Can we (or should we) open source the legal profession?
When it comes to sharing of knowledge and tools, there is a common perception that legal professionals are not very good at sharing of its valuable knowledge and resources. Whereas the common perception (one which I used to hold) is that the tech industry is much better at sharing than the legal profession. For me, the foundation of this belief is the prevalence of open source development and distribution of code which has taken over the world in the last decade, to the point that many of the modern software applications would not run at all without them. In contrast, the legal profession seemed to have a common practice of rivalry and secrecy.
I had an interesting discussion on this topic of sharing of knowledge and open sourcing with John Scrudato last week. On top of his day job, John has been working on an open source tool for lawyers to annotate and share labelled data. It was John that pointed out that legal professionals do share a lot of valuable knowledge and resources, but that they may not be as portable as code – all the case laws, template documents, opinions, and commentaries that legal experts release to the wild are as powerful in their own arena as any open source codebase in fulfilling its purpose.
In this 30 minute conversation, John and I chat about the unusual topic of the role of open sourced software for legal. Our conversation traverse these topics:
what would make legaltech “seamless” for lawyers
pros and cons of point solutions vs platforms
do legal professionals “open source” their resources
why would someone open source legaltech code
what lessons about sharing can the legal profession learn from the tech industry
why “selfish altruism” makes sense for technology entrepreneurs
does open sourcing threaten lawyers – collaboration, competition or “coopetition”
open source projects for law, like:
the Accord Project – smart legal contracts
the Atticus Project – open source contract labels
LexNLP – information extraction
where does John’s “open source low-code tool” fit into the legaltech ecosystem
how do we leverage legal data for data-driven law
how can we encourage people to share code for the legal profession
One of our favourite quotes that captures why open source makes sense for legal use cases was:
“... there is a donut hole in the middle, where you have some kind of automation project or problem, where you can do it with code and you can do it pretty easily, and you can do it on a scale where for human it would be a pain in the butt. Right, like, I’ve got 500 or 1,000 documents, no one is going to build a business off that, and it might be one of those situations where it’s so specific that a multi-purpose platform probably, no matter how good it is, isn’t going to be suited to it. That’s the donut hole.”
Transcript of conversation
(Note: transcript is automatically generated, and may contain misspellings and typographical errors)
00:00
so today we're talking with john
00:01
Scrudato a legal innovator
00:04
who is doing something really cool in
00:06
the open source space for legal
00:08
that's right absolutely a lot of legal
00:11
tech
00:11
claim to be designed for users
00:14
oh god it is it is going to be seamless
00:18
seamless this kind of like like almost
00:21
imaginary
00:22
world where law firms and
00:25
corporates don't have legacy systems
00:28
that are clunky that are stuck together
00:31
with duct tape
00:32
yeah well that's not always true though
00:34
to be fair
00:36
certainly not some organization often
00:38
true often true
00:39
yes but this claim that legal tech is
00:41
seamless
00:42
yeah i think it's greatly exaggerated
00:45
sure
00:46
um but i would like to ask you
00:49
what would make legal tech scenics i
00:51
think there's a lot of answers to that
00:52
question right i think
00:54
you know there's there's internal um
00:57
requirements as you hinted at right
00:59
there's
01:00
different systems that exist that you
01:02
need to work with for better for worse
01:04
that
01:04
i think from an engineering perspective
01:06
there's oftentimes a tendency to sort of
01:09
uh you know assume the best case right
01:11
assume well
01:12
just don't do it that way anymore just
01:14
you know don't use this legacy system
01:16
don't don't collect that data point
01:19
don't
01:19
make people fill out this form with this
01:22
particular format
01:23
do it differently um and i think that
01:27
impulse is frankly an obstacle from
01:30
the sort of the human center design
01:32
perspective
01:33
and that you can say that to an attorney
01:36
but
01:36
when they're working at 110 pace
01:40
and you go well look i know you're
01:43
working flat out
01:44
and your clients are happy with your
01:46
service and you're making a lot of money
01:48
but there's some other metric that i
01:51
think is important
01:52
and i think you should change what
01:53
you're doing and do it my way even
01:55
though i've never done what you do
01:56
you should give it a shot yeah i think
01:58
that's a difficult sales pitch
02:00
yeah and i think you know that's one of
02:02
the reasons why uh
02:03
sometimes it can be hard to get people
02:05
to change because for them
02:08
i think it's sort of an adversarial
02:09
opening an entree to
02:11
innovation right yes right like i mean i
02:13
think you can get people there
02:15
and on some level like you do have to
02:16
get people to change what they do
02:18
for some things but i think
02:20
unfortunately a lot of times
02:23
i think sort of the first impulse is i
02:25
know you work a certain way
02:26
but here's this product and it's going
02:28
to make you work completely differently
02:29
and you're going to love it
02:30
and i just i think that's a difficult
02:33
sell for a lot of people
02:34
so so essentially transformation getting
02:36
in the way of adoption
02:38
yeah you have to be an ally right like
02:39
you have to come to someone and be like
02:41
look i know
02:42
how you work and here's this tool that
02:44
isn't going to take
02:45
30 minutes your time that you can kind
02:47
of just come at
02:48
and for whatever this little thing is
02:49
that you're trying to fix it kind of
02:51
makes sense to you
02:52
right you that that in that opening
02:55
entree that
02:56
intro needs to be something that
02:57
honestly i think you need to be able to
02:59
explain in 30 seconds
03:00
i mean it's it's any elevator pitch
03:02
right
03:04
people's attention span is getting
03:05
shorter i think you've got oh yeah
03:07
absolutely
03:07
absolutely that's everyone's attention
03:09
span and then attorneys are so busy
03:11
yeah and they're again in these in this
03:13
day and age with you know
03:15
the coronavirus and everything even more
03:18
so
03:18
yep that you really you've got like you
03:20
know 10 seconds to basically
03:22
show not tell here's what this is and
03:24
here's why you should care
03:26
right yeah right now there are two
03:28
breeds of legal tech
03:29
there's platform place and point
03:31
solutions
03:32
yeah and it almost sounds like you're
03:35
saying
03:37
go with the point solutions just give me
03:39
discreet little things that solve very
03:42
specific
03:42
pain points yes or no um
03:47
i think from the standpoint of adoption
03:49
100 point solution
03:51
if your goal is to show something to
03:54
someone that will be intuitive and that
03:55
they'll use
03:56
and that isn't going to be a huge
03:58
barrier to adoption i think definitely
04:00
the point solution
04:02
as point solutions proliferate the
04:04
problem you have
04:05
is that you have people now have data
04:07
silos right and
04:08
they have systems where okay i put the
04:11
data into this system and now i gotta
04:12
take the same data and i have to put it
04:14
into this system
04:15
and then i have to put it into this
04:16
system and you know that becomes a pain
04:19
and there are solutions to that
04:20
obviously i mean you know there's
04:23
all kinds of enterprise search products
04:25
there's you know different kinds of
04:26
automation platforms but
04:29
i think at least my own personal take
04:31
here is
04:32
a good point solution with a somewhat
04:35
open architecture
04:36
yep that can be integrated either to
04:39
other point solutions directly or some
04:41
kind of intermediate platform yep
04:43
i think is probably the route where
04:45
you'll see the most impactful
04:48
innovation unfold over time is this a
04:51
nice segue into
04:52
your project the one that that you've
04:54
been working on
04:56
open legal i mean what i was gonna i was
04:58
gonna go back to the early days of open
05:00
source
05:01
wow history lesson but i think i think i
05:03
think i think it's important though
05:05
right i think
05:06
i think so because it's a cultural
05:07
difference between the technology sector
05:08
and the legal sector
05:10
and i think there's there's actually um
05:13
structural barriers yeah to open
05:15
sourcing in legal
05:17
i think that's right but again i think
05:20
the weird thing is in other ways
05:22
it actually is very much in line with
05:24
the legal ethos to
05:25
share yeah so i think you know to that
05:28
point about you know what is what is the
05:30
culture do people share
05:31
i think it's true that law firms as an
05:34
entity
05:35
don't have a natural um
05:38
desire to share a lot but they do right
05:41
you have partners who they they share
05:43
bar committees
05:44
they write books they participate in
05:47
industry consortia
05:48
so they they do right and i think
05:52
that inclination and even if it's driven
05:54
by self-interest right like
05:56
even if you want to be known as you know
05:58
the expert
05:59
on securities law when it pertains to uh
06:02
you know i don't know uh startups and
06:03
technology right like yep
06:05
there's nothing wrong with that it
06:06
doesn't change the fact that that
06:08
sharing is beneficial
06:10
and it's already natural but when it
06:12
comes to software
06:13
i think you're absolutely right that you
06:15
know as an industry
06:17
i think it's very much foreign to think
06:20
that
06:21
law firms or lawyers or legal
06:23
technologists
06:24
or any combination of those would
06:26
develop and share
06:28
software i think i think that's right
06:31
interesting
06:31
because i was thinking it's it's
06:34
actually the opposite
06:35
okay um that the know-how
06:39
lawyers have of what do you do when you
06:42
see a certain set of
06:43
circumstances yeah these are the facts
06:45
these are the laws what do you do that's
06:47
the value
06:48
lawyers had true true um
06:52
that part that know-how yeah it's
06:56
pretty secret it's closely guarded by a
06:59
lot of the lawyers
07:00
yeah so that's the part i think you know
07:03
is the challenge for open sourcing
07:04
legally
07:05
let me give you a couple of examples of
07:06
where i think i think you're right to a
07:08
point
07:09
yep and but and again we'll get to the
07:11
software piece of it but i think you see
07:12
a similar dynamic
07:14
in uh open source software as well where
07:16
there's
07:17
the part that you share and then there's
07:19
the know-how that you don't right right
07:21
and i don't think those are mutually
07:22
exclusive well google's not sharing
07:24
everything right
07:24
exactly right like look at facebook look
07:26
at google like facebook open source
07:27
react which
07:28
for people who are not familiar with it
07:30
is an amazing open source
07:32
front-end javascript framework that
07:34
makes it possible to do some incredible
07:35
stuff
07:36
on the front end and it's free it's free
07:38
completely completely free there's no
07:39
strings attached
07:40
yet facebook gave it away right yep and
07:43
i think if we look at
07:44
law you see again i think you see a
07:45
similar dynamic
07:47
so one example the national venture
07:49
capital association in the united states
07:51
yes they have an incredible set of form
07:54
documents
07:55
that have been put together by some of
07:57
the greatest minds
07:58
in securities law and corporate law in
08:01
the united states
08:03
and if you want to create um a venture
08:06
deal
08:06
if you want to pay for it you use those
08:08
documents yep and sure you might have
08:10
your version
08:11
but they're based on those documents i
08:13
think you see that with isda right is
08:15
that has its own set of form documents
08:17
and they they do the same thing yep i
08:19
think if we look at
08:20
bar committees you see similar dynamics
08:24
and even something as um
08:27
for a lawyer at least as mundane as case
08:29
law right like you don't think of that
08:30
as a contribution
08:32
yeah but in a way it is right because
08:34
you're giving away knowledge
08:35
right giving away you you read you read
08:38
the holding you read the case law
08:39
there's an incredible amount of
08:40
knowledge that's in those
08:42
those books in those holdings that
08:46
you can go and look through and that is
08:48
shared
08:49
and i think as an as not an industry as
08:52
a profession
08:53
where we have an obligation to society
08:57
right and we're trying to yes we are
09:00
running businesses we're also trying to
09:02
serve a valuable social role
09:04
i think it's very much a natural
09:06
inclination that there's some give and
09:08
take
09:09
and it doesn't mean you give away
09:10
everything but you give away something
09:12
so i think we're there i think that
09:13
mindset is there it's just how we think
09:15
about
09:16
the what like what do we give away so
09:19
that's interesting now
09:20
let's go back to the software side sure
09:22
yeah
09:23
what are you trying to get people to be
09:26
comfortable
09:27
giving away where where is that line for
09:29
you
09:30
yeah so i think you know for me i took a
09:32
couple of uh lessons again from open
09:34
source software and
09:36
even just looking at some of the stuff
09:37
that google and facebook have done we
09:39
were talking about
09:40
react we're talking about uh pi torch
09:42
and tensorflow
09:43
which are platforms where people do
09:45
research
09:47
into machine learning that they can then
09:48
actually deploy into production settings
09:51
and i think one of the things uh that
09:54
google and the ai researchers that work
09:56
in them have seen is
09:58
giving people the tools to create
10:00
doesn't necessarily
10:02
is not a negative for somebody who is
10:04
trying to sell
10:06
uh deep learning frameworks or trying to
10:09
sell
10:09
the the data and the know-how right yep
10:12
i think there's some recognition that
10:15
having more people
10:16
know how to do things in a certain way
10:19
that creates standardization that
10:21
creates
10:22
structure that then speeds up everyone's
10:25
workflow
10:26
is beneficial to everyone yep and so one
10:29
of my frustrations
10:30
in dealing with a lot of legal problems
10:32
is
10:34
it it rhymes but it's not the same
10:37
right like you have interest you're
10:40
smiling so i think you kind of know
10:41
where i'm going i know where you're
10:42
going and i also really like that
10:43
phraseology of it rhymes
10:45
yeah that's not quite the same right
10:47
there's a lot of that in law like
10:49
it's similar situations but requiring
10:52
different answers exactly yeah exactly
10:55
and i think that's
10:56
that's the thing i've seen right is
10:58
you'll get and i this was from my days
11:01
when i practice right like you'll get
11:02
a deal and you're kind of working with
11:04
the same kind of documents
11:06
that you did before but it's a little
11:07
bit different and
11:09
if you want to find a tool to help you
11:12
kind of automate it
11:13
a problem you often run into is everyone
11:16
has their own little way of doing
11:17
something
11:18
right yes so if you want to do it your
11:20
way yep
11:22
then you're unlikely to find someone
11:23
else who's built a tool for you
11:26
that does it exactly your way yep so
11:29
what i thought to myself is
11:30
is it possible to start to break these
11:33
processes and these workflows down
11:36
into their constituent pieces and find
11:38
the parts they're shared
11:40
where they're not really adding a lot of
11:43
value
11:44
but they're they're requisite steps to
11:46
doing some of the more interesting
11:48
automation projects
11:49
doing some more interesting data
11:50
extraction projects where it's not the
11:52
know-how directly
11:53
it's the enabling infrastructure okay
11:57
so i'm gonna use a lego block analogy
11:59
okay
12:00
um because i'm a visual kind of person
12:02
okay it's kind of
12:03
easy to grasp okay so if we were to say
12:08
that the way that the products of
12:10
lawyers
12:11
is like the completed lego model you're
12:14
your death star
12:15
yeah your you know history sure i think
12:18
i have more millennium falcon guys
12:20
fair enough yeah that's a millennium
12:23
falcon
12:23
[Laughter]
12:25
absolutely and um
12:28
what you're saying is you are looking
12:31
for
12:32
something or you're trying to get people
12:34
to share something that
12:35
is the process for breaking down those
12:38
outputs
12:39
into the lego blocks or are you saying
12:42
sharing of the lego blocks
12:45
it's it's a little bit of ball um i
12:48
think there's different situations where
12:50
it's helpful to do either
12:53
to make it a little more concrete though
12:54
so one one of the projects
12:56
i've been working on is just kind of a
12:58
workflow tool that makes it possible
13:00
to take really small snippets of python
13:03
code yeah
13:04
and string them together and then
13:05
basically let someone submit a document
13:07
yep and then it will work on the
13:08
document and give you the resulting data
13:11
or document
13:11
yep and so an example of that might be
13:13
and and this is a really
13:15
a very specific application but you have
13:17
a
13:18
true form document where you know it's a
13:20
form but it's your form so
13:21
the likelihood that someone else is
13:22
going to have a tool that works with it
13:24
it's very small
13:25
and you can use say something simple
13:27
like regex to get data out of it
13:29
you can write a script that's 10 lines
13:31
long that can extract the data from that
13:33
now the classic problem you have is if
13:35
you had that script
13:37
and you wanted to give it to a lawyer
13:39
god knows you're not going to give them
13:40
a python script and say all right here's
13:42
what you do
13:42
go to your command line you type python3
13:45
dash
13:46
you know extract such and such dot pie
13:48
no one's gonna do that
13:50
you start giving me a spanner and saying
13:51
fix that car yeah
13:53
okay it's crazy it's crazy but nor would
13:56
you
13:56
go to a most likely you wouldn't go to a
13:58
software development shop and say look
14:00
i've got a thousand documents and i want
14:02
to do some data extraction
14:03
and i need you to build me a web portal
14:06
to do this
14:07
right because it wouldn't make any sense
14:08
right so what i thought to myself is all
14:11
right well what is the actual piece
14:12
there that's unique
14:13
yep it's the 10 lines python code it's
14:16
not the platform
14:17
right the platform could be the same
14:19
every for if you had a million different
14:21
templates
14:22
it would be the same ui for that example
14:25
it's the same toolbox it's the same
14:26
toolbox so what i want to do is
14:28
basically
14:29
make it possible to hone that knife edge
14:32
those 10 lines of code and deploy them
14:35
for as many different scenarios
14:37
as you would want in a way that you
14:39
don't have to constantly redevelop what
14:40
you're deploying it to
14:41
so when you say your 10 lines of code
14:44
and
14:45
are you saying you're sharing the 10
14:47
lines of code
14:48
or are you saying you're sharing the
14:50
platform that enables
14:52
lawyers to write the 10 lines you've
14:54
helped me crystallize my response
14:55
and again the response is both but now i
14:57
can actually explain that in a clear way
14:59
yes in some cases it's the one in some
15:02
cases the other
15:03
as someone who wants to deploy those 10
15:05
lines of code for my own purposes and
15:06
i'll be honest there's a selfish aspect
15:08
to this
15:08
it benefits me and it benefits everyone
15:11
to have
15:12
that infrastructure available and robust
15:15
yep so if i share that platform and it's
15:17
maintained and there's eyes on
15:19
by more than me yep it's a it's a
15:21
stronger platform
15:22
it's more mature it's more stable and
15:24
presumably people will extend it the
15:25
react of the world
15:27
right exactly exactly i mean it benefits
15:29
facebook that react
15:30
is so much more um integrated
15:33
in everything else and i think there's
15:35
an altruistic aspect to that and there's
15:37
a selfish aspect
15:38
the selfish aspect is i've dealt with
15:40
this problem enough that i know that i'm
15:41
going to have to keep dealing with it
15:42
and i don't want to do it by myself
15:45
there's that and the second piece is
15:48
let's say
15:49
that maybe i was an entrepreneur and i
15:51
had
15:52
i wanted to make it possible to use my
15:55
template
15:56
i i don't know what kind of business i'm
15:57
running maybe i'm automating forms right
16:00
and i want to have a nice data extractor
16:02
for my platform yep why should i
16:06
hey why should i write that
16:07
infrastructure and b why not make my
16:10
my spec available for free share those
16:13
two lines of code
16:14
so then everyone can use my template yep
16:16
everyone can extract data from my
16:18
template
16:19
we're all confident in it and no one has
16:20
to reinvent the wheel
16:22
and so i think there's there's that
16:24
answer is it's both
16:26
and depending on who you are it benefits
16:29
you
16:30
to do one or the other or both and i
16:32
think
16:33
there's a lot of situations like that
16:35
and if we start
16:36
collaborating and cooperating we will
16:39
all benefit
16:41
and that's what i want to see so so it's
16:43
a very altruistic
16:44
goal altruistic and selfish well
16:46
selfridge altruism
16:47
i wasn't about to call you selfish um
16:50
it's okay i can take it
16:51
i can take it this is new york after all
16:53
i've been called a lot worse actually
16:55
so it's fine well i'm still not gonna
16:58
call you that
16:59
fair to your face exactly
17:03
um the open sourcing of software that
17:06
enables the sharing of legal knowledge
17:09
yeah does that threaten lawyers and
17:12
and why or why not this
17:15
is not a problem that's unique to
17:16
lawyers this is a challenge that every
17:19
industry that's had a decision point of
17:22
do we collaborate
17:23
or do we compete or do we call um
17:26
what is the word uh coopetition
17:28
competition right
17:30
i mean at imperfectly stated but i i
17:33
think there's sort of like three
17:34
different ways you can go there right
17:36
it's we have the the closed wall
17:38
and we just don't talk to each other
17:39
there's the hey we're just gonna work
17:41
together i love you you love me like
17:43
we're all on the same side which is
17:44
clearly not the case
17:45
right and then there's something in the
17:46
middle and i think you know the question
17:48
is often is is
17:50
what is your core competency yep what
17:53
is your secret sauce and i think
17:57
you know if if there is something
18:00
inimical to the software
18:02
that you are that's making you
18:05
competitive and your competitor's not
18:06
competitive
18:08
then in that case i think you really do
18:09
have to stop and think you know is this
18:11
something i want to share
18:12
but then i have to ask in the example i
18:16
gave
18:17
what is the cost benefit if you have a
18:20
form you've got
18:21
say you have 500 copies of your form and
18:23
you have a script that can extract data
18:25
from it perfectly
18:26
right and you only have 500 examples of
18:28
it and you're not planning on starting a
18:30
data business
18:30
around your form right and you're not
18:32
planning on creating
18:34
a mc law firm where you're going to have
18:36
10 million copies and you want it to
18:37
become the standard
18:38
and you want everyone to go through you
18:41
what is the benefit of maintaining that
18:43
infrastructure to deliver that extractor
18:46
to your attorneys and i mean frankly
18:50
i really don't think that cost benefit
18:51
makes sense to do all that
18:53
maintain it yourself and eat the cost
18:57
i think in that case again being
18:58
selfishly altruistic
19:00
it's a clear situation where if you can
19:03
split that cost 10 ways
19:04
that's a benefit it's a total it's it
19:06
there's i wins all 10 people who
19:08
participate all 10 people all 10 people
19:10
and if you split it 100 ways
19:12
even better right and if it happens that
19:14
in the course of splitting that one very
19:16
specific thing
19:17
it also ends up doing five other things
19:19
that are also useful to you
19:20
yep that's a bonus right so i you have
19:23
to think carefully
19:24
but there's definitely situations
19:26
absolutely people don't really talk
19:28
about
19:29
this idea in the legal space yeah um i
19:31
think
19:32
of everyone i've met you're probably the
19:35
only one
19:36
who's who's kind of like not just
19:38
talking about it but also
19:40
doing something about it yeah well try
19:43
me part of it part of it is
19:44
i'm a damaged person and then this is
19:46
what i do for fun
19:47
so i'm like well i might as well give it
19:49
i might as well get some other people
19:50
involved
19:50
yeah um actually a correction a
19:52
correction you're not the only
19:54
i was going to say there are others i
19:56
was going to say there are others
19:57
yeah um because i mean atticus project
19:59
there's one yeah um
20:00
there's blacks blackstone is um
20:04
the lost side of england and wales just
20:05
put that together yep yep um
20:07
and there's a couple of others who've
20:10
also
20:11
given their their code or their yes data
20:14
yeah so i want to you know a couple
20:15
projects worth mentioning
20:17
um the accord project yes hybrid smart
20:20
contracts
20:21
they have speaking of react they have
20:23
some excellent ui elements
20:25
it's really impressive actually and then
20:27
the the spec they have for how to split
20:30
the logic versus the human readable
20:33
contract versus the data model yep
20:35
is really i think really well thought
20:37
out yep you know then you have
20:39
some folks who are working on things
20:41
that are
20:43
kind of a little more low-level but lex
20:46
predict
20:47
has lex nlp which is a great piece of
20:50
software yep um the library is
20:52
it's uh afro gpl licensed so perhaps
20:56
less of a great product but for anyone
20:59
who just wants to
21:00
use it and rely on it yeah amazing yes
21:03
if you want to build a business off it
21:05
that could be an issue but if you just
21:06
want to use it it's amazing it's really
21:08
good
21:08
um they also have open edgar which is
21:10
also lex predict which is another
21:12
fantastic piece of software
21:14
um lex predict another one so uh or is
21:17
it no contract suite
21:18
contract sweden contract suite yeah buy
21:20
lex predict yeah
21:22
um and then of course you have kind of
21:24
more smart contract
21:26
open source projects which are i think a
21:28
lot of people wouldn't see them as legal
21:30
but at i there is a slow
21:33
bleed over i think into legal
21:36
applications
21:37
so i put them in there too so so where
21:40
do you see yourself
21:42
fitting into this ecosystem where do you
21:43
see your kind of
21:46
well extractor would you like um
21:49
i think where i see mine fitting in is
21:53
it's sort of the swiss army knife of
21:56
automation
21:57
right so i think we talked about this
21:59
before there's sort of a lot of point
22:00
solutions and there's a lot of platforms
22:03
and building a platform that can do a
22:04
lot of different things well is very
22:06
challenging yeah
22:07
that said there are people who are doing
22:08
it and it's a lot of work
22:10
it's impressive and i'm i'm looking at
22:13
people like you
22:14
but i mean that's you know that's that's
22:16
there's a place for that
22:17
i have an amazing place for that and i
22:19
think there's that kind of product and
22:21
then for the point solutions where
22:23
there's some task that's shared across a
22:26
large number of people yep
22:27
such that it's worth building one way of
22:30
doing it for a large number of clients
22:32
it makes sense there's the donut hollow
22:35
in the middle where
22:36
you have some kind of automation project
22:38
or problem where
22:39
you can do it with code and you can
22:42
probably do it pretty easily
22:44
and you could do it on a scale that for
22:46
human would be a pain in the butt
22:47
right like you know i've got 500 or 1000
22:50
documents
22:51
no one's going to build a business off
22:52
that and it
22:54
might be one of those situations where
22:55
it's so specific that a multi-purpose
22:57
platform
22:58
probably no matter how good it is isn't
23:01
going to be suited to it
23:02
yep that's the donut hole you've got
23:05
documents coming in
23:06
and you want documents coming out or
23:07
data coming out and you can do something
23:10
relatively simply yeah with code yeah
23:13
that's where it is so i think it's sort
23:15
of maybe a
23:17
open source low code tool but
23:19
specifically for
23:21
analyzing and automating contracts so
23:23
there's a there's a term i've been
23:25
trying to
23:26
float around it's it's got zero adoption
23:28
so far nobody's using this
23:30
yeah but it's a term i call legal data
23:33
okay
23:33
um and it's
23:36
the sorts of information that exists and
23:39
matter pretty much only to lawyers yeah
23:42
so not commercial data not things like
23:46
dates numbers interest rates so on which
23:49
is
23:49
interesting to non-lawyers as well yeah
23:52
but more things like legal concepts
23:55
legal drafting
23:56
yeah specific phraseologies that have
23:59
consequences to lawyers
24:01
or meaning specific to lawyers yeah
24:04
um and it sounds like your tool strides
24:07
both of these camps yeah i think that's
24:11
right i like that term actually legal
24:12
data right yeah
24:13
yeah no i correct credits absolutely no
24:17
i legal data and data driven law like i
24:19
think
24:20
are it's it's a great concept and i
24:22
think we're going there in some capacity
24:24
right and
24:25
the the challenge becomes you need a
24:27
slightly different tool kit
24:29
and i think you need to look to data
24:30
science more than you necessarily look
24:33
to
24:33
practice or even traditional enterprise
24:35
software for
24:36
what kinds of tools do we need when we
24:38
get there yep and that's
24:40
some of the inspiration i've taken for
24:43
the some of the tools i've developed for
24:45
open source legal one of which
24:46
is the low code tool you know i think
24:48
there's some um
24:50
excellent tools in that frame for data
24:53
scientists
24:54
and there's sort of a similar design
24:55
similar architecture where you can
24:56
design data flows
24:58
and you kind of have these very discreet
24:59
little steps that that
25:01
analyze and extract data and it's the
25:03
same thing with what i'm trying to do
25:05
but it's
25:06
designed around the idea that you want
25:07
to work on a document
25:09
contract and it also has sort of a
25:11
self-service aspect where
25:13
once you've developed this you can make
25:14
it available to someone
25:16
who is a non-technical user who can just
25:18
go in and pick from a menu
25:19
and say you know what do i want to do i
25:21
want to extract
25:23
concepts i want to extract certain
25:25
clauses from this type of document click
25:27
add a document play done yep that's
25:30
that's the goal and it's a way
25:32
it's almost an interface to sort of hide
25:34
the data science yes
25:35
from the lawyers yes but still give them
25:38
the ability to work with them well it
25:39
doesn't matter to them right yeah
25:41
exactly i was watching this great google
25:44
uh lecture it's by a lady who um i
25:48
forget her name which is terrible but
25:49
she explained how
25:50
for data scientists it's like you want
25:52
to know how the microwave works
25:54
yeah but for everyone else you just want
25:57
to get your hot food
25:58
yeah yeah exactly well exactly it's like
26:01
what i said before
26:02
you know you can do magic with 10 lines
26:04
of code but an attorney doesn't
26:05
want to see the 10 lines of code no they
26:07
want the magic they want the magic
26:09
and there's we can give it to them so
26:11
let's give it to them you know yep
26:12
and and i think again in a lot of
26:14
situations
26:16
the data scientists can build the tool
26:19
the lawyer could use the tool but the
26:21
tooling to bring the two together
26:23
doesn't exist and i frankly i don't
26:26
think
26:27
and i could be wrong and maybe i'll
26:29
regret this
26:30
but i don't think that piece is
26:33
necessarily
26:34
a business or if it is i think there's a
26:36
much more transformative
26:38
role that we all can play as legal
26:40
innovators in making
26:42
that infrastructure robust and available
26:44
across the industry
26:45
and then just doing the magic yeah
26:47
that's where the cool stuff happens
26:49
you need people as well as the
26:52
technology
26:52
yes i think that's your greatest
26:55
challenge right how do you
26:57
get people inside of law firms
27:00
yeah to go i know how to write those 10
27:02
magic lines right
27:03
right that that's it's that so that's
27:06
part of the reason why i think
27:07
open source also works really well here
27:09
yes and i think it's
27:12
law firms as a general rule there are
27:14
exceptions but
27:15
most of them don't have data scientists
27:18
or
27:18
if they do they have a couple right and
27:21
i think
27:21
overwhelmed they're overwhelmed right
27:23
and i think you know again looking back
27:25
to the early days of open source
27:27
software a lot of this stuff
27:28
that today is world changing started out
27:31
as
27:32
a hobby project it started out as
27:34
someone had an idea
27:35
and it was it needed a lot of work and
27:38
needed a lot of people
27:39
and if you told someone in 1992 you know
27:43
linux is going to run the world
27:45
right i mean maybe maybe linus stravals
27:48
would have said well
27:49
you know linus excuse me i would have
27:51
said yeah absolutely
27:53
but i think no very few other people
27:55
would have agreed
27:56
no yeah people would be like you're
27:58
crazy microsoft is going to dominate
28:00
forever
28:01
today 2021 desktop linux while i use it
28:05
is not it's not it's not mainstream
28:08
and frankly it might never be mainstream
28:10
however it dominates the server market
28:13
it dominates handsets yep android is
28:16
linux
28:17
i mean it's linux plus some stuff but
28:19
it's linux
28:20
and again i think we go back a lot of
28:22
the stuff that today
28:23
is world changing started out as a hobby
28:26
project
28:27
or a meeting of the mines that at the
28:29
time
28:30
no one would and put money into because
28:32
it was well maybe
28:33
one day this will be useful would you
28:36
give me some money please
28:38
and no one is going to give you money
28:39
for that yeah but
28:41
there is a benefit in having
28:43
infrastructure
28:44
that is non-monetary there's you know
28:48
kinship friendship uh there's just
28:51
simply because you like it it's fun it
28:53
makes you feel good makes you feel good
28:54
there's that
28:55
there's reputational and i do think long
28:57
term if we
28:59
want the industry to move we need the
29:01
tooling to move it
29:03
and contributing some things
29:07
to the industry altruistically i think
29:09
also
29:10
can if if your angle is personal benefit
29:13
i think there's that
29:15
and i think there's again there's also
29:16
just
29:18
we can we can do cool stuff and if we
29:21
make this stuff happen if we interest
29:22
people yep
29:24
we can do that so to your point about
29:25
how do you get the people
29:28
try to open source stuff get people who
29:30
have an inclination
29:32
who want to play around with this stuff
29:33
say look we're both going to do this
29:35
anyway
29:36
so i've done some of it you do other
29:38
pieces and now you know we can both do
29:40
the fun stuff
29:41
it's it's like putting a pie on a window
29:43
scene and then like
29:44
just letting the smell catch the
29:46
audience yes
29:47
yes exactly exactly that is really cool
29:50
sean
29:50
i am so grateful you make the time to
29:53
have a chat i know i haven't given an
29:54
intro
29:55
oh yeah so i'm going to give the intro
29:57
at the end and then cut it back to the
29:59
beginning
