Thread by @JeremyRubin

Responding to @VitalikButerin's response to my piece in @TechCrunch -- thanks for taking the time to thoughtfully respond!
My Piece: techcrunch.com/2018/09/02/the…
Vitalik's Response: reddit.com/r/ethtrader/co…

First, a summary of @VitalikButerin's response: my article is mostly correct, but Vitalik plans to make Eth required for the network to function via rents or via min burned fees. HD-PoS not doable because consensus pricing is hard.

@VitalikButerin thinks economic abstraction can work for end users if min fees can be made required to be paid by the miner's block, not the individual txn.

I agree with @VitalikButerin to a point. Let's go over where we agree first:
- I agree that wrappers work to emulate abstraction from an end-user's perspective (mentioned in my article)
- I agree that the minfee/burn and rent mechanisms have a component of upwards price pressure

That's about where my agreement ends. First, let's take a look at @VitalikButerin's specific proposals and then we talk more broadly.

Contract issuers care about two things: security and expense. Every change made should be audited against these two goals. Other factors, like ease of use, are important but fit in to the other two categories.

Therefore we should ask:
Is the system made more secure after the change?
Is the system made less expensive after the change?

Often times, these are at odds. Security can cost $$$.

Making Eth required for Ethereum can't be automatically considered in the context of the argument I make assuming Ethereum succeeds, because we need to consider that this may cause the system to decline in adoption.

Enshrining Eth seems to improve neither security nor expense. In fact, it seems to make these things worse. Here's why:

Making Eth required for transactions makes transactions inherently more expensive than they should be as it biases pricing upwards. This is anti-user, but perhaps good for the price of Eth in short term.

Immediately, requiring miners to have large chunks of Eth decreases mining decentralization (if contract users are abstracted to use whatever they want). Under PoS, making users burn their Eth will increase staking centralization.

Without mining, using non-HD PoS, Eth dependence imposes a new constraint -- now it's hard for a contract to deal in value greater than that of the Eth market cap.

If assets on Ethereum grows in excess of Eth (let's say 100x Eth), then PoS is relatively insecure because one could use 1% of that asset to purchase enough stake to roll back txns in those assets.

This brings us back to expense -- the solution is that either you need HD-PoS, or Eth will have to be worth more than all the assets managed by the network -- unlikely, and too expensive.

Even if the last claim doesn't fully check out (don't have a model for the exact ratio of value Eth would have to be to ensure security), inarguably you get higher quality PoS the more at stake (why have staking incentives otherwise)

If HD-PoS cannot be done, as @VitalikButerin asserts, then this is bad news for Eth. Most contracts would likely de-platform from Eth in favor of systems that don't rely on Eth economics for security and siphon off value.

Notably, Bitcoin and Stellar both have this property. I mention them because I'm familiar with them -- Ethereum could adopt parts of their design if they are good. Usual conflict of interest disclaimers apply.

With Bitcoin, security is paid for via an 'energetic one-way function'. Fees could be paid in anything, even out of band, but the security of the network depends on the energy not on the price of BTC (when subsidies end).

Notably, Bitcoin Core implements the prioritisetransaction API which allows a node operator to artificially fee bump transactions according to the node operator's preference or out of band payments.

Stellar consensus is reputational (non-economic). You pick a circuit of validators you like, and you maintain consensus with it. If XLM price goes to zero, the system doesn't fall apart because there isn't mining/staking rewards.

Unfortunately (in my opinion), Stellar does have minimum fees required to be in Lumens. I'm looking to change this though! It's less problematic in Stellar because of the native DEX for pricing and we only need fees for anti-DoS.

So for both Bitcoin and Stellar, the integrity of the chain is independent of the asset price. If BTC and XLM go to zero, the chains could still operate.

For example, the recent decreasing of the block reward & delay of the difficulty bombs...

If these methods are used, it's like a pilot pushing the eject button. It guarantees that the ship/network goes down, while short term providing a slower fall/price support for the pilot/big holders. Feels immoral.

At a higher level, the philosophy of forcing users to value Eth runs counter to libertarian ideals. Economic abstraction is a free market. Free markets are good -- cryptocurrency itself, is a free market for money.

This also explains why a Tax can be effective in price support: you have a coerced buyer! You can get paid in Eth, but you have to pay income tax in dollars. But tax is not voluntary. If you don't pay tax, you go to jail. Bad things happen

The state has a monopoly on violence, and they will use that violence (with a measure of restraint) if you don't pay your taxes. (Note I'm using taxes in the abstract, to include compelled monetary conventions like petrodollar)

If I choose not to cough up a fee in Eth, what happens? What should happen? If we are trying to build a freer more open society, we should carefully avoid attempts to use coercive measures.

Without a state backing Eth, it's silly for Ethereum to try to emplace Eth as a required fee paying mechanism. There are other ways to be valuable.

I hope that @VitalikButerin thinks about the direction for the Ethereum network carefully -- I know there are a lot of users and investors depending on him.