[ExI] comments please regarding snoopy-doopy.gov
spike66 at att.net
Tue Oct 29 17:47:09 UTC 2013
From: extropy-chat-bounces at lists.extropy.org
[mailto:extropy-chat-bounces at lists.extropy.org] On Behalf Of Kelly Anderson
Sent: Tuesday, October 29, 2013 9:53 AM
To: ExI chat list
Subject: Re: [ExI] comments please regarding snoopy-doopy.gov
On Tue, Oct 29, 2013 at 9:26 AM, spike <spike66 at att.net> wrote:
All this they did, wrecked the system and influenced government for perhaps
a decade, all in order to gobble up private data on the proletariat.
>.Spike, what is the basis for believing that the health care site failed
because of data collection and not because of poor engineering?
Hi Kelly, sure it was poor engineering and a flawed underlying concept, but
what I meant was, all they needed to do was create a spreadsheet and have
the insurance companies fill it in, then post that to anyone who wants to go
look at it. They didn't need to ask where I live, my legal name, the amount
on my last W2, for two of those three things are likely to change for a lot
of people next year. A buttload of employees will be moved from 40 hrs/wk
to 30hrs/wk so their company need not shoulder that expense, and in many
cases, that change will also result in a change of address. So why do they
ask it? And why the security questions? Once I drop those into the public
domain, those are three pieces of information I can never again use to
secure an account.
I claim that nooooone of what caused HealthCare.gov to fail was necessary.
Anyone who did dump all that info on the government was imprudent or had
exactly nothing to lose (because they had no money and/or were desperately
ill). These are examples of the zombies, who the insurance companies do not
want. They want the track team, who have money and seldom get sick. If you
dump all that info, you tell the government and everyone who wants to know,
which team you belong to, the track team or the zombie squad. I don't want
the whole world knowing that, even if I am lucky enough to be on the track
team, thank evolution.
The claim has been made that the government needed to know this in case you
want to apply for a subsidy. But if you already know you are not eligible,
why isn't there a bypass feature, or a privacy setting so that you can
create a fictitious account? Or just an estimated benefits page? Or just a
comparison sheet? How easy would that be? Hell I could derive that, if
they would just let me shop anonymously.
Kelly, what this was about in my case is I wanted to apply a mathematical
technique I know of which I developed in my career, superposition of
probability distributions. It would allow me to get the available subsidies
in classes or categories. I could then estimate using demographics and my
estimated probability of opt-outs how much this whole scheme is likely to
cost us (by "us" I mean taxpayers) to subsidize the zombies, and how much
would be offset by the tax penalties paid by the track team, corrected by
the percentage that choose to pay it, since the IRS cannot legally collect
it. (Is that a kick in the ass, or what?)
>.My understanding of the Healthcare.org web site problem increased greatly
yesterday, when I came to understand that they are using web services to
gather information from individual states, individual insurance companies,
and from the bureaucracy itself.
Sure but again please, why do they need to know all this? All they would
have needed is to create a pot of money, toss in there the tax penalties
collected from donors (the opt-outs who chose to pay) then divide that pool
among the qualifying destitute and send subsidy checks to the insurance
companies of the opt-in poor, they're done! It doesn't require all the
snoopy-doopy info-harvesting, doesn't even need external funding.
>. All of this data and logic is needed to perform the complex operation of
coming up with a price (A really stinky architecture, btw) and all of this
data and logic is stored and maintained by each party individually.
No sir. The insurance company gives the price. The info was needed to
calculate the subsidy, and that is a critical difference indeed. They could
estimate the subsidy and just say right out that they won't know what is the
subsidy until after it is collected, because they don't know how big is the
pot of money from opt-out donors. They could estimate it, but that's all.
If they calculate it, they are guaranteeing it, which they cannot really do
in any case.
>.This means that if there is one weak link in the chain, you don't get an
answer. This is an Obaminable architecture.
>.A much better approach would have been to say:
1) You host your data and functions here (rackspace, whatever)
2) You answer the questions like so (standard APIs.)
3) test the hell out of it
No sir again to all the above. The right approach would have been collect
data from the insurance companies, knowing they will get their own stuff
exactly right (they are in the biz of being right), make a single digit
estimate of the available subsidy, based on the number of opt-outs who
choose to pay the tax, and state right up front that it is an estimate.
Very little code would be needed to do that, an no testing at all full stop.
>.This isn't rocket science, but the way they implemented it as a
distributed system maintained by 53+ separate entities, it only takes one
entity being a dumb ass to bring parts of the system to its knees. -Kelly
Agreed to that. The whole scheme was based on such deeply flawed reasoning,
it calls into question the fundamental judgment of any government which
would get so deep into it without seeing that which is perfectly obvious to
any outside observer. Am I hallucinating, or is not this design obviously
flawed? To me it is as clear as designing a jet liner with only one wing.
Everyone can see it can't fly like that.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the extropy-chat