Blockchain is fascinating and flawed

Like many, during last year's run up of Bitcoin I became motivated to get a better understanding of the underlying blockchain technology. Based on the hype, and as the word "blockchain" began finding its way into every startup and Wired article, I expected to gain an appreciation for something that was truly going to be game changing.

At its core, blockchain (and Bitcoin in particular) advances some clever ideas about how to decentralize things that require trust - like a ledger of financial transactions (or health records, or edits to a document, etc). The system protects the integrity of the ledger by making it mathematically improbable that a bad actor could add a fraudulent transaction to the ledger before one of the "good guys". The idea that a financial system could trust its users to collaboratively maintain its records sounds crazy, and the innovation necessary to make it possible is impressive.

The lottery

But that's where the impressiveness ends for me. From there you quickly realize that the "Proof of Work" approach that Bitcoin employs means that the integrity is supported by an irresponsible use of electric energy. Essentially, in order to prevent "bad guys" from adding fraudulent entries into the ledger, blockchain relies on a lottery of sorts. Each addition to the ledger requires solving a processor demanding math problem. As long as more good guys are playing the lottery than bad guys, the good guys are statistically guaranteed to solve the problem first and win the lottery. Except instead of buying a lottery ticket, we turn loose expensive machines on computationally expensive math problems that require massive amounts of electricity. "Wait", you say, "surely computing those math problems provides an ancillary benefit for - say - cancer research or SETI?". Nope, they are just computationally intensive and environmentally reckless. Bitcoin alone is estimated to use the same amount of energy as the country of Iraq, and .25% of total worldwide energy usage. As Moore's law continues, the math problems have to get harder to avoid making it too easy for the bad guys. So the usage won't decrease over time.

Security Issues

The "Militia" problem

Another way to think about Proof of Work blockchains like bitcoin is to imagine your local bank used a militia to protect its vault. Every time a threat arises, the militia is called and the integrity of the bank is based on the number and responsiveness of the "good guys"showing up to the parking lot before the bad guy(s) can break in. This works great against a singular attack and as long as the militia is strong. But if enthusiasm wanes or the opponents launch coordinated attacks the balance could easily swing the other way. If enough bad actors overwhelm the good, in what is known as a "51% attack", they can steal coins by rewriting the ledger however they wish. Ethereum, arguably the second most popular blockchain coin to Bitcoin, was victimized with such an attack in 2019, to the tune of $1.1M.

With blockchain, the good actors are often motivated to participate because of rewards they get for doing so. In Bitcoin, this is the mining model where "good guys" get coins for helping to add to the ledger and this reward motivates the ratio of good to bad in a way that preserves the militia. The question is, does this notion scale when we try to project use of Proof of Work blockchain to other use cases and industries? Can we keep enough good actors motivated to maintain the mathematical advantage? Can we afford to have more databases that require 1/4 of percent of the earth's entire energy output to do what a well secured database at a traditional bank can do?

The "Weakest Link" Problem

While the technology allows laymen to transact directly with the blockchain, for most luddites the steps required to do so are outlandish. So most people who buy or sell bitcoin do so through a nice exchange like Coinbase, Kraken etc. These entities making transacting with bitcoin more like using your online bank, which is great from a user experience standpoint but introduces an important security vulnerability. In this scenario, the vulnerability level of the exchange is as important as the vulnerability level of the blockchain. Infamous hacks at major exchanges like Mt. Gox, where $350 million USD were stolen, show you how the security of the blockchain is only as good as the weakest link - your exchange.

Scalability Issues

The "Excel" Problem

Maybe you're old enough to remember before the days of Google Sheets or Microsoft Office 365, sending excel files around your office with appended version numbers. Asking colleagues if they had the "latest version" and shipping increasingly larger documents around was a constant hassle.

Blockchain's decentralized database is very reminiscent of this approach. Instead of your bank having a single database of all transactions, everyone who transacts has a copy of the database and deltas are shipped everywhere. The Bitcoin ledger is already 20GB, and recomputing the ledger for new machines takes hours.

The "10 minute" problem

It famously takes nearly 10 minutes for a transaction to be added to the bitcoin ledger due to the size of a transaction batch and the time needed for the lottery "race" to complete. In a world where a 400ms response time from a credit card machine feels interminable, 10 minutes is a ludicrous amount of time to transfer money between parties.

Conclusion

Blockchain enthusiasts will surely offer counterarguments to some of these concerns, and cite ongoing solutions that are underway. For me, none of them change the fundamental structure of blockchain in a way that seems to enable it to become the game changer it was once hailed as.

Efforts like "Proof of Stake" promise to supplant the energy-hungry "Proof of Work" model, but have plenty of flaws and often rely on some degree of centralization - the exact trait the original blockchain vision was meant to eschew. Models are also being introduced to try to speed up the transaction time but they introduce new vulnerabilities that need to be solved. And don't even get me started on what wide scale quantum computing will do to blockchain (read: destroy it) , though in fairness that moment is a long way off and will create havoc for security in general.

I wonder with all this in mind, if we won't one day look back with a chuckle at blockchain - a flashy curiosity that never went anywhere. Blockchain : the fidget spinner of technology ideas.