Here's a video of Alex's and my presentation at lawTechCamp 2012 (held on May 12, 2012 in Toronto, Canada) titled My Co-Worker is a Robot: The [Coming] Age of Legal Informatics. Really enjoyed the event. And we were lucky to have a number of good audience questions during our presentation. Enjoy and please let us know your comments!
Here are videos of a number of other presentations at the event.
Alex and I (Noah) are speaking at lawTechCamp in Toronto this Saturday, May 12. We're trying to make our presentation My Coworker is a Robot: The [Coming] Age of Legal Informatics great, and there are a bunch of other interesting-looking sessions on the schedule. Our presentation is at 9 am and is accredited for 1 CPD professionalism credit. The event is free but advance registration is required. I went last year and many sessions were interesting (if you care about legal tech (which, if you're reading our blog, you probably do)). Introduce youselves if you do come!
Do you know how accurate your due diligence contract review is? Seriously, do you? Does your team find every relevant change of control provision? 75% of them? Under 50%? How do you know how accurate it is?
If you're at a good firm, with junior associates hired from top schools and good quality training, you might feel the juniors are finding all relevant provisions they have been tasked to find. But this is a feeling, not a fact. Even good people get tired and distracted. And provisions can be buried in agreements. Those involved in closely supervising diligence know that provisions often get missed. Perhaps you have midlevel associates do a second level review from scratch. But they get rushed and may also missed buried provisions. Again, how do you know how accurate your due diligence is?
Missed provisions in due diligence investigations can be a big deal. You may be giving inaccurate advice if your are basing it on incorrect facts. And incorrect facts can come from faulty due diligence review.
We can help. DiligenceEngine finds 90% or more of almost every provision it covers. We know this based on our automated testing system. While 90% is not the perfect accuracy lawyers strive for, it is very accurate. And actually tested and known. Why not run the system on your agreements and get a more accurate idea of how accurate your diligence is. We can facilitate this process if you're interested—get in touch! You have nothing to lose, apart (perhaps) from a false sense of security in your due diligence quality. Or not—your current process may be great. At least you will know.
;; Clojure Software Developer @ Toronto Startup
(ns com.diligenceengine)
(def startup "DiligenceEngine Inc.")
(def about-us
"We're a Toronto-based startup using modern
computer science techniques to simplify corporate
legal practice. Our small tight-knit team works to
combine modern web technologies with data from
sophisticated machine learning backend services.")
(def looking-for
"We're looking for someone that shares our
vision and passion for building quality software.
As an exceptionally-talented engineer who is
excited to work with a team of like-minded
individuals, you will build out features that
bridge the data engine and web interface.")
(def topics [:clojure :web-dev :functional-prog :postgresql :rabbitmq :java-jvm :unix])
(def requirements [:comp-sci :comp-eng :soft-eng :other-related])
(def why-work-with-us
["Awesome co-workers of course!"
"Challenging & interesting problems"
"Responsibility & ability to make an impact"
"High-quality mentoring to help you learn"
"Competitive salary"])
(def show-off
"Please send us your cover letter, resume, and
any references to software you have written or
contributed to (e.g. pointer to some of your open
source projects on Github):
jobs@diligenceengine.com. Include your name as the
email subject.")
(def bar [33 115 107 115 105 114 101 116 115 97 32 110 105 32 101 109 97 110 32 114 117 111 121 32 101 115 111 108 99 110 101 32 100 110 65])
(defn pretty-print [i]
"I print out stuff really pretty for your i!"
(print (apply str (interpose ", " (map name
i)))))
(defn print-ad []
(-> "Hey, we're from " (str (.toUpperCase
startup) "! ") (str about-us " " looking-for)
(println))
(println "\nYou can expect to work with the
following, so familiarity with any helps: ")
(pretty-print topics)
(println "\n\nYou should also be either working
towards or already possess a degree in:")
(pretty-print (take 3 requirements))
(print " or ") (pretty-print
(drop 3 requirements))
(println "\n\nWhy work with us:")
(doseq [reason why-work-with-us]
(println " * " reason))
(print "\n" show-off "")
(println (apply str (map char (reverse bar))))
(println "\nHappy Coding!"))
DiligenceEngine Inc. is a Toronto-based startup focused on building lawyer automation software. We understand the importance of design in creating the ultimate web experience for our users and, as such, are looking for a Web Designer to work alongside our Web Developer! You will take the lead on driving our brand across all channels - web, video and print. Specifically, you'll be designing further pages of our website, contributing to the user interface design, creating promotional videos, and designing other marketing materials involving print.
You should be comfortable working with HTML / CSS, and have a really solid understanding of web interfaces and usability. Our small and dynamic team will provide the overall creative feedback and direction, as well as a competitive salary.
This position is for a three-month period beginning June 1st or sooner, and you must be a returning student in the Fall.
Feel free to wow us with your resume and portfolio link at jobs@diligenceengine.com!
Note: We are no longer accepting applications for this position.
Google announced a partnership with NASCAR for their 2012 April Fools joke: Google's self-driving cars were to compete in NASCAR races. This set-up a Watson on Jeopardy! moment for computer-driven cars. In fact, the Google/NASCAR announcement was a prank. But not one that was far off reality. Google's self-driving cars apparently work well. They have driven over "200,000 miles without a single machine-caused mishap." While a safe driving record doesn't seem key to NASCAR success, Google cars might work very well in NASCAR races. Self-driving car systems could be trained with data from previous NASCAR and other auto races to understand how the computer and its opponents might act in given situations. And even how specific NASCAR opponents are likely to behave. This could give a Google NASCAR entrant an edge against human competitors.
Should race car drivers someday face computer competition, they may find the path to success through being themselves assisted by a computer. A computer could rapidly calculate options and likely outcomes. Giving the driver ideas. Winning drivers will have to move beyond calculation to provide something extra. This is similar to what has happened in some types of chess (as related in more detail in MIT's Eric Brynjolfsson & Andrew McAfee's excellent "Winning the Race With Ever-Smarter Machines"), where human + computer chess teams have been successful.
Lawyers should consider this situation. Computers may soon safely drive cars on the road and in races. And write news articles. And do document review. And do due diligence. Now is a good time to think what this will mean for your practice. And consider how to work with coming technology to do higher quality work and thrive as a business.
(photo: http://www.flickr.com/photos/pranavian/6487719111/)
Has your once-good ability to remember phone numbers declined to the point where you can dial few without looking? Do you now also forget facts, dates or the exact wording of legal rules that you once remembered well? Why is this? General memory loss [due to aging]? Or because of technology's coddling effect? This is not a medical blog, so let's put aside memory loss to consider whether technology is at fault. And whether that means technology is making us dumber.
We don't have to remember many numbers, dates or facts in today's technology-supplemented world. They are quickly avaliable online, or accessed automatically by systems we use like cellphones. Because we don't need to remember these details, we are likely reducing the portions of the brain responsible for storing this information. Human brains have been found to be plastic–continually adapting to their circumstances, evolving to meet demands. The brain can reengineer functions after parts of it are damaged in incidents like strokes. And parts of the brain with high demand overdevelop. As Dr. Pascale Michelon relates in the SharpBrains blog:
For instance, London taxi drivers have a larger hippocampus (in the posterior region) than London bus drivers (Maguire, Woollett, & Spiers, 2006). Why is that? It is because this region of the hippocampus is specialized in acquiring and using complex spatial information in order to navigate efficiently. Taxi drivers have to navigate around London whereas bus drivers follow a limited set of routes.
Plasticity can also be observed in the brains of bilinguals (Mechelli et al., 2004). It looks like learning a second language is possible through functional changes in the brain: the left inferior parietal cortex is larger in bilingual brains than in monolingual brains.
Plastic changes also occur in musicians brains compared to non-musicians. Gaser and Schlaug (2003) compared professional musicians (who practice at least 1hour [sic] per day) to amateur musicians and non-musicians. They found that gray matter (cortex) volume was highest in professional musicians, intermediate in amateur musicians, and lowest in non-musicians in several brain areas involved in playing music: motor regions, anterior superior parietal areas and inferior temporal areas.
So tech may be changing our brains. We are losing the ability to remember things we no longer have to remember. But that doesn't necessarily mean we're getting dumber. It means our brains are changing to do what we need them to do. Instead of remembering facts, today's more valuable skill is remembering where to quickly find facts. We can now access much, much more information than we ever could have remembered. For example, I knew a touch about neuroplasticity when I went to write this piece, but was able to quickly call up a lot more information. I might have done badly on an in-class exam on neuroplasticity, but that isn't the test I took.
Brain plasticity is a happy concept as we confront a world where computers can do more and more advanced human tasks, from driving, to winning at Jeopardy!, to document review and legal due diligence. Our brains can adapt to work with new technology. And figure out how to thrive despite traditionally-human work being done by machines instead.
Since this is the DiligenceEngine Blog, it seems like a good place to consider whether use of our system (which finds diligence-relevant provisions in contracts and puts findings into summary charts (sort of like an automated junior associate)) will ruin junior associates' ability to identify change of control or amendment provisions, forever stunting their growth? Stay tuned and I'll cover this in a future post.
Information overload is one of the world's major commercially-addressable problems. More information than ever is accessible. But it's hard to find what we're looking for. Many companies have stepped into the breach, ranging from search providers like Google or e-discovery predective coding document review companies, to news aggregrators like Alltop or newsana, to more individualized recommendation sources like bloggers or Twitter posters (though the latter two can also be part of the problem). And us. We all share a focus on helping people find the information they need, whether it is Google with its broad mission of organizing the world's information or us with our much narrower focus on getting relevant sections of contracts to people who have to review them.
Sometimes, though, too much information is fun. Like for one of my former law firm colleagues who came across the extremely entertaining Pub Crawl Holdings 10-K while doing EDGAR research (for unfamiliar readers, Form 10-Ks are annual reports that some companies must file and EDGAR is the Securities Exchange Commission system where information like this is filed and searchable electronically). We are not putting this 10-K up to mock Pub Crawl Holdings. Rather, we see this 10-K an example of the sorts of information very expensive lawyers have to hunt through to find valuable material that will help their clients. And it happens to be entertaining. So now, without further ado, excerpts from the Pub Crawl Holdings, Inc. 10-K.
Pub Crawl Holdings 10-K
Pub Crawl Holdings (see their logo here) is
an Internet-based company specializing in providing information on Happy Hours, drink specials, nightly specials and “pub crawls” for bars and restaurants in San Diego, California. A “pub crawl” as used throughout this statement refers to an organized event whereby a group of people collectively visit multiple bars in a single night. Pbpubcrawl.com was launched January 6, 2010 as an informational website for people in the Pacific Beach and Mission Beach areas of San Diego, CA. The website contains information on the bars and restaurants in the Pacific Beach and Mission Beach areas, and offers these bars and restaurants a chance to showcase their businesses at reasonable advertising costs.
Revenues have been low so far:
the Company generated revenues of $884 from advertizing and marketing sales compared with $356 from the period from May 27, 2010 (date of inception) to June 30, 2010.
Those numbers are not in thousands.
In case you wondered about their real estate situation, their response to Item 2 informs us that:
We currently are using a portion of our Chief Executive Officer’s home as our corporate headquarters located at 802 Sunset Court, San Diego, CA 92109, and we are using the space rent-free. As of the date of this filing, we have not sought to move or change our office site. Additional space may be required as we expand our operations. We do not foresee any significant difficulties in obtaining any required additional space. We currently do not own any real property.
Good news if you're looking to buy in: they
intend to contact an authorized OTC Bulletin Board market-maker for sponsorship of our securities on the OTC Bulletin Board.
But shares are closely-held:
As of October 6, 2011, there were 5,000,000 shares of the registrant’s $0.001 par value common stock issued and outstanding and were owned by approximately 1 holder of record, based on information provided by our transfer agent.
Pub Crawl Holdings doesn't anticipate dividends in the foreseeable future, but the $74,250 they incurred in 2011 professional fees is paying dividends to us, the reading public!
Though I didn't do a form check, Pub Crawl Holdings seems to thoroughly describe their business and have a perfectly adequate 10-K (which it looks like they didn't get any SEC comments on). They appear to be pursuing a "get a minimum viable product out and iterate" strategy, which is popular and has thoughtful supporters. And makes for much more fun reading than the average 10-K!
(photo: http://www.flickr.com/photos/foolstopzanet/217082717/)
I'm happy to report that we now have a website that says what we do. Which is build software that helps users review contracts and organize their findings. More specifically, our system reads electronic versions of contracts for certain provisions and puts its findings into summary charts, which users can then manipulate and store online or download. Our website says it more elegantly than I can here, so check it out. Essentially, you could think of our system as like a junior associate on a diligence project, except that it's way quicker—documents get reviewed in about 30 seconds each. Regular readers know that we think junior lawyers have nothing to worry about from our system and tech like it (see, e.g., this and this). And in fact we see a main use of the system as helping junior associates prepare more accurate diligence summaries much faster. And way less painfully. I know I would have loved it as a junior, and I hope today's juniors will be as enthusiastic about the system as we are.
Has DiligenceEngine Launched?
We now have test users on the system and are using their feedback to make DiligenceEngine thouroughly amazing before getting it out to more people or actively promoting it. So no, we haven't yet launched. Stay tuned—we hope you'll be hearing lots more about us soon! And let us know if you would really like early access—we may be able to make it happen in limited circumstances.
Why Wait to Have a Descriptive Website and Launch Open Access to the System?
We know that the trendy thing to do in tech these days is to get an imperfect product out as early as possible, get feedback, and iterate. We think that may be the right idea for a lot of products. But we knew that the technology behind what we were trying to do was tricky, and thought it was better to wait to be sure we could do what we were trying to do before telling people about it. We're not crazy—we talked to a number of lawyers before building the system. And a number of them wondered if our product was even possible. It is. We think DiligenceEngine as it now stands is pretty terrific. But we're waiting to launch. The lawyers we know are busy people. And we remember our Biglaw obsession with getting clients great work product, think a lot of our future users will have the same perspective, and would like to make sure the product we launch is outstanding.
Increasingly powerful technology is transforming industries worldwide, and there's more to come. Even white collar work like that done by lawyers and doctors is more-and-more susceptible to tech-driven change. MIT's Eric Brynjolfsson & Andrew McAfee's excellent "Winning the Race With Ever-Smarter Machines", the cover story of the MIT Sloan Management Review's Winter 2012 issue, explains why technology is advancing so fast, how it will advance even faster, and how man plus machine is more powerful than each on their own.
Brynjolfsson and McAfee commence by relating how quickly technology has advanced in recent years. They include the example of driverless cars: in 2004, it was hard to imagine computers able to drive; in 2010, Google announced that seven of their test cars had driven 1,000 miles without human intervention and over 140,000 miles with only occasional human control. Brynjolfsson and McAfee assert that such dramatic progress is due to the cumulative impact of Moore's Law played out over time. Moore's Law relates that number of transistors that can be placed on an integrated circuit doubles every year or two (there is a lot more detail on the Law available). More broadly, Moore's Law has come to stand for computing power's rapid and dramatic increase. Brynjolfsson and McAfee add that quick doubling may not seem like a big deal at first, but then it does, and then it just keeps happening; doubling represents much more than straight-line growth. If computers able to drive was unimaginable six years before it happened, it should take far less than six years for comparable innovation to happen now. They continue,
These results will be felt across every task, job and industry. ... computers increase productivity not only in the high-tech sector but also in all industries that purchase and use digital gear. And these days, that means all industries; even historically low-tech American sectors like agriculture and mining are now spending billions of dollars each year to digitize themselves.
Professionals including lawyers will be affected by increasingly-powerful technology. Brynjolfsson and McAfee have written, including in a heavily-reviewed recent book, about man's "Race Against the Machine". While humans competing with computers in tasks computers are good at is generally a losing strategy, Brynjolfsson and McAfee suggest that that man plus machine is often much more powerful than each alone. And that human + computer combination may unlock immense advances and value gains in coming years.
Software is increasingly able to do legal work. This is occurring at multiple levels of practice, including services like LegalZoom or Rocket Lawyer preparing simple personal legal documents and software doing document review better than humans can. There is much more of this to come, and lawyers and firms should consider how to respond if tech pops up that can do their job faster, more accurately, and (often) for dramatically less money. Brynjolfsson and McAfee's writing suggest that the solution is to work with technology to do even better work than was possible before. Lawyers have already seen how they can leverage technological tools, ranging from word processing programs to legal research search engines, to do higher quality work in less time. Attorneys who use, instead of fight, new technology will come out ahead. And maybe make more money being more efficient in the process.
