What with the recent discovery that VW has been using software to cheat on emissions tests, there has been a sudden and widespread conversation taking place on how we can interrogate algorithms.
In an New York Times op-ed from yesterday, Zeynep Tufekci weighed in on both the VW scandal and another recent software problem of public interest, namely voting machines. She concludes that “…the public can’t always know if the device is working properly — but we can check its operation by creating auditable and hard-to-tamper-with logs of how the software is running that regulators can inspect.” She also notes that slot machines in casinos have regular such inspections, so it’s not impossible.
Another New York Times article profiles Columbia Law professor Eben Moglen, quoted as saying that “proprietary software is an unsafe building material,” because “you can’t inspect it.” That was in 2010. Ironically, the article explained, the reason automobile manufacturers gave for not allowing inspection is that individuals would set up their cars to cheat on admissions tests. Of course, that doesn’t explain why you wouldn’t open up the algorithms at least to regulators.
The inspection of algorithms is a concept that’s probably new to a lot of people, first because algorithms are marketed as “objective” and “fair,” second because they are almost by construction too complicated for an average person to understand.
But, as we’ve seen in this example, those are simply not good enough reasons not to do it anyway. There’s a trade-off when we take advantage of automation and algorithms: we get efficiency and scale, on the one hand, and on the other we lose control. In fact, we don’t really know what’s happening and when.
The very least we could do is ask them.
Do you guys remember the event I went to last week? It was the kick-off event for a new group at Harvard called Gender Inclusivity in Mathematics (GIIS), and it went well. It was written up in The Crimson, together with an action shot of me and Moon giving out tough love advice:
One of the organizers, Cherie Hu, wrote a blog yesterday inspired by last week’s event which ponders identity and mathematics.
Also! I was please to be featured in a story that ran in Le Monde last week with my friend and mathematical colleague Leila Schneps, who recently wrote a book with her daughter about bad math in courts systems. It’s in French, obviously, but if I remember my schooling correctly I am quoted in the article talking about destructive algorithms like the Value-Added Model for teachers and even political micro-targeting. No picture there though.
This is a guest post by Ronald Sinai, the founder and CEO of Nova Legal Funding, a national lawsuit funding company based in Los Angeles. Prior to entering the legal finance field, Ron was a student at the University of California, Berkeley where he earned a Bachelor’s degree in 2014.
Personal injury law is a big business in the United States. With every traffic accident or slip and fall, a ‘lawsuit economy’ emerges in expectation for the looming monetary compensation for the plaintiff. Attorneys, medical treatment centers and litigation service providers make up the bulk of this network. This post will tackle the latest, fastest growing and most disruptive industry to enter the lawsuit economy: pre-settlement funding companies.
Pre-settlement funding is a financial lifeline for plaintiffs involved in personal injury litigation. It’s a cash advance on the future proceeds of a settlement for people who can’t wait years for their cases to finalize. Plaintiffs often use the advance to pay for living necessities, medical bills and other immediate financial obligations. Repayment to the funder is wholly contingent on the case being settled out of court or won in trial. The plaintiff repays nothing if the case is lost, making it a risky non-recourse investment for the funding company.
As a principal in a funding company, I felt obliged to contact Cathy after hearing her speak on Slate.com’s Money Podcast and reading her negative blog post about my industry. While it may surprise you, my intent in reaching out was not to correct her. In fact, her worries are completely valid and for good reason. Little-to-no oversight by regulators has allowed bad players in the legal finance industry to employ business models that place profits over ethics. Over time, such practices justifiably resulted in a bad name for the industry, with some comparing this otherwise justice-equalizing tool with payday loans.
It’s important that we don’t throw the baby out with the bathwater. Simply because a lack of regulation led to a swarm of bad actors doesn’t mean pre-settlement funding should go away. When done correctly, embracing the fundamental benefits of lawsuit funding can help our justice system become more equitable and accessible to everyone.
On the value of plaintiff funding
Funding does more than ensure fair and complete compensation for the injured—it ensures that our justice system is blind to an individual’s economic standing.
Pre-settlement funding empowers small plaintiffs against big insurance companies. Just as attorneys litigate and treatment centers heal, the funder adds value to the case by granting the plaintiff financial stability. Solid financial footing helps them reject early lowball settlement offers from insurance companies, who always seek to take advantage of a person’s vulnerable economic position.
No matter how obvious the negligence or big the damages, insurance companies always delay compensation. It’s the oldest and most effective trick in the book: gain leverage by inducing desperation. The adjuster capitalizes on this desperation by offering the plaintiff a lowball offer in exchange for a quick and early payout. Funding the plaintiff takes the leverage away from the adjuster, which results in a big win for the small guy.
On the problem with plaintiff funding
No matter how great the premise of consumer legal finance might be, a lack of oversight will continue to allow bad players to charge unreasonable rates and make the service abusive rather than valuable.
As an operator in the space, I know what deals are being made and under what terms. I’ve seen countless of funding agreements from dozens of companies since the inception of my business. Plaintiffs who previously received funding from other sources come to us in hopes of refinancing their expensive paper. I don’t have hard data, but the average rate seems to be 3-4% compounded monthly or 35-40% every six month period. That’s not including a possible broker fee (10-20% of funding) and an application fee of $250-$400.
Even worse, most companies charge rates that are uncorrelated with the risk profile of each individual case. In other words, a litigious person with a questionable slip and fall at Wal Mart will get the same rate as a victim who was rear-ended while stopped at a red light.
The “fund-everything-at-high-rates” business model
Companies that charge high rates have the luxury of relying less on proper underwriting than they do on volume. This business model is attractive for many reasons. First, higher rates makes it easier to swallow loser cases, which reduces underwriting requirements and drives an increase in deal flow. Relaxed underwriting also means hiring less attorneys, which leads to a massive reduction in overhead.
Secondly, a lack of proper underwriting makes the process hassle-free for the plaintiff’s busy attorney, who wants nothing more than to get the funding process over with. Believe it or not, attorneys whose clients bug them enough for cash will take a quicker funding process over lower rates every time. Same goes for the plaintiff: they like what is fast, not what is affordable. They only realize the mistake when it comes time to repay the funder.
The fix? Enforce rate caps, force careful underwriting.
Not all funders behave recklessly. At my firm, for example, we offer a fixed payoff schedule for 10-15% each consecutive 6-month period. At these rates, our underwriting has to be lock-tight in order to turn a profit. A call with the attorney needs to be arranged, all liens must be inspected, and a variety of documents must be submitted by the law firm before a deal is made.
This forces us to only fund meritorious claims where plaintiffs suffer serious damages, and negligence is clear. By this logic, rate caps will do more than stop unreasonable rates. They will also make it impossible for frivolous lawsuits to get funding and hurt our economy.
Readers, Aunt Pythia is quite pleased with herself this morning. She has come up with an amazing solution to the problem of teenage dissipation and slovenliness.
Now, don’t get Aunt Pythia wrong: she’s got some amazing teenagers. They even do their own laundry, and take turns doing the dishes (when prompted!). But one thing they haven’t been able to do, no matter the level of coaxing, is to put away their clean clothes in their dresser. What invariably happens is they put their clean clothes in a bag, which gets turned over onto the floor in the following morning’s search for a clean sock.
Bottomline: their floors are always entirely covered with clothes.
Solution: get rid of their dressers altogether and replace them with a large “clean laundry” bin. These are the bins I bought which have just been delivered:
Strangely enough, their father doesn’t seem as excited as Aunt Pythia about the “clean laundry bin”. Something about the aesthetics, or the size. His tune will change when there’s no laundry on the floor, though, I assure you. I promise to update you on this miraculous cure to all things slipshod and/or lackadaisical.
OK, on with the advice! And after you enjoy said advice, please:
ask Aunt Pythia any question at all at the bottom of the page!
Dear Aunt Pythia,
I’m stuck in an interpersonal pickle and I need some insight from someone totally removed from the situation. Most of the time I have a pretty strong moral intuition but this has me at a loss.
I’ve known this woman, “Beth,” since high school. She has always been a difficult person to be a friend to, and I think I’m reaching my limit. (We’re both in our early 30s now, to give you an idea of the timeline.) Beth is difficult because she is a self-centered person, which is exacerbated by mental illness. Beth has been on medication for OCD since high school and for bipolar disorder since college.
While she is currently seeing a psychiatrist, she definitely never visited a mental health professional in high school and probably didn’t in college, either. According to her, she “diagnosed herself” with OCD, depression, and bipolar, then talked to her GP (a close friend of the family), who agreed with her assessment and wrote out the prescriptions. I don’t know how prevalent this kind of “self diagnosis” is, but I think this part of her background is relevant, so I’m including it.
For what it’s worth, I don’t doubt for a moment that she suffers from mental illness. I just worry that she is getting the wrong treatment, since she doesn’t seem any “better” after ten years of this particular cocktail of medication. (But I haven’t said any of this to her, and wouldn’t dare, because IMO that would be presumptuous and maybe she’s coming off worse online than IRL. That’s the job of a mental health professional.)
At the moment I am one of two people she talks to who aren’t her family (husband, in-laws, mother), her psychiatrist, or her current lover. (She has been having an affair for almost a year; this is not an open/”monogamousish” marriage.) I feel morally obligated to remain in her life to at least some degree, since I imagine she is probably very lonely, especially since she is in the middle of an argument with her only other friend. This “only two friends” situation is also something she’s told me; I’m not making any suppositions here. Otherwise, I would have cut ties a while ago.
I don’t like the person I become when I talk to her and I don’t think I have the right skillset or knowledge to help her. The only thing that happens as a result of our conversations is that she gives me minute-by-minute updates on her moods/activities, trash talks her husband, relates the sexcapades she’s having with her lover, and asks me for advice that she doesn’t follow. Occasionally she shares random news link with a few throwaway comments on them, and once in a while she asks me what I’m doing, but after a few lines of conversation everything is back to her.
Most people I think I could say, “I want to support you, but I’ve got a lot of stuff going on my self and it’s taking all of my cope just to deal with that. I’ll let you know when I’m feeling better.” or “You know, you tell me a lot about what’s going on with you, but you don’t seem to be displaying any interest in my life. I know that you care, of course, but it would be nice if you could show me that you do.” and, while it would sting, they would be able to handle it. But she is fragile enough that I think even that would crush her, considering that she is angry at her only other friend for essentially saying just that.
The silver lining in all of this is that I am hundreds miles of way and will remain there for the rest of my life, so I only have to interact with Beth online. At the moment I am basically checked out. I’ve limited myself to blase responses like “that sounds annoying” or “that’s good” to most things and outright ignoring what I think is the most harmful/unhealthy stuff she says, or the things that sound like a bid for attention or validation. Is this the best I can do? Should I tell her I need some alone time (or full-on ghost her) and reduce her social outlets by half? Am I overestimating my own importance? Am I underestimating her resilience? Am I making myself a martyr?
Thank you for your input.
A few things. First, sympathy: your friend sounds really hard to deal with, and it’s kind of you to stick with her.
Second, I agree that she sounds like she has real problems, and I’m no professional so I wouldn’t hazard a guess what her problem is, but I’d suggest you spend some time looking at personality disorder profiles. I say that because it has helped me enormously in the past; when you encounter someone with a personality disorder, you feel bewildered and confused – and sometimes even partially responsible to help – but then, reading about the disorders, and the support groups for people who are married to people with them, you realize that you are not alone in your confusion, and that you are not capable of curing them.
Finally, advice. You are at risk of getting so fed up with your friend that you leave her entirely. Instead of letting your last ounce of true goodwill drip out of you slowly, I suggest you tell her about the difficulties you’re having, and asking for her help to remain friends, while you still can do it. Too often, people only express frustration at the point of no return, so the underlying message is, “you cannot convince me to be your friend anymore, it’s too late.” I would love to see your message be something more like, “you need to be a friend to me as well or else you’ll lose me.” It’s a much kinder message.
So, if you can do it, tell her truthfully what’s frustrating you, and be sure to tell her that you still want to be friends, and see what happens. In other words, don’t be a martyr, and don’t underestimate her resilience. If she cannot hear you, and gets upset and refuses to talk, then wait a few months or a year or two and get back in touch, because people often need time to recover, and their disorders often oscillate in terms of severity. Above all, keep careful track of what you’re thinking and doing versus what she accuses you of thinking and doing, because you’ll need to stay calm and reasonable, and that might be hard, but it’s what a good friend does.
Dear Aunt Pythia,
I have been close friends with a guy from undergrad for six years. We met my freshman year and became best (platonic) friends that year. He was dating a girl from his hometown, but they were never very close. He felt obligated to stay with her for intense family reasons, but the emotional bond between them was minimal. They fought often and had very little in common. There was no sexual relationship.
They broke up during our sophomore year and he approached me about starting a relationship. I was in a bad place and was not ready to be in a relationship. They got back together about six months later.
We remained very close throughout college – ran together, studied together, went backpacking together. We both told each other everything. I thought that we were really just friends, and that the people who thought we were dating or should date were reading into things (professors, friends, etc. frequently assumed we were).
After graduation, we remained very very close and he remained dating his girlfriend, still under strict family pressure. I realized after we both graduated that I was in love with him. I was/am very physically attracted to him and emotionally bonded with him. I didn’t say anything to him. We both started doctoral programs in New England (in the sameish field) and are both two years in. We don’t see each other much (about every 2 months), but talk on the phone once a week, write, and text often.
They broke up about four months ago and I’m at a loss of what to do. They definitely won’t be getting back together, but at this point, I’ve lived in stagnation for so long that I’m afraid to tell him. I don’t want to lose my best friend, and the long wait has left me more scared than ever. I don’t even know if I want to tell him. What do I do? Help me, Aunt Pythia! I dreamed of this for so long, but now I don’t know what to do.
Perplexed and Frozen
OK, so two comments. First, nobody writes to Aunt Pythia so that she can say, “don’t go for it, it’s a trap!”. That doesn’t happen. So obviously what you’re looking for here is the green light. They don’t call me Aunt “Go For It” Pythia for nothin’.
Second, I’ma give you the green light here. Not necessarily because I think it will work out – although it well might! – but mostly because I need you to move the fuck on. Holy crap, lady, you gotta get your love life moving here, and it’s been according to my calculations 6 years of this platonic friend crap at least. You didn’t mention how many love affairs you’ve been having on the side in the meantime, so I’m going to imagine at least a few, but jeez. How can you be so patient?!
As for my advice, it’s the oldest and simplest plan in the book. Invite your friend to stay with you for the weekend, get everyone out of the area with strict instructions never to return, and drink a ton of booze. Easy peasy lemon squeezy, with an emphasis on the lemon squeezy. And please do it quick, my patience is completely worn out. And then please write back and tell me what happened.
Dear Aunt Pythia,
I’m a liberal arts/sciences undergrad focused on the obstacles to just climate and economic policies. I’m also interested in economics/finance, the political process, and social justice, among other things.
I want to get work experience related to my interests before I graduate, so I’m planning not to take classes in Spring 2016 so that I can do an internship (or several), but I’m not sure about how to find the right opportunities. I’ve reached out through some social connections to folks who might be interesting, but I should do more.
Do you have any tips for finding internships? Or even better, do you know of any great people who could use a smart research assistant this coming Spring? I do good research.
Thank you, Idealistic Human
Great idea, and I’m sure my commenters will weigh in with ideas. Personally I’d find underfunded organizations that do good stuff and I’d simply ask them if they need help. The ones that advertise for internships are way too overstaffed and organized.
Dear Aunt Pythia,
So, as a not young in body person (half a century, woohoo!) am a little surprised to find myself:
- with a job after almost 12 months out of work,
- excited like I was starting fresh, and
- worried about the future – aka ai/robots getting the work.
The job I am about to start shouldn’t last more than 5 years. The goal is to set up a reporting system for a variety of KPIs drawing on data from a variety of external organizations.
On the one hand, if I don’t manage to automate most of this, I would see it as a failure. On the other hand, what work will be left for others when I succeed?
I will be fine. After 5 more years of earning, I should be mortgage free and healthy savings. Should I feel a bit bad that I am helping software eat the world?
Frumpy Old Graduate Excitedly Yearning
A wise man (Suresh Naidu) once said to me, “protect the people, not the jobs.” I think he’s right. We are going to have to deal with the robot/ automation revolution sooner or later, and so instead of pushing to avoid automation, a futile gesture to save unnecessary and outdated jobs, we should be thinking about pushing for free college and training for the jobs of the future with all the money we’re saving as a result of this nifty automation revolution.
So, in short, no, don’t feel guilty. But be sure to do your part in figuring out what the future should look like for young people once you retire. Be an advocate for a fair and equitable future!
Readers? Aunt Pythia loves you so much. She wants to hear from you – she needs to hear from you – and then tell you what for in a most indulgent way. Will you help her do that?
Please, pleeeeease ask her a question. She will take it seriously and answer it if she can.
Click here for a form for later or just do it now:
For the past two days I spent my time bored out of my mind at jury duty. And it’s not even that unpleasant or uncomfortable, and it even has pretty good wi-fi, but for some reason, seated as we are in a big room with 90 other people or so makes you kind of nuts. It’s like you’re on a two day plane ride to nowhere.
For reading I had with me A Confederacy of Dunces, which I’m reading for the second time, and which is great background for the story I’m about to tell you.
On the first morning of jury duty, you get to see your cast of characters, and it’s kind of amusing. In my case, we had an extremely overworked clerk named Bill, who was doing the job of three people, telling us how to fill out our forms in precise and extremely detailed patient language, repeating everything 5 times for clarity and emphasis. And I would have started to wonder at Bill’s constant repetition, except that in spite of it, there were a few people who would manage to get confused and go up to him – invariably in the middle of a task – and ask him questions.
One woman in particular seemed to do this a lot, and she was loud as well, and almost seemed hostile. It seemed like she was objecting to the form itself, and wanted to find a way to trip up Bill or something, as a way to get back at having to be at jury duty. To the credit of Bill, he was always extremely polite to her. That guy is a saint. But it didn’t prevent her from looking around at the crowd of people, as if she wanted confirmation that her plight was unreasonable.
So yesterday rolled around, the second day of interminable waiting, and it was much worse than the first day. Because, after all, we all knew how boring it would be, and we were all hoping we wouldn’t be called to do our actual civic duty. Being prepared to do it was surely enough. The guy next to me kept mumbling, “I gotta get outta here” under his breath, while shaking his leg furiously.
At around 11am, something happened that kind of broke through the tense fog of boredom. Namely, about half of all the cell phones started to beep loudly. It was an Amber Alert (since resolved). We all pressed “OK” and the beeping din subsided.
Except not for long. I guess people who are on different networks get their Amber Alerts at different times. So for the next 10 minutes or so, random cell phone beeps would happen and be resolved. For all but one phone, everyone’s noise eventually went silent for good.
That last phone, however, was left unattended, which meant that every 3 minutes or so, it beeped loudly for 15 seconds. The fourth or fifth time this happened, the loud lady from the previous day started loudly complaining, “THAT NOISE IS ANNOOOOOYING ME! CAN SOMEONE TURN DOWN THAT NOISE?! IT’S SO ANNOOOOYING!”
Some combination of how pent up everyone’s frustration was, and this loud woman, and probably also the book I was reading, made me start laughing uncontrollably at this point, which was slightly contagious but didn’t stop the loud woman from complaining.
A bunch of people started to explain to her about Amber Alerts, but she just kept telling everyone how annoyed she was (loudly). Finally, one of the clerks at the front, who had (very reluctantly!) decided to show up to work today, told her there was nothing she could do and could the woman settle down.
Well, that made her quiet for about 15 minutes, but it didn’t stop the Amber Alerts from sounding every 3 minutes. And every time they started again it was difficult not to laugh. After the fifth time, some guy who had been in the bathroom for the first kerfuffle made the mistake of saying to the group, “I think someone needs to look at their phone and deal with it,” which was the cue to the loud woman to start wailing again about the noise, and it made a bunch of us start laughing again (I admit I was the worst). This time the loud woman added some sarcastic comments about how SOME people seemed to think her suffering was FUNNY, which made me simultaneously laugh harder and consider suicide. The lady at the front asked us all to settle down again and we did.
At this point it had been going on for almost an hour, everyone was hungry, and it was nearing lunch time. Finally, just as we were being dismissed for lunch, someone sitting next to the loud lady proclaimed, and I literally have no idea why it took them so long, “Hey lady, that’s your phone!,” which she denied, but then the woman up front said, “Lady, that better not be your phone! Take out your phone and check it right now!”
Readers, it was her phone.
I’ve been getting lots of people writing to me recently with data science jobs that I think look like nice jobs, with nice people, but I am personally not going to apply for. Instead of forgetting about them, I’ve decided to advertise them myself.
The first position is from Harmony Institute, a research center trying to understanding the social impact of media. They are looking for a Technical Data Analyst. Here’s their blurb:
HI is looking for a media-savvy researcher with an understanding of social science/humanities research and a passion for empowering causes for social good. The ideal candidate is interested in studying complex phenomena — such as information diffusion, media framing, and emotional responses to stories — from multiple, often indirect perspectives. You know how to hack code and wrangle data, but you’re just as concerned with big questions as big data.
In this role, you’ll be working closely with a team of social scientists to study various aspects of the social impact of media. Most projects have a critical data component, and you’ll be the go-to person for a broad range of data tasks: identifying indicators and datasets of interest based on research questions; collecting data from available APIs, databases, and even web scraping; verifying, documenting, and sharing the processed data; and either leading or facilitating the data analysis and visualization of results. You’ll also collaborate with and get support from our product focused data scientists and software engineers. Fortunately, HI has built up a large code base (in Python) to help you with your work, so you’ll rarely be starting from scratch. You can look forward to improving your programming and data skills through diverse, interesting research projects; being exposed to many different perspectives among media makers, funders, and researchers; and advance the field’s understanding of media’s impact on society.
The full description is here, and they’re also looking for a data engineer.
The second position is from Paperless Post, which you’re doubtless familiar with if you’ve ever been invited to a kid’s birthday party. They are looking for a data scientist, and here’s their blurb:
The data team at Paperless Post helps people throughout the company use data to make better business decisions. We’re a team of five people (three women, two men), and each person supports various business and product teams. We spend around three fifths of our time working on projects for teams, and two fifths working on data team projects.
We do all types of data work: data engineering, ETL, analysis, building production data systems, building visualizations, building dashboards, creating KPIs, and more. As the main data contact for a team, you get a lot of autonomy and flexibility when helping them best use data.
We’re looking to grow our team, and are looking for both junior and senior data people. Paperless Post is a great company to work for as a data person. We have a product that people enjoy and pay for, and we don’t sell our user data. We care about professional development, and are supportive of interests outside work. Within the company, the data team is viewed as a valued partner to the teams we work with.\
For questions, contact firstname.lastname@example.org.
If you have a data science position that is not in finance, or ad tech, and is with super nice people, I plan to do it at least one more time, so email me at the gmail address listed on my About page.
Whenever I hear an argument about the possibility of raising the minimum wage to $15 per hour, it sounds like this. Person A, who is for it, makes the case that it’s too difficult to live on minimum wage earnings, and it doesn’t make sense for someone working full time to struggle so much to feed their kids. Person B, who’s against it, says that 15 is too high, that too many employers will be unwilling to pay for unskilled workers at that rate, and they will replace such people with machines instead of doing so. Essentially, they argue the bad will outweigh the good.
Full disclosure: I am often Person A. I once figured out that if you take someone’s hourly wage in dollars, and you multiply by 2, then you get their yearly wages in thousands of dollars. That means an income of $100K per year is $50 per hour. That means an income of the current New York minimum wage, $8.75 per hour, is a measly $17.5K per year, which would be absolutely crazy to try to live on, according to my reckoning. In other words, I think about what I could theoretically live on, if I had a minimum wage job, and I have extreme sympathy for people who try to.
Let’s get back to Person B’s argument. It’s weird because it sounds like Person B is arguing for the sake of the poor, but they’re ignoring the vital question of what is a living wage. Let me give you an analogy.
You have a sick population, and they all need 3 pills per day to stay well. The pills are expensive, though, and so the people in charge of pill distribution give most people 2 pills per day. They argue that, if they gave out 3 pills per day to everyone, some people would have no pills. For the sake of those theoretical people, then, they give out only 2, and everyone remains sick.
In other words, for the sake of holding on to crappy jobs that pay below living wages, and where the employees need food stamps to survive, we don’t raise the bar so they can actually sustain someone in a basic way. It’s almost like we’re desperate to hold on to them because otherwise our unemployment rate would be higher.
I say, figure out what a living wage is, and raise the minimum wage to that level. I actually don’t know what the magic number should be, exactly. Is 15 big enough? Maybe it is, in some places, but maybe in others it’s actually smaller. It doesn’t have to be the same throughout the country. But for as long as we live in a country where the model is that a job is supposed to support you, we should make sure it actually does.