Linzhi ASICs
Jan 15 · 7 min read

It’s The Unknown We Fear: Response to MoneroCrusher

Once in a while when reading articles, you come across the one you feel is different. A like-minded soul searching for something? Or maybe not? You are willing to investigate…

ASIC resistance is a state of mind, not something you can achieve with a final software implementation.

MC: There will always be a specialized device that is going to beat the big decentralized GPU mining community.

Yes.

MC: The more efficient machine will always replace the less efficient one.

Yes.

MC: The current ASIC market weakens the security, integrity and decentralization of PoW based cryptocurrencies.

We believe the opposite. Our customers agree and never brought up this subject. You say “current” market, there is hope to change what bothers you. The market is a result of our work.

MC: All GPUs are an assembly of ASICs themselves.

Yes.

MC: The EF & the Ethereum Core Developers risk tainting their image permanently.

This should be our top concern, second being environmental impact.

MC: This is literally a billion dollar business.

It is, and unlike free software source code, business considerations are by default (though not necessarily so) intransparent. The current intransparency of this business is where we believe is a chance for improvement. Not just on the software side — transparency on the software side is pretty good.

MC: PoW & ASIC-resistance: If you want it, you have to fight for it — play the game!

Again?

MC: General purpose. Adapting to various chaotic and changing circumstances with little effort.

All hardware is the same. New ASICs are better at adapting to chaos than GPUs, because GPUs are stuck technically and economically in a fixed architecture. This is a key point from our view.

MC: Monero hasn’t seen a large ASIC takeover of the network since its first anti-ASIC fork.

You probably mean a “non-Nvidia, non-AMD” takeover. That seems to not have happened yet.

MC: ASICs on the Monero network.

You mean non-Nvidia, non-AMD.

MC: … they have just become smarter about it and made it less obvious

Sounds like a sad story.

MC: Ethash has proven itself to be extremely good at keeping ASIC efficiency gains pretty close to a GPU’s efficiency.

It wasn’t worth yet to make better ASICs. The Bitmain, Innosilicon etc. ASICs are already beating GPUs easily. If you already win the race, you don’t need to spend more money to make your race car faster.

MC: Efficiency gains much smaller with Ethash.

Efficiency gains are more expensive with Ethash. Cost is the only difference.

MC: Takeover of the network would cost hundreds of millions in funds

We think it’s much less, but our customers take that risk. We see strong demand for PoW asics, of any algorithm, from small home miners to mega farms (100MW+).

MC: Risk is immensely increased for ASIC manufacturers and their investors.

Risk is increased for our customers, not us. Unless you believe you can destroy PoW altogether, then the risk is for us as well and we need to focus on non-PoW chips.

MC: Carbon footprint will also be kept lower.

We should talk about this more, it should be one of our highest organising principles. The original ASIC-Resistance thoughts from the 2014 Yellow Paper ought to be brought forward to reflect what we learned in the last 5 years. ASICs use less energy. You don’t resist ASICs because they are less energy efficient? We think the reason for ASIC-Resistance was that people thought they lead to centralisation. We think that’s a fundamental misunderstanding: The cause of centralisation is cost. ASICs reduce the energy requirement of the network prior to PoS.

MC: Say currently set 1 (Ethash) consumes 1 GW of power, set 2 (ProgPOW) would increase it to 1.3 GW of power.

If your numbers are true, that alone should make us reject ProgPOW. What world are we in in 2019 that we can afford to waste 300 MW of constant load for …. for what? Can a large farm help the community with solid numbers? This is important!

MC: solution … fork … unpredictably every 6 months

Will reward secrecy.

MC: GPU miners don’t have to worry about anything.

It was wrong in the past, it stays wrong in the future. GPUs are a good tool to start a network, and for development. For large systems, GPUs can’t maintain their cost/efficiency/performance advantage.

MC: Equivalence of 30% of the maximum output of the Clinton Nuclear Power plant.

We think 75% of mining uses green energy, and may actually be consuming excess capacity on the grid, and even supporting the profitability and as a consequence the development of renewable energy generation capacity. It may be net beneficial for the renewables space, but more research is needed in this area.

MC: Opposing arguments to this could be: “the more energy is expended, the more secure the network becomes”.

We believe security comes from the asymmetry between hash and verification. Hashing is hard, and thermodynamically provable hard, while verification is easy. Ease-of-verification is an important decentralising force, unfortunately also hard to quantify. The argument of expending more energy to make the network more secure must be rejected.

MC: If you want Proof of Work to work in an originally intended decentralized fashion.

We also believe efficient decentralization is the main goal, and free competition between asicmakers (including GPU) is the best way to reach it.

MC: You’ve got to put work in (pun intended).

We always loved the work part in Proof-of-Work. Hardware is hard.

MC: AMD always allowed VBIOS modding on their cards

Yes.

MC: Nvidia cards have much better alternatives if mining dies down: > reselling or repurposing for HPC/DL/ML.

Yes. Important to think about these things. Waste reduction is important. Frivolous bricking of miners is littering.

MC: What ProgPow factually does … They want to have the benefits of low risk and at the same time the rewards of high risk.

Common sense tells us there are probably reasons for that.

MC: Mining centralization and ASICs will ultimately cripple the decision-making process of the currently independent Ethereum Foundation and ETH Core Developers.

Our proposal is:

  1. Actively govern the competition of multiple asicmakers (including GPUs) with each other.
  2. Announce PoW algorithm changes a long time in advance (we proposed 2 years), to remove most incentives for special interests to influence the EF/devs. The longer you announce in advance, the less incentive there is. The shorter you announce in advance, the more you reward the wrong thing.

Just because we are asicmakers doesn’t mean all of our ideas are bad. We see some problems earlier than others, and some later than others. That’s where dialogue starts. Maybe asicmakers do not have to be bad and can be governed just like anything else.

MC: Signal to community that ASIC resistance is a goal and that it is a current problem.

ASIC resistance is the problem. The solution is ASIC friendliness.

MC: If evidence arrives of massive ASIC takeover.

You are calling for asicmakers to self-mine and do secret deals with large farms.

MC: That’s the ultimate step towards psychological warfare on ASIC manufacturers

We are already here.

MC: and their respective investors

The asic investors care about whether there are enough paying customers. asicmakers care about the future of PoW.

MC: Only here to suck out every last drop of everything that makes cryptocurrency such an amazing concept.

We think we are equal.

MC: (that being mainly decentralization).

ASICs don’t lead to centralization, cost advantage does.

MC: ASIC manufacturers are not expected, nor incentivized to truly change their practices anytime soon, if ever.

Step up and demand change. We are in Shenzhen, we can translate all documents bilingual. We can create global rules, communicated in English and Chinese, for a healthy PoW system, with focus on good governance, environmental impact, waste reduction. In the case of Ethereum, maybe only until PoS arrives. In the case of others, maybe longer. But the “it’s only for a little while until PoS arrives” argument shouldn’t mean that we do a lousy job until that day. We believe the following list is a good first set of guidelines for PoW asicmakers:

  1. cost transparency
  2. production transparency
  3. sell to developers and community first
  4. volume discounts topped out at 20%
  5. recycling, protection against loss of value
  6. self-mining only when unable to sell and fully transparent
  7. transparency in developer relations and funding
  8. qualify customers, no sales to ponzi, MLM, etc.

MoneroCrusher and Linzhi can work together on this. We have to overcome stereotypes, but today can be the day we start. If someone tells us it can’t be done, we know it can. We are ready.

MC: brick current fixed function hardware aka ASICs

Bricking miners is littering.

MC: minus the ASIC manufacturers, but nobody cares about those

Why not?

MC: they knew what they had signed up for when they secretly created ASICs

We focus on our customers. As long as they are happy, we are happy.

MC: A project that states an anti-ASIC stance in its founding document, the Whitepaper.

We believe it’s unfair to hold authors of a document written in 2013/2014 accountable for every last word in every paragraph forever, and in a fast-changing high-tech environment in particular! The development and realizations of the past 5 years could not have been predicted by anyone. Never stop learning.

MC: Chaos and ASICs are like oil and water — they just don’t mix.

We love Approximate Computing.


Never stop learning. Linzhi Shenzhen Telegram: https://t.me/LinzhiCorp