> Michigan Integrated Data Automated System, or MiDAS
They should have listened to King Midas cautionary tale.
This kind of system will work better if it worked both ways. A system that notifies you when you are eligible for unemployment and when you are not eligible anymore. Instead it seems it only tries to identify fraud from incomplete data.
I know that automatic taxing and other government automation is not implemented because there are opposed lobbying efforts. But, the current state is a waste of money and negatively affect USA citizens mental health.
If the USA wants to continue being competitive in the world economy it needs to step down from its pedestal and work hard to improve its infrastructure, digital and physical.
I’ve known a number of folks that have been involved in lawsuits with considerable merit, and not one received anything even close to addressing the damage they suffered; often after years of bruising court battles.
In a quick skim I see the real problem is mentioned several times through the article.
The system was designed to "save costs" by removing humans. Not by feeding the humans a curated and highlighted set of cases to review for confirmation.
A proper system would hold for human review as well as have physical mail confirmation sent to relevant addresses.
How can someone not take the gov. to court, show the judge the numbers, and the judge put a hold on Midas?
And how is such massive fraud only highlighted on some random blog?
This is also quite a good example of illustrating how cable news is so rigorously narrative-driven: their websites right now may have 10 links relating to their 'favorite local story they chose to make a national story' which plays into some kind of politics, instead of having references to more nuanced things like this. But without some 'politician they don't like to blame' it's hard to make into clickbait.
I believe these more 'operational' realities of governance are far more impactful on our civilization than the more ideological things we argue about all day. A highly functional and efficient bureaucracy makes political arguments effete.
I really feel as if software quality is at the heart of a lot of this.
I’m a bit despondent over the casual approach to quality exhibited by a lot of corporations, and the stakes are getting higher.
I dread what will happen when the software controls even more critical infrastructure.
The software is getting so complex that I can’t see how it can be effectively tested. Meanwhile, the entire IT industry is on a wholesale mission to remove experienced engineers, and replace them with outsourced talent, or people right out of bootcamps.
In this case, the people affected were probably not the ones that would (or could) raise a lot of hell. Also, some managers probably got bonuses or raises, based on this.
It’s funny. I remember, many years ago, a story about errors in supermarket pricing, and how pricing errors invariably favored the supermarket. It was determined that was because the supermarkets only considered it a problem if they had to pay.
> I’m a bit despondent over the casual approach to quality exhibited by a lot of corporations, and the stakes are getting higher.
I don't think that there is any evidence that the quality of the software was low. The problem was the design of the system and the willingness of the developers to implement it.
I have a "holistic" system approach. I believe that every system; whether software, hardware, design, written materials, whatever, needs to be treated as a "whole," and that "whole" often includes components that may not be considered relevant to implementors.
Take, for, example, end-user documentation. That's often considered "not mah job, man" by engineers; which is appropriate. It is not their job, –and, this is important– if some knucklehead manager makes it their job, they are gonna get a real turd sandwich for end-user documentation.
What needs to happen, is that the manager needs to hire a real tech writer (not a repurposed administrator -unless they get lucky), and then supply that writer with the tools and authority necessary to do a good job.
Stop me if you've heard this before.
A lot of startups don't have this luxury. Almost every person in the company is wearing half a dozen hats. The CTO may be the main (sometimes, only) engineer. For whatever reason, they don't have headcount money for a tech writer, so the CTO, who is, to be fair, an amazing and energetic engineer, writes the user docs.
Unfortunately, she is also intimately familiar with the system, and has a very advanced baseline. She does not approach the documentation with an eye towards new users, coming in hot and two engines shot to hell. She assumes that they are engineers of her caliber, and that they "of course" know the basics that she glosses over.
So we have this marvelous system, that was produced by marvelous people, and is presented to the user, wrapped in a turd sandwich.
12 comments
[ 2.9 ms ] story [ 11.4 ms ] threadAutomated justice was once thought to be the next great thing. I'm not so sure.
They should have listened to King Midas cautionary tale.
This kind of system will work better if it worked both ways. A system that notifies you when you are eligible for unemployment and when you are not eligible anymore. Instead it seems it only tries to identify fraud from incomplete data.
I know that automatic taxing and other government automation is not implemented because there are opposed lobbying efforts. But, the current state is a waste of money and negatively affect USA citizens mental health.
If the USA wants to continue being competitive in the world economy it needs to step down from its pedestal and work hard to improve its infrastructure, digital and physical.
It's pretty insane.
It’s not helped by corporate PR efforts that gin up fervor about “frivolous” lawsuits.
Like so: https://www.vox.com/policy-and-politics/2016/12/16/13971482/...
I’ve known a number of folks that have been involved in lawsuits with considerable merit, and not one received anything even close to addressing the damage they suffered; often after years of bruising court battles.
The system was designed to "save costs" by removing humans. Not by feeding the humans a curated and highlighted set of cases to review for confirmation.
A proper system would hold for human review as well as have physical mail confirmation sent to relevant addresses.
How can someone not take the gov. to court, show the judge the numbers, and the judge put a hold on Midas?
And how is such massive fraud only highlighted on some random blog?
This is also quite a good example of illustrating how cable news is so rigorously narrative-driven: their websites right now may have 10 links relating to their 'favorite local story they chose to make a national story' which plays into some kind of politics, instead of having references to more nuanced things like this. But without some 'politician they don't like to blame' it's hard to make into clickbait.
I believe these more 'operational' realities of governance are far more impactful on our civilization than the more ideological things we argue about all day. A highly functional and efficient bureaucracy makes political arguments effete.
I’m a bit despondent over the casual approach to quality exhibited by a lot of corporations, and the stakes are getting higher.
I dread what will happen when the software controls even more critical infrastructure.
The software is getting so complex that I can’t see how it can be effectively tested. Meanwhile, the entire IT industry is on a wholesale mission to remove experienced engineers, and replace them with outsourced talent, or people right out of bootcamps.
In this case, the people affected were probably not the ones that would (or could) raise a lot of hell. Also, some managers probably got bonuses or raises, based on this.
It’s funny. I remember, many years ago, a story about errors in supermarket pricing, and how pricing errors invariably favored the supermarket. It was determined that was because the supermarkets only considered it a problem if they had to pay.
I suspect this is similar.
I don't think that there is any evidence that the quality of the software was low. The problem was the design of the system and the willingness of the developers to implement it.
> The problem was the design of the system and the willingness of the developers to implement it.
I would consider the design of the system as a "quality" metric.
Usability, accessibility, training and support are all what I consider to be important components of "quality."
We have a big problem (not just in our industry) of "Not mah job, man."
Take, for, example, end-user documentation. That's often considered "not mah job, man" by engineers; which is appropriate. It is not their job, –and, this is important– if some knucklehead manager makes it their job, they are gonna get a real turd sandwich for end-user documentation.
What needs to happen, is that the manager needs to hire a real tech writer (not a repurposed administrator -unless they get lucky), and then supply that writer with the tools and authority necessary to do a good job.
Stop me if you've heard this before.
A lot of startups don't have this luxury. Almost every person in the company is wearing half a dozen hats. The CTO may be the main (sometimes, only) engineer. For whatever reason, they don't have headcount money for a tech writer, so the CTO, who is, to be fair, an amazing and energetic engineer, writes the user docs.
Unfortunately, she is also intimately familiar with the system, and has a very advanced baseline. She does not approach the documentation with an eye towards new users, coming in hot and two engines shot to hell. She assumes that they are engineers of her caliber, and that they "of course" know the basics that she glosses over.
So we have this marvelous system, that was produced by marvelous people, and is presented to the user, wrapped in a turd sandwich.