precious resources more efficiently
Full blocks are fine
High miner fees imply a sustainable future for bitcoin
gmaxwell on why full blocks are good
The whole idea of the mempool being "filled" is wrong headed. The mempool doesn't "clog" or get stuck, or anything like that.


What is segwit

luke-jr's longer summary
Charlie Shrem's on upgrading to segwit
Original segwit talk at scalingbitcoin hong kong + transcript
Segwit is not too complex
Segwit does not make it possible for miners to steal coins, contrary to what some people say
Segwit is required for a useful lightning network It's now known that without a malleability fix useful indefinite channels are not really possible.
Clearing up SegWit Lies and Myths:
Segwit is bigger blocks
Typical usage results in segwit allowing capacity equivalent to 2mb blocks

Why is segwit being blocked

Jihan Wu (head of largest bitcoin mining group) is blocking segwit because of perceived loss of income
Witness discount creates aligned incentives