On January 20 and 21, 2017 this piece titled GitHub for Lawyers was thrust into the top-ten on Hacker News. That in and of itself is notable because it evidences the acceleration of increased attention that legal tech is receiving.
GitHub for Lawyers in Hacker News Top-Ten
There are scores of concepts, projects, products, and companies jockeying for ascendant slots in multifarious niches — everything from blockchain to research analytics. The catalog has grown legion, and the full roll is not likely comprehensively enumerated anywhere. But the prior, present and future art inventories compiled by Legalese provide a fair exposition of what has transpired, is transpiring, and might transpire in this space. (Legalese is a Singaporean software startup building and compiling a language and app environment for smart/automated contracting. The sheer number of words set down on its site reveals the endeavor will not be a simple one. Full disclosure — I sit on its Slack team.)
I am going to break a bunch of hearts here and say it like it is. Technologists are overly sanguine about what code can do for law. The best practice of law, contrasted with law per se or the rote delivery of templates, draws on emotional intelligence no less than it draws on any codified authority, and perhaps more so, or much more so. The practice of law is as complex as the navigation of love’s intricacies, if not its perils. Human psychology, and the fathomless variables instantly and constantly processed by the matrix of five senses, and their tie-in to experience, training, knowledge, skill, intelligence — nay, wisdom — all superimposed over and contextualized within the interminable flux of external events, not least of which is the incessant evolution of law itself, for which no statistical controls exist, coupled with counterintuitive stratagem and figurative language, etc. The equation for talent is infinite; the numerics are as irrational as the people involved, and exponentially so as the persons and entities multiply; and computational capacity, thus its authority, degenerates at the nexus of it all. Quite simply, outside of a limited universe of circumstances, clients do not hire lawyers to spit out formulas — they hire them to solve entirely human problems. Law is not chess; there are not a finite number of moves, even if that number approaches the limits of comprehension. It is often more like marriage on the cusp of divorce, wherein mathematics is distinctly incongruous. The notion that there exists a solution for legal that parallels Turbo Tax/accounting, or Photoshop/design, or Autodesk/engineering, overlooks the mathematical foundation of the latter three fields and the social nature of the first.
A recent paper from Karen Levy at Cornell really seizes on this point. I relate the abstract here in pertinent part:
This paper critiques blockchain-based “smart contracts,” which aim to automatically and securely execute obligations without reliance on a centralized enforcement authority. Though smart contracts do have some features that might serve the goals of social justice and fairness, I suggest that they are based on a thin conception of what law does, and how it does it. Smart contracts focus on the technical form of contract to the exclusion of the social contexts within which contracts operate, and the complex ways in which people use them. In the real world, contractual obligations are enforced through all kinds of social mechanisms other than formal adjudication—and contracts serve many functions that are not explicitly legal in nature, or even designed to be formally enforced. I describe three categories of contracting practices in which people engage (the inclusion of facially unenforceable terms, the inclusion of purposefully underspecified terms, and willful nonenforcement of enforceable terms) to illustrate how contracts actually “work.” The technology of smart contracts neglects the fact that people use contracts as social resources to manage their relations. The inflexibility that they introduce, by design, might short-circuit a number of social uses to which law is routinely put.
This is quite the reason why lawyers practice law and never plainly do it, let alone master it. Every experienced attorney worth his or her salt knows that artificial intelligence would make incompetent counsel in all but a few circumstances. Coders invested in legal tech products will try to have us believe this position is premised on a self-interest in inefficiency, and some of the 50 comments to the Hacker News listing touch on that mentality. Those of that mindset, whether lawyers or developers, must have very narrow exposure to lawyering, and it is an ignorance they apparently will be made to greet only upon failure. What developers in this space are attempting to do is not merely compare law to code; they are trying to artificially imitate the practice, which is a different enterprise altogether — one situated in a weird place between idolatry and anthropocentrism, to think there is nothing human we as humans cannot recreate.
Accordingly, this piece reads a bit like nonsense (and to compare Musk with Einstein is thoroughly misplaced). Furthermore, headlines like this one are disingenuous — reading into the article proves it is not only lawyerly time that was saved, but general staff time, and that some of what the noted software “innovations” were doing was resetting passwords and checking the text of documents. That sounds at the margin of sophisticated lawyering, if anywhere near it at all, whether lawyers were doing any of that work or not. Knocking lawyers is a perennial and ubiquitous pastime, and headliners love doing it as much as anyone — perhaps now more than ever before, when there is a fashionable wholesale attack by one industry against all others, i.e. software.
Now, that is not to say there are not a few circumstances valuably to be addressed. The application of AI to a narrow subset of services may well prove lucrative, utile and convenient, and the economics of it may well support a thriving industry. I encourage that as I might any other efficiency, and I will not expend another moment delimiting the bounds of these ventures — they are conspicuous.
So the ever-burgeoning interest in the intersection of law and code is not groundless. It likely signifies the common process of crystallization, after which the crystal can and will be sold. Consistent with the allure of what shines in the mind, one of the most visible writings I have produced to date, despite that it is essentially just a snippet, is again the subject of this note — GitHub for Lawyers. Here are some stats from just that one twenty-four hour window during which the piece ran in the top-ten on Hacker News.
Realtime Stats at Height of Activity
At the height of activity the Google Analytics control panel showed over 100 visitors online at any given time — visitors of course came and went, but the page views piled up per minute as they did (seen in the panel marked with a red star). GoSquared registered about 150 visitors online at any given time. The total tally over the twenty-four hour period that the piece ran on the front page of Hacker News was about 19,000 page views. Here is a look at the viral arch, and how it wound down over almost exactly that twenty-four hour period.
The Viral Arch and Winding Down
Below you can get a better sense of engagement (seen in the panel at the top left marked with a red star). About half the visitors at any given time had either opened a tab to read it later or were engrossed in the content of the site. I would guess most of those were opting to read later, as the 43% active stat tends to depict. Less than a one-tenth were having a quick look. The residual bulk, nearly another half, would seem to have been reading the post, or digging further into other content on the site, as the most visited pages image details below. I find this breakdown acceptable — it depicts an appreciable readership.
Time and Degree of Engagement
Now let us have a look at some aggregate demographics — most conveniently depicted by Google Analytic’s iOS app. Keep in mind that data on a top-ten listing may reflect a fair cross-section of Hacker News readers altogether — useful information all on its own. The first image shows a Slack integration that notifies me when a certain threshold of visitor count is reached. That prompts me to investigate spikes in activity and suitably respond to a specific interest. The next three in the series show: (1) the cities from where most visitors hailed; (2) the most popular sources of visitors; and (3) the most visited pages on the site. Note that the “direct” source is most likely also fully Hacker News.
Next let us look at the technology used. What you will notice probably will not surprise you — most visitors used a desktop/laptop and Chrome, but if it was a mobile/tablet device in hand, the iPhone was vastly dominant.
There are a couple of other interesting things about this piece running in the top-ten. It was by far the oldest piece in the top-ten (2013), and it was the oldest piece in the top 100. That again speaks to the staying power of the concept. But there is more to say — on the substance of the piece, and how it was harnessed in discussion by others.
The community’s comments to the piece tended to focus on legal tech products and that market — all good and relevant, but not exactly what GitHub for Lawyers is about. The piece is actually about how legal drafting is procedurally and logistically consistent with software code drafting, and that one of the most popular tools used by developers for version control (Git) could just as well be put to use by those desiring to manage a legal drafting project.
In particular, I strike a note on public participation in the legislative process (see also this piece in the same vein, which, not coincidentally, received similar attention on Hacker News). Our current model for representation by Congress and the Senate is exactly that model deployed when our federation of states barely broke the teens in number, was concentrated on the eastern seaboard still distant from even the Mississippi River, was populated by something like 1/50 the population represented coast-to-coast and in Alaska/Hawaii today, and when transportation technology consisted of horse and saddle, or waterfaring steam vessel — not even train. There were then good reasons for representative bodies — it was a convenience, if not a necessity of federal governmental management.
But each of the reasons on the foregoing list has lapsed. So why not encourage citizen groups to form their own lobbies, draft legislation using version control tools, and advance causes through bill-drafting efforts? Do not ever be fooled into believing the power does not remain in your hands. Perhaps now more than ever it truly does.
Update: March 21, 2016
The New York Times recently published an article concurring with the thrust of this post, saying, in sum, “like it or not, a robot is not about to replace your lawyer.” You can read that piece here.