This is just an outstanding article that I think all should read.  I can see why the author is remaining anonymous as he would probably be fired as banks and insurers defend their models and software.  Models sometimes need to be fixed and here’s another great article from a former Wall Streetimage Quant that tells it like it is as Cathy O’Neill tells you she was given the job to fix models and what was desired was to add more complexity and she states we don’t need more of that, but rather models that don’t lie.  Today she also has a great post on the creepy job hiring models.  I covered some of the effects of that with flawed data in my Attack of the Killer Algorithm series last year.  Here’s the bio on the author of this outstanding article discussing why tech projects fail and it just goes to show they know exactly what they are doing and how the consumer gets taken over and over. 

“The author, a senior IT executive at one of the nation's largest banks, shares his experiences under the pseudonym Coverlet Meshing. He has spent the last two decades in the financial services sector, picking a fight with anyone who doesn't understand that banks are actually software companies and need to invest in engineering as a core competency. His cheery outlook and diplomatic nature are rarely reflected in his writing.”

Number one on this list, what you read here too is that ROI numbers are mostly fiction and I say that a lot when folks starting claiming “trillions” in all these silly press releases we see out there.  You see them too and know who they are that keep the media full of them.  As I keep saying it’s age old scheme of “throw some numbers at them”…people will suck it up, no matter what it is so it’s easy to mix n the garbage with the good stuff.  Think about this when the next big press release rolls out with the illustrious word “trillions” appears relative to savings as the software will keep prices rolling.

“The most complex variable in the ROI equation -- one that's usually ignored -- is the cost of the business re-architecture required to consume a proposed technology. If you take away nothing else from this article, know that technology demands business transformation, and that's usually the largest hidden cost”.

Number two is great as well and again echoes what I say as far as ROI driving tech investments…it’s the steroid marketing for a lot of it, as he says dishonest marketing…and the the author says he worked at many start ups too so I think that lends some credibility here, so he has more than one point of view and is what I call a “hybrid” who can see a big picture.

The third is what I write about a lot with no accountability with technology..boy do we know that one.  There are areas to where accountability is difficult too when working with several data platforms to where you are using code you didn’t imagewrite and have to depend on the other one’s code but again the two, three or what ever number of contributors should have some kind of group responsibility and we see none of that.  Just look at all the errors on credit reports and what is provided to you and you don’t get the whole package as it might expose even more errors and they all want to retain their “good’ reputations.

Further more he addresses how companies become slaves of sort to the big IT folks out there like IBM and others, if they can afford it.  This is a good quote as well below, keep your workforce dumb, no different than what the goal is for consumers as you can model and manipulate money easier when nobody understands or knows what you are doing and the end result is focus on cost reduction…exactly what the government and all the digital illiterates are doing to all of us right now, becoming obsolete as they can’t or won’t see the reality of how technology is shaping everything out there. 

“Finally, bringing in outsiders keeps your workforce dumb. It locks you into a vendor interested in getting you hooked on its proprietary black box, the corporate equivalent of a gateway drug.”

Instead we get the crappy OMG emotional “drag you down” stories or something about some Hollywood Star.  I read the Huffington Post articles via Twitter and every day there’s an article with something about Kim Kardashian…every day without fail…that’s what you get fed and other media outlets do that too and of course some of that is our culture but darn run it into the ground for clicks and ad exposure.  It makes money for the media and some are really struggling so this is what you get to keep traffic moving on their sites and it kind of takes away from the focus on the good stuff out there. 

Again we can perhaps look at this as some very good information or go out and sucker up to the OMG stuff in the media but this author has it right and has the hybrid background to write about it.  Will anyone listen to reality or will we suck up to the garbage forever?  I call that phenomena Algo Duping and there’s a series of videos at this link that will catch you up and maybe you can figure out where you have been duped with models and math…we all have but when you admit it and see the reality it takes on a new image for sure and you will get mad.  So far not many are getting mad so the duping continues  on and on.  We are under the Attack of the Killer Algorithms and nice for this banker to confirm this as he lives it day in and day out, and is probably making a lot of money but has a conscious and a burning desire to try and educate, for those that want answers and can come out of the sometimes brutal world of software.

One of the videos linked above says it all as we have Mike Oslinski who developed the software all the banks used for the sub prime loans…”you can do anything with software” he says and he’s right (want to talk bitcoins maybe) …you can dupe and make money too as he talks about how the banks took his basic software and modeled and demented what he wrote for profit.  He wrote for efficiencies and it’s important to note how others take such software and rewrite code and models for profit.  It’s the same thing as buying Word and using it for ulterior motives..it’s not the developers of the software, it’s the folks with how they modify and use it to make money with models that lie.  See what Mike and others have to say to include some of the smartest and best Quants around..they tell you how it is.  I love Paul Wilmott describing how “we will just pick 6 for the number”…as he says no science there at all for that model. 

Quants: The Alchemists of Wall Street Video Documentary - Why It Needs to Matter What Companies Do and Not Focus Only On the Price of Stock With So Called Value - Attack of the Killer Algorithms Chapter 44

Looks what’s happening in science and medicine journals…rings clear again what Mike said “you can do anything with software”…

Pseudo-Academia World of “Spam Type” Medical Journals That Will Publish for A Price Polluting Research Society–To Make a Buck, Back To Data Selling-Duped, Duped and Duped Again

So now that we have had this discussion, any wonder why insurance premiums are not decreasing…you are paying for a lot of software in your insurance payments, a lot of it.  Sure there’s some that is really needed too and I don’t dispute that but we have insurers going off the deep end with never having enough and report after report will show up in the news to substantiate their models and algorithms.  Take this example with Blue Cross buying up Master Card and Visa consumer buying data…”looking for consumers buying clothes that are one size larger”..well that’s what Blue Cross said but don’t forget they get everything else in there as far as what kind of gas you put in your car, where you eat out at restaurants, what you spend at drug stores, the grocery store and so on.  Was that the best explanation that Blue Cross could come up with for paying the banks for this data?  I’m sure it cost a lot and then some of the data when combined with other data bases will all of a sudden give birth to some new “analytics” that perhaps their intelligence division will market for sale.

Insurance Companies Are Buying Up Consumer Spending Data-Time is Here to License and Tax the Data Sellers-As Insurers Sell Tons of Data, Gets Flawed Data When Data Buyers Uses Out of Context Too

If you want to read more then use the link below on Big Data as it explains the culture here with data selling and how it has become an epidemic with nobody ever being satisfied and the silly degrees companies and banks are going to hold on to that one tiny hope they will find a “miracle model” and code to run it.  Again I fully understand the use of linear data for efficiencies and information but so many are out there winging it on the hope that the magical model will be created, again we come back to Algo DupingHalf of all the analytics purchased will be a waste due to the fact that “the perfect model” as the marketer sees is will be pushed to the limits with press releases and other media coverage to get you all excited when nobody has validated or been able to confirm the value of what they are selling..the way it works and has been working..read up on Algo Duping for the real perspective here.  You haven no idea to what you are buying a lot of the time and just accept whatever sales pitch is given. 

Big Data Revolution-Dangers of Using It To Drive Ourselves Off a Cliff -Debunking Myths of Algorithms Replacing Humans In Decision Making Processes in the Healthcare Business (Video)

Shoot when I read this article about the IRS, first thing that came to mind here is are they trying to predict who will and will not file taxes and who might cheat from social media?  Where did they buy their data from?  Experience tells me that there’s an IT Infrastructure  in there somewhere with a model that some Quant or Programmer developed.  It might be rough as most of what the government creates is rough in fact as they get duped too on some of this. Economists for that matter today too are pretty much guessing and using tools written by a Quant or Programmer, so again who’s giving them the tools and do get software tools with accurate models..something to think about..If I were an economist I might ask that question.

IRS Using Facebook and Twitter as Part of Their Intelligence–Opportunities to Sell Data Continue As Private Industry and Government Cash In On Intangibles Leaving Tangibles in the Dust…

Overall this was a very good insight written here by the anonymous banker and glad we have some folks out there that see the reality and can share a little bit from their bird’s nest up there.  Wished we had more of these folks instead of those who create complex models and some that just flat out lie.  Here’s the link to the entire article, see what you think. 

Again when your health insurance premiums come due for payment, just think of all the software you pay for as well as your actual care and how much is really needed…with all of this our manufacturing tangible business certainly is given the back seat sadly, which is what we need for jobs and boosting the economy. 

Three years ago I wrote a couple of satire posts of sorts and I’ll be darned if they don’t apply today…A Department of Algorithms and a new 12 step program for data addiction..I revisited the two post recently.  BD

Does the US Government Need a Department of Modeling and Algorithms–Is Data Addiction and Abuse the Next Up and Coming 12 Step Program, Some Classic Posts & Topics Revisited

Finally my suggestion to get into the data selling epidemic that is creating a pseudo algorithmic world that we can’t exist in..license and excise tax the data sellers to get some control and regulation here…otherwise nobody will ever be accountable and the Algo Fairies keep flying:)  You need IT Infrastructure spelled out with digital laws that show certification on some of these models today, so we know we have a chance at not being soaked.  BD

Privacy Wanted–So Let’s Require Those Who Sell Web Data to Register and Tax the Transactions and Publicly Disclose Who They Sell To With a Federal Registry

http://www.informationweek.com/global-cio/interviews/why-tech-projects-fail-5-unspoken-reason/240152282?pgno=2

0 comments :

Post a Comment

 
Top
Google Analytics Alternative