Two weeks ago, we wrote an intro-duction to digital currencies in which we pointed out how fast these new instruments have proliferated. Nearly a thousand are currently available — and almost a hundred more are now listed on coinmarketcap.com than when we ran the article. (A lot of readers found the piece helpful; if you haven’t seen it, please feel free to request a copy.)
As we’ve discussed in recent weeks, these various forms of “gold for the digi-tal age” take the basic innovations of bit-coin — encryption and a decentralized ledger system — and modify them to add desired features, such as improved privacy or better transaction processing speeds. So far, bitcoin is still the largest, with a market cap of about $39 billion as we write — but the next largest compet-itor, ether, has gone from a $700 million market cap to a $19 billion market cap since January 1 of this year. Ether has risen on the strength of its underlying platform, which is designed to facilitate the execution of electronic contrac-tual agreements. Among the hundreds of other “alt coins,” there are certainly many with advantages that will be com-pelling, although most will be “zeroes.”
exist completely outside the regulatory framework that governs the initial pub-lic offering of stock in newly listed public companies.
In our previous coverage of digital cur-rencies, we have pointed out some of their potential pitfalls for would-be investors and speculators. For example, unexpected technological developments that could threaten digital currencies’ cryptographic foundations; or the need of most users to interact with the world of digital currencies through a system of intermediaries which may not be as secure from fraud as the pure technical design of these instruments would sug-gest. There is also the specter of inter-nal disagreements leading to “forks” in which currencies split into rival versions in response to some technical challenge or problem. (Ether has already weath-ered one of these, with its platform split-ting into ethereum and ethereum classic. Bitcoin itself is nearing such a moment, with rival camps unable to agree on an approach that will speed transactions.) These growing pains will result in vola-tility, as if there weren’t already enough of that for digital currency speculators
New coin launches are proving to be interesting in their own right — particu-larly in their wider ultimate implications for financial markets. So far this year, new coins have raised more than $1 bil-lion. Many investors in bitcoin and ether are buying new coins at their launch, or “initial coin offering” (ICO) using their existing cryptocurrency stash — a kind of diversification. And of course, ICOs
The rise of prominent ICOs, and the first stirrings of digital currency hedge funds, suggests to us another issue on the horizon. Digital currencies are above all, as we observed two weeks ago, a gauntlet thrown down in front of government and financial regulators. Regulators’ responses thus far have been either hesitant or permissive. As digital currencies become more prom-inent — as the news flows, transaction
volumes, market capitalizations, and hedge fund activities grow — regulators around the world will become less sanguine about the existential challenge being made to their power. How will the communi-ties of the various cryptocurrency systems respond? Will they accommodate — or will they entrench?
We do not know how this, the most basic crypto-currency conflict, will play out — but speculators should be cautious and aware of the regulatory mood. We believe that cryptocurrencies will be able to survive regulation, since evading it has been part of their core raison d’être. But the process in which they accommodate themselves to a reg-ulatory presence, or fail to do so, will shape their future — and determine if they will become more predictable virtual commodities, or remain a “wild west,” with the excitement, freedom, and risk that that entails.
Certainly, cryptocurrencies will make for interest-ing times.
Investment implications: As we have noted in our previous coverage of crypto-currencies, we don’t believe that they’re investments at this stage — the risks and uncertainties are too large. Speculators should be comfortable with those risks and uncertainties. Some level of technical expertise is necessary for intelligent spec-ulation. Speculators should also monitor the regulatory atmosphere carefully, since regulators’ attention, no matter what the eventual outcome, could cause severe vol-atility, or even the collapse of coin ecosys-tems.
Robots Diagnose Skin Cancer As Effectively As Doctors
Several weeks ago, we wrote a piece on the cur-rent acceleration of artificial intelligence and the new chips and innovative chip architectures behind it. Recently we reviewed a stunning arti-cle in Nature, “Dermatologist-level classification of skin cancer with deep neural networks.” The paper was co-authored by a group of Stanford Univer-sity scientists which includes Sebastian Thrun, the founder of X, Alphabet’s semi-secret “moonshot” lab which is now a standalone company under the broader Alphabet umbrella [NASDAQ: GOOG].
The paper’s authors wanted to build an artificial intelligence system that could assist human doc-tors in correctly identifying melanomas. 20% of Americans will be diagnosed with a skin malignancy during their lifetime; melanomas are represent some 5% of skin cancer but account for 75% of skin cancer deaths, and identifying them early is critical. Melanoma identified in its earliest stages has a 99% five-year survival rate; that drops to 14% if it is caught in its latest stages.
Using a GOOG-designed “convoluted neural net-work,” or CNN, the team took a clinical dataset of nearly 130,000 images, including photographic and dermoscopic images, and used it to train a ma-chine-learning system. Dermatologists labelled the images according to a tree-structured taxonomy of 757 disease classes and 2,032 specific diseases.
This disease taxonomy proved powerful in train-ing the neural network. In distinguishing between benign, malignant, and neoplastic lesions, the neu-ral network achieved 72% accuracy, versus 66% for human doctors. In the second-level nodes of the taxonomy, where finer distinctions are being made
between disease types, the system achieved 55% overall accuracy versus 53–55% accuracy by der-matologists.
The study concludes: “We match the perfor-mance of at least 21 dermatologists tested across three critical diagnostic tasks… This fast, scalable method is deployable on mobile devices and holds the potential for substantial clinical impact, includ-ing broadening the scope of primary care practice and augmenting clinical decision-making for der-matology specialists.”
Investment implications: As we observed in last month’s piece on artificial intelli-gence, the application of AI requires three basic components: data, hardware, and training systems for the artificial neural networks. Access to data — public, private, or proprietary — will become a key eco-nomic variable for company performance. Companies with data and the capacity to generate it, as well as companies with the political savvy to make use of exter-nally generated public and private data, will benefit. The key in hardware is the arrival of chips which discard the leg-acy chip architecture in favor of new architectures intrinsically suited for deep learning application — neuromorphic chips. Look not just for manufacturers, but for those companies that can harness network effects to win dominance in the adoption of their chips and their program-ming ecosystem. Finally, look for com-panies able to implement fast and cheap training systems at scale.
Blockchain May Ease Compliance and Help Developing Countries
The post-crisis world has seen big increases in regulatory burdens for financial institutions. This is something we’ve often written about in these pages from the perspective of the small and mid-sized U.S. banks who have suffered from the increased burden of regulatory compliance — and many of which have merged or folded as a conse-quence.
However, it’s not just within the U.S. that this trend has been playing out. It’s a problem that has beset banks in many of the world’s smaller countries. In a process of “derisking,” large international banks are pulling back from many smaller economies, par-ticularly in eastern Europe and other parts of the developing world. It makes sense from a business perspective. Dealings with smaller countries are less profitable and still present large fixed costs in terms of compliance with anti-money-laundering regulations, terrorism, and conforming to sanc-tions restrictions. But it presents considerable trouble for smaller economies, and ironically, may actually create more clandestine activity by driving market participants into cash and informal financial networks.
The trend can also spell trouble for international charities. Critical activities can be challenged by
financial hurdles — there are already unfortunate stories about relief efforts in Syria and Afghanistan being blocked by financial intermediaries, with aid unable to reach its recipients.
Last week, we talked about the potential benefits of one of bitcoin’s underlying technologies, the “blockchain,” outside the realm of digital curren-cies. The current case — where small participants in the global financial system find themselves mar-ginalized by growing compliance costs — is a per-fect example. JPMorgan Chase is one of the largest international clearers of dollar transactions — and like others, has been withdrawing from smaller markets due to the economics of compliance. But JPMorgan hopes that blockchain technology will be able to encode a traceable, transparent record of transactions — and allow compliance without the huge back-office overhead.
Investment implications: Many institutions are investigating the use of blockchain and distributed ledger technology. The use of this technology to streamline compliance will be a key application, and will reduce compliance costs for some of the institu-tions most burdened by post-crisis regula-tory growth.