Re:Show HN: Stripe Atlas with a Lawyer
by anonbanker on Friday February 26, 2016 @07:09PM
(#11185307)
This is the first stage: lawyers are now going to the cloud for their tasks. Legal offices can now be run without a physical office. Lawyers can potentially deal with more cases. It's blue-ocean strategy that just found a way to disrupt a market; Legal assistants can be reduced (or eliminated), Eventually, the need for paralegals will disappear (or pivot, hopefully). Then even Lawyers will be eaten by software.
I give it fifteen years before Judges are hearing arguments from Watson (or [hopefully free software] equivalents) more than they're hearing humans.
I'm looking forward to running process on/through a judge that's software-based, myself, as even their jobs will be swallowed soon enough. Judges are pretty much compilers for legal processes. it'd be interesting to deal with a GNU one.
Re:Show HN: Stripe Atlas with a Lawyer
by hobbyjogger on Tuesday March 01, 2016 @10:42AM
(#11203315)
Meh, not so fast. Here's the Stanford CS professor who runs the Stanford Center for Computers and Law[1]:
"One technical problem with Computational Law, familiar to many individual with legal training, is due to the open texture of laws. Consider a municipal regulation stating "No vehicles in the park". On first blush this is fine, but it is really quite problematic. Just what constitutes a vehicle? Is a bicycle a vehicle? What about a skateboard? How about roller skates? What about a baby stroller? A horse? A repair vehicle? For that matter, what is the park? At what altitude does it end? If a helicopter hovers at 10 feet, is that a violation? What if it flies over at 100 feet?
The resolution of this problem is to limit the application of Computational Law to those cases where such issues can be externalized or marginalized. We allow human users to make judgments about such open texture concepts in entering data or we avoid regulatory applications where such concepts abound.
A different sort of challenge to Computational Law stems from the fact that not all legal reasoning is deductive. Edwina Rissland [Rissland et al.] notes that, "Law is not a matter of simply applying rules to facts via modus ponens"; and, when regarding the broad application of AI techniques to law, this is certainly true. The rules that apply to a real-world situation, as well as even the facts themselves, may be open to interpretation, and many legal decisions are made through case-based reasoning, bypassing explicit reasoning about laws and statutes. The general problem of open texture when interpreting rules, along with the parallel problem of running out of rules to apply when resolving terms, presents significant obstacles to implementable automated rule-based reasoning."
[1] http://logic.stanford.edu/complaw/complaw.html
Re:Show HN: Stripe Atlas with a Lawyer
by anonbanker on Tuesday March 01, 2016 @04:26PM
(#11206210)
What you state is essentially banking on the limitations of AI programming, on the hopes that what keeps us "human", and therefore, important as a cog in the machine of law/commerce, will never be discovered or replaced.
That might take longer than 15 years, but it's definitely within our lifetimes. and it'll be a beauty to behold.
Re:Show HN: Stripe Atlas with a Lawyer
by hobbyjogger on Tuesday March 01, 2016 @07:45PM
(#11207395)
I didn't state anything. That's a direct quote by the guy who runs what is, by leaps and bounds, the most advanced project applying AI to legal analysis.
They're doing some incredible (and incredibly interesting) things, but even they don't see AI taking over courtrooms any time soon. Think about the two problems Professor Genesereth noted above (even in the context of the most laughably simple example imaginable). These problems aren't just of a different magnitude compared to the problems we are working on with AI right now--they're of a different character entirely.
Do you know of any AI programs that can generate analogies to every possible thing/situation/event and choose the most "well-suited" among those alternatives based on a holistic, "all important things considered" balancing approach? Do you know of any AI programs that are able to rank human values/morals in a heirarchy, in a way that would be acceptable all (or even most) of us? Until the answer to those questions is "yes", you're not going to see AI sitting behind the bench in an American courtroom.
A brief aside: the common law is used as a lazy evaluation strategy[0], whereby we can defer filling in the full content of any idea/word/phrase/whatever until it actually becomes a problem. So the legislature can pass a law "no pets allowed" instead of spelling out a full and comprehensive list of every imaginable item to be banned: "No Mesocricetus auratus [golden hamster], Mesocricetus brandti [Turkish hamster], Mesocricetus newtoni: [Romanian hamster]...and so on for the next 10,000 pages."
AI could easily handle some of this filling in of content but not so much items like "reasonable search" or "malicious intent" or even, apparently, "vehicle." Those types of decisions often fall, not to logical deduction, but to induction from incomplete, conflicting and messy human theories about goodness and justice and truth. Just as most wouldn't outsource their decisions about who to love or what god to believe in, they're not going to let AIs make these types of decisions (assuming, of course, that AIs are even eventually capable of doing so--which remains an open question).
Re:Show HN: Stripe Atlas with a Lawyer
by dragonwriter on Tuesday March 01, 2016 @08:13PM
(#11207501)
Quite simply, there is no consensus on the right rules to use to apply the existing body of case, statute, Constitutional, etc., law -- which is why judicial appointments are a contentious, political thing.
Without such a consensus, even if all the AI-related problems are solved we couldn't design AI judges -- because the requirements would not be well defined.
Re:Show HN: Stripe Atlas with a Lawyer
by justinlaing on Friday February 26, 2016 @12:00PM
(#11182442)
You should add more information to your site. More about what exactly you'll do. Comparison to other ways people get these things done. Maybe some testimonials.
You need to instill an incredible amount of trust for someone to sign up for this type of service online. I don't think your website is quite there yet.
Stripe has an incredible brand that many people already trust because they are using it as the way they get paid.
You have a high bar to get over to rival that type of trust as an unknown company.
Re:Show HN: Stripe Atlas with a Lawyer
by will_brown on Friday February 26, 2016 @12:59PM
(#11182854)
I will implement testimonials and comparisons. As a lawyer the trust has always been inherent with clients, but this is also why my reputation and reach is limited geographically and Stripe Atlas can provide this as a company with instant credibility globally. It is like Oz, I have performed this exact transactions for 100's of corporate clients since 2010 and they haven't even begun, but I can't compete with their credibility.
At the end of the day I could probably commoditize the service more:
-Package: New DE C-Corp; Business Bank Account; Stripe Account...and promise turn around in a few days.
But that is not the reality. Each client/customer is unique. Example, when all the founders are non US citizens? Forget a few days, in the most extreme cases I have seen it take months, and I think Stripe Atlas will probably focus on that founder make up. Or if one DE C-Corp needs to be qualified to do business in CA or another State and that aspect is neglected delaying opening the bank account or worse setting the stage to pierce the corporate veil. Some may want an LLC or an account with a different bank, so instead I just have clients contact me, we schedule a meeting (in-person, teleconference), the client speaks directly with the lawyer (me), and in I develop/propose a customized path with a cost budget and timeline...that's my current process.
I have sent an email to someone at Stripe Atlas from the original HN thread, I have asked if I could independently provide my personalized services to a sample of the first 100 free Stripe Atlas customers. Allow Stripe Atlas to judge their customer feedback for themselves and maybe they would wish to bring me aboard.
I give it fifteen years before Judges are hearing arguments from Watson (or [hopefully free software] equivalents) more than they're hearing humans.
I'm looking forward to running process on/through a judge that's software-based, myself, as even their jobs will be swallowed soon enough. Judges are pretty much compilers for legal processes. it'd be interesting to deal with a GNU one.
"One technical problem with Computational Law, familiar to many individual with legal training, is due to the open texture of laws. Consider a municipal regulation stating "No vehicles in the park". On first blush this is fine, but it is really quite problematic. Just what constitutes a vehicle? Is a bicycle a vehicle? What about a skateboard? How about roller skates? What about a baby stroller? A horse? A repair vehicle? For that matter, what is the park? At what altitude does it end? If a helicopter hovers at 10 feet, is that a violation? What if it flies over at 100 feet? The resolution of this problem is to limit the application of Computational Law to those cases where such issues can be externalized or marginalized. We allow human users to make judgments about such open texture concepts in entering data or we avoid regulatory applications where such concepts abound. A different sort of challenge to Computational Law stems from the fact that not all legal reasoning is deductive. Edwina Rissland [Rissland et al.] notes that, "Law is not a matter of simply applying rules to facts via modus ponens"; and, when regarding the broad application of AI techniques to law, this is certainly true. The rules that apply to a real-world situation, as well as even the facts themselves, may be open to interpretation, and many legal decisions are made through case-based reasoning, bypassing explicit reasoning about laws and statutes. The general problem of open texture when interpreting rules, along with the parallel problem of running out of rules to apply when resolving terms, presents significant obstacles to implementable automated rule-based reasoning." [1] http://logic.stanford.edu/complaw/complaw.html
That might take longer than 15 years, but it's definitely within our lifetimes. and it'll be a beauty to behold.
They're doing some incredible (and incredibly interesting) things, but even they don't see AI taking over courtrooms any time soon. Think about the two problems Professor Genesereth noted above (even in the context of the most laughably simple example imaginable). These problems aren't just of a different magnitude compared to the problems we are working on with AI right now--they're of a different character entirely.
Do you know of any AI programs that can generate analogies to every possible thing/situation/event and choose the most "well-suited" among those alternatives based on a holistic, "all important things considered" balancing approach? Do you know of any AI programs that are able to rank human values/morals in a heirarchy, in a way that would be acceptable all (or even most) of us? Until the answer to those questions is "yes", you're not going to see AI sitting behind the bench in an American courtroom.
A brief aside: the common law is used as a lazy evaluation strategy[0], whereby we can defer filling in the full content of any idea/word/phrase/whatever until it actually becomes a problem. So the legislature can pass a law "no pets allowed" instead of spelling out a full and comprehensive list of every imaginable item to be banned: "No Mesocricetus auratus [golden hamster], Mesocricetus brandti [Turkish hamster], Mesocricetus newtoni: [Romanian hamster]...and so on for the next 10,000 pages."
AI could easily handle some of this filling in of content but not so much items like "reasonable search" or "malicious intent" or even, apparently, "vehicle." Those types of decisions often fall, not to logical deduction, but to induction from incomplete, conflicting and messy human theories about goodness and justice and truth. Just as most wouldn't outsource their decisions about who to love or what god to believe in, they're not going to let AIs make these types of decisions (assuming, of course, that AIs are even eventually capable of doing so--which remains an open question).
[0] https://en.wikipedia.org/wiki/Lazy_evaluation
Without such a consensus, even if all the AI-related problems are solved we couldn't design AI judges -- because the requirements would not be well defined.
You need to instill an incredible amount of trust for someone to sign up for this type of service online. I don't think your website is quite there yet.
Stripe has an incredible brand that many people already trust because they are using it as the way they get paid.
You have a high bar to get over to rival that type of trust as an unknown company.
At the end of the day I could probably commoditize the service more:
-Package: New DE C-Corp; Business Bank Account; Stripe Account...and promise turn around in a few days.
But that is not the reality. Each client/customer is unique. Example, when all the founders are non US citizens? Forget a few days, in the most extreme cases I have seen it take months, and I think Stripe Atlas will probably focus on that founder make up. Or if one DE C-Corp needs to be qualified to do business in CA or another State and that aspect is neglected delaying opening the bank account or worse setting the stage to pierce the corporate veil. Some may want an LLC or an account with a different bank, so instead I just have clients contact me, we schedule a meeting (in-person, teleconference), the client speaks directly with the lawyer (me), and in I develop/propose a customized path with a cost budget and timeline...that's my current process.
I have sent an email to someone at Stripe Atlas from the original HN thread, I have asked if I could independently provide my personalized services to a sample of the first 100 free Stripe Atlas customers. Allow Stripe Atlas to judge their customer feedback for themselves and maybe they would wish to bring me aboard.