51–60 of 82 posts
I think it's already been pointed out but I'll clarify a little more. You've hired an independent developer, at a reasonable rate. Hiring a lawyer from a law firm, comes at the rate of the firm. He bring to the table the resources of a large law firm, and along with the cost of operating a law firm. In addition to his nice salary. If you had hired an tech consulting firm you'd be looking a similar hourly rate if you…
thisThere exist at least some software developers who bill $700 an hour or north of that, by the way.
Is it feasible to program a machine learning system (such as Watson?) to run a straightforward legal case, with a natural language interface? Presumably the majority of "bread and butter" legal cases don't require an innovative step? Rather, they involve rolling out a set of standard arguments and countering a set of standard arguments put forward by the opposition, a bit like a chess game? I'm talking straighforward…
Is it feasible to program a machine learning system (such as Watson?) to run a straightforward legal case, with a natural language interface? The law is full of context and interpretation details. One example my business constantly runs into is weather or not we have to register as a 'foreign business" when we do business in other states (my business is an engineering consulting business that does work in many states…
> for some states, you only have to register is you maintain an office presence; for some states you have to register only if you spend a significant amount of time in the state, [...]
then this is doable, in principle. But the overall project and required workload would indeed be vast.. I wonder how much one might have to work to have some kind of a working PoC.. NLTK and their computational linguistics book / python tutorial might be a good start. e.g. http://nltk.org/book/ch10.html - we can do things like
dt = nltk.DiscourseTester(['A student dances', 'Every student is a person'])
dt.add_sentence('No person dances', consistchk=True)
- throws error 'Inconsistent discourse ..[list of predicates etc]'
dt.retract_sentence('No person dances', verbose=True)
- outputs
Current sentences are
s0: A student dances
s1: Every student is a person
But scaling this / making it work for something more than these simple linguistic tasks.. quite an undertaking! In any case, the field in question (and the task at hand) is very interesting, and I hope that lots of interesting stuff will be done in comp.linguistics / nlp in the near future.
I truly hope he is not a great iOS developer because that rate is criminally low. There is something magical that happens in the 100 pennies between 99 per hour and 100 per hour, a mental chasm that is in reality no different than 98/hr vs 99/hr but we make it into a big deal. Why he chose that number, and why people choose any number for an hourly rate is always a curious thing to me. We pick numbers that we think a…
I'm a top Android developer from the Czech Republic and usually charge about $25/hour.
Now suppose you hired Ramsay, but instead of a dinner party it was for your kid's birthday party--and all you want Ramsay to do is boil Oscar Meyer hot dogs and serve then on Safeway hot dog buns with Heinz Ketchup (and only that) on them, and later scoop some store-bought ice cream onto store-bought cake.
This is something you could hire a neighborhood teen to do for $20. Do you think Ramsay is going to do it for $20, because you are asking him to do something much simpler than make his signature food?
Of course not. He's charging for his talent and skill. If you want to misuse that by having him work on something that is trivial and can be done easily by someone with much less talent or skill, that's your problem.
I suspect that's what you are doing with the lawyer. A partner in a major firm has the skill and talent to work on very large and complicated matters that involve big companies and a lot of money--things like major mergers and acquisitions, IPOs, and such.
You are asking him for something that could probably be drafted by a paralegal from his office, and then checked over by a junior associate--the legal equivalent of boiling hot dogs.
I've known network consultants who charged banks and other Wall St firms $2600/hr for troubleshooting. A network outage could easily cost some of these firms millions of dollars per hour, so paying $2600/hr to fix such an outage is petty cash in comparison.
Hire the $2600/hr guy and be done in ten minutes because of his fifteen years experience and deep knowledge of the system; spend $433 very quickly. Pick the $50/hr rent-an-intern and you're down for a day while they're googling the fixes. Spend about the same amount, but avoid the losses from 7+ extra hours of downtime.
Time is money, after all.
Same thing with an after-hours/emergency call to a plumber. Yes, you'll spend twice as much - but that's nothing compared to what it will cost to repair the structural damage to your house because your toilet overflowed and you ended up with three inches of standing water soaking into your everything.
Is it feasible to program a machine learning system (such as Watson?) to run a straightforward legal case, with a natural language interface? Presumably the majority of "bread and butter" legal cases don't require an innovative step? Rather, they involve rolling out a set of standard arguments and countering a set of standard arguments put forward by the opposition, a bit like a chess game? I'm talking straighforward…
Is it feasible to program a machine learning system (such as Watson?) to run a straightforward legal case, with a natural language interface? The law is full of context and interpretation details. One example my business constantly runs into is weather or not we have to register as a 'foreign business" when we do business in other states (my business is an engineering consulting business that does work in many states…
Like Jeopardy, the task involves a lot of "soft" information, but results in an easy to measure pass/fail decision. It could be run in real-time, as an advocate in a courtroom, or off-line as a legal adviser. Also like Jeopardy, it offers the opportunity of a public contest, and the associated PR.
[1] http://www.research.ibm.com/deepqa/faq.shtml
---
Edit:
It seems like at least one person inside IBM is already thinking this way. To quote [2]:
"Law, Patent and Trademarks"
"Not only could a Watson-like capability minimize filing through existing databases on laws, prior cases, rulings, hearings, opinions, it could also be used as a method of testing witness questions. or suggest a series of inquiries and questions for litigation. It could be used to simulate certain judges, prosecution and defense lawyers, based on prior cases."
"A Watson-like system could generate questions for a prospective patent claim based on it's ingestion of the entire patent and trademark database."
[2] https://www-304.ibm.com/connections/blogs/davidian/entry/whe...
Prices are of course about supply and demand. The supply of business lawyers is limited, which keeps prices (relatively) high. What keeps supply limited? In practice, bar admission requirements are not the operative bottleneck. There are about twice as many JD's each year as there are jobs for them, and the prices for entry-level legal work (across the whole industry) have fallen far below that of entry-level enginee…
Prices are of course about supply and demand. The supply of business lawyers is limited, which keeps prices (relatively) high. What keeps supply limited? In practice, bar admission requirements are not the operative bottleneck. There are about twice as many JD's each year as there are jobs for them, and the prices for entry-level legal work (across the whole industry) have fallen far below that of entry-level enginee…
Not everyone gets the most expensive product or service. Shouldn't people be able to get cheaper lawyers? Why don't some of the unemployed lawyers create budget practices? No person or business would be willing to use them?
E.g. see: http://www.bizjournals.com/boston/print-edition/2012/04/27/l.... The firm in the article is a 135 lawyer outfit (so solidly on the larger end of mid-sized) with offices in Boston and D.C. and four other Northeastern cities. Partners bill out between $300-$500. (No affiliation, just Google-ed).
Experienced solo practitioners serving individuals in personal matters might charge $100-$150. And if you just need someone with a JD, you can get one for $10/hour on Craigslist.
Then on the other end of the spectrum, you have firms like Wachtell Lipton, who basically only do billion dollar M&A deals. They don't even charge a fee, they charge a percentage of the deal value. People hire them because when you've got a $5 billion deal, you don't care if your legal fees end up in the tens of millions of dollars.