Review: Finding Signal in a Noisy World - Data Always

Jeff Booth recently challenge another guest on a podcast to dissect his article Finding Signal in a Noisy World and find flaws.

TLDR;

Jeff Booth’s attempt to disentangle bitcoin from crypto disappoints, delivering more noise than signal. Rather than make a compelling case for bitcoin beyond altcoins, the piece glosses over the intrinsic shortcomings found in modern Bitcoin and rehashes rudimentary Bitcoin talking points that demonstrate Booth’s mere surface level understanding of the wider cryptoasset environment and his lack of game theoretical intuition.

The absence of real-world data and analysis in the article is a stark, yet common, contrast to the ideals that Booth preaches. It demonstrates once again that although Bitcoin thought leaders emphasize that their followers should verify facts rather than trusting others, they still expect their followers to follow them blindly—providing little to no supporting analysis to justify their ideas.

To his credit, Jeff Booth is directionally correct in most areas:

The public discourse around proof-of-work is broken.
The world and financial system have seen dramatic rises in misinformation.
The most valuable blockchains are those that prioritize security and decentralization.

Here are my positions on those topics.

I believe that calls to abolish or outlaw proof-of-work are ridiculous, not only for their futility, but also because the mainstream media has drastically exaggerated the environmental impact of Bitcoin mining. I’ve gone as far as open-sourcing analysis that demonstrates that the Cambridge Bitcoin Electricity Consumption Index is, and will continue to be, an overestimate. Despite its trade-offs, proof-of-work is a good consensus mechanism.

Independent of political and technological slant, it would be nearly impossible to find someone who disputes that we have seen a rise in misinformation over the past decade. The cryptoasset space is no different, and although misinformation seems to inversely correlate with market cap—many smaller tokens are almost entirely misinformation, Bitcoin is not immune and often falls prey to false narratives. For my part, I attempt to minimize misinformation by focusing more heavily on data than narrative, however the two can never be fully separated. Responsive commentary, this writing for example, will naturally slant less data-oriented than I prefer my work to be. I do not like to write this way.

Security and decentralization are particular optimizations for a blockchain, and in my opinion the most important and the most interesting. There may be value in fast centralized chains for niche use cases, but as the foundation for either global money or the base layer of technological innovation, security and decentralization are fundamental requirements.

Unfortunately, Booth’s description of why Bitcoin is the way forward is fraught with misunderstanding.

Security and Decentralization

Booth’s writing hinges on the assumption that the Bitcoin Network is unquestionably secure, however this assumption not rooted in data or logic. Rather than provide a dependable voice in the debate around the future of Bitcoin’s security model or discuss the intricacies of the actual dynamics behind network security, Booth explains bitcoin mining as if he were trying to placate a toddler that asked how digital money is made. He quotes terra [sic] hash numbers knowing that no reader, nor the author, has a true understanding of what 185 million terahashes per second means in practice. Hash rate metrics are a distraction and a deception: a measure designed to increase perpetually and give the appearance of heightening security.

The security of the bitcoin network is not measured in hash rate!

There are two metrics that determine how protected the network is: the amount of electricity consumed by miners and the number of modern ASICs required to gain a majority of network hash rate. Hash rate statistics are simplistic approximations to these metrics; they do not account for the increasing efficiency of ASICs or varying energy costs. Hash rate is only a valid proxy if you are defending against an adversary with static technology.

Looking at the electricity spend that can be supported by the current mining dynamic, we see a healthy climb, followed by some recent stagnation. Despite the rolling-annual value being about to fall significantly below the rolling-cycle average for the first time ever, there is nothing to panic over yet. Levels remain sufficiently healthy, although we will continue to see the capitulation of mining companies unless price rebounds in the near-term. This is fine.

However, concern is warranted when we break mining revenue into its corresponding parts. The stagnation in blockspace demand is very heavy, and the extreme volatility within cycles calls into question the stability of the mechanism as the block subsidy continues to be reduced.

At current levels, transactions could support a single terawatt hour of mining per year. The cycle peak for usage was approximately 140 terawatt hours per year, so to simply grow demand to a point that could support the current crop of miners would be an incredible feat.

Not unlike absolute hash rate statistics, electricity metrics are sufficiently abstract to be confusing, so although it is a less accurate proxy, that in a rising energy cost environment will make the network look more sustainable than it actually is, if we denominate fees in dollars, total Bitcoin transaction fees in 2022 will be their lowest full-year value since 2016.

The regression in demand for blockspace, despite massive mainstream growth, seriously calls into question the long-term stability of the security model. Without rising transaction fees, to maintain the level of support for miners in the face of block subsidy cuts, the price of bitcoin is required to rise by 19% per year denominated in energy.

This is not indicative of unquestionable security. It is too early to propose changes to the security model, but assuming its indefinite security is irresponsible.

Jeff Booth’s understanding of the rollover of mining devices is naïve and disconnected from how the dynamic plays out in practice. Booth claims that:

the process of paying for additional equipment which is then obsoleted over time as new equipment becomes superior, is costly. This has the effect of supporting new entrants/ideas in the market. In other words, its very nature reduces the monopolistic tendencies of a market to consolidate around a few large miners and price others out.

In practice, the inverse is actually true. The constant need to replace old machines with newer machines heavily favors large centralized mining operations. Large industrial miners have more negotiating power with suppliers, earn higher profits due to economies of scale and preferential power purchase agreements, and as businesses better manage their fleets of mining devices. As their machines age, many large miners dump their older generation equipment onto the market, bought up by smaller operations and home miners (lacking purchase negotiating power this is often their best choice), who then struggle to break even on their investments.

The economies of scale described above are centralizing forces that lead to the most likely short-term path of capture for Bitcoin. Hash rate controlled by publicly traded miners has grown extremely aggressively over the past three years—both by going public and through expanding their operations, and with their ease of access to cheap money, relative to private and retail miners, this trend is set to continue. These businesses expose the Bitcoin Network to risks of regulatory capture, as their primary requirement is to abide by US laws. When a Bitcoin unfriendly administration decides to take action, if public hash rate continues to trend up and breaches 51%, the administration will be able to attempt to force miners to build only KYC blocks, and only on-top of KYC blocks. This effective cartel of miners will form the longest, and therefore, according to Booth, the most trusted, chain.

Booth fearmongers about the introduction of oracles, but this regulatory capture dynamic is a stark contrast to similar threats in a proof-of-stake system. Staking services tend to only be a fraction of the business of centralized custodians and therefore these custodians are able to cut off a limb and continue to operate; Coinbase for example has committed to shutting down and exiting their Ethereum validator business if forced to comply with censoring regulations (even if withdrawals are not yet enabled).

Bitcoin miners have no second business income to fall back on, therefore their choices would be either full liquidation at firesale prices (while facing lawsuits for breach of fiduciary duty) or censorship.

This is a terrifying timeline that is becoming more likely every day.

Booth speaks with an air of confidence on security, suggesting that the Chinese Bitcoin mining ban proved the network’s resilience to “nation-state attack[s]”, but this is twisting reality. China did not attack the Bitcoin Network, they attacked their own Bitcoin miners. An attack on the network would have been a nationalization of hash rate, but it should be clear from the past year that China’s goals were to oppress their own population and to lighten the load on their energy infrastructure.

The current security of the Bitcoin Network is nowhere near sufficient to repel an attack from a motivated large economy. The US Military budget is approximately $800 billion per year, or approximately 100x the annual electricity spend of the network. Allocating only a couple percent of their operational military budget, the United States could easily cripple the system. Bitcoin is a fly on the wall, currently not worth the energy to swat it in comparison.

Scaling

Booth presents the following chart as showing “lightning adoption since inception”.

I believe that Jeff Booth is a good actor, so I’m going to give him the benefit of the doubt, but capacity and usage are very different metrics.

There is a narrative in which these two metrics correlate: increasing usage of the Lightning Network drives fees higher, making it more profitable to run Lightning nodes, thus spurring higher capacity, but this narrative is only the dominant factor in a vacuum. The reality is less endearing to the adoption narrative: spikes in the capacity of the Lightning Network correlate far more heavily with decreasing yield on Bitcoin lending platforms. As an alternative, although poor, source of yield, the majority of these capacity jumps that were touted as adoption by the community were likely reallocation of funds from platforms like BlockFi as yields fell the past two years.

I like Bitcoin and I want the Lightning Network to succeed, but ignoring its non-utilization and presenting disingenuous metrics is not going to magically solve any problems. Lightning usage statistics are relatively obscure, but the data points we have are not promising. Sergej Kotliar, the CEO of Bitrefill and one of the biggest advocates for the Lightning Network, gave this talk worth watching at Pizza Day Prague 2022.

Eric Wall provides a succinct summary in this Twitter thread, with the key points being:

Historical Lightning Network adoption expectations were far too optimistic.
Despite largely being a failed and pointless cryptoasset, Litecoin does more volume than the Lightning Network even though their ecosystem is Bitcoin first.
There is a large cultural divide between modern Bitcoin philosophy and Bitcoin the tool.
The data is largely irrefutable that a Bitcoin-only world is a fantasy. The people who still believe it to be a possible outcome are data illiterate.

Closing Remarks

Early in his writing, Jeff Booth invokes Charlie Munger’s famous quote:

show me an incentive and I will show you the outcome;

suggesting that altcoins are “perverse … schemes” to enrich insiders, with Bitcoin being comparatively pure. Although I do agree in part with this characterization, Bitcoin maximalism is far from clean. Satoshi was not against altcoins, predicting their natural rise, for their goal was to create a decentralized and censorship resistant p2p cash system.

Most modern Bitcoin maximalists are investors, not cypherpunks. The effect has been to shift the importance from censorship resistant money to profiteering—using techniques like pandering to the SEC to regulate other cryptoassets or embracing the concept of the US Military mining Bitcoin. True cypherpunks don’t care which consensus mechanism wins in the end; that is a trait of the modern investor class who are primarily concerned about the value of their holdings.

Bitcoin has fallen, what remains is an empty shell, and frankly it’s embarrassing.