Comment by menacingly | Hacker News Reader

Comment by menacingly | Hacker News Reader

Civilization at scale is very difficult. It seems to have something in common with software. Naive ideological purity produces folksy, approachable solutions that fail to address just enough edge cases as to not function past a certain point.

The more I encounter scale, the more terrified I am of it. We don't seem to have any great answers. I spent my youth frustrated that the suits were going to complicate my elegant simplicity, that some series of meetings would pile on needless complexity, but the miserable truth is they aren't exactly wrong.

Sentences that begin with "Why don't we just-" are usually fatally shortsighted. Startups that appear to have finally cut through the burdensom red tape of the crusty establishment are doomed to become their parents: at scale, they start to do what they were fighting in the first place. Google looks like Google for a reason.

We love a fresh start, because we see in it the promise of a new take, someone applying the simplicity to a problem we intuitively feel is there, but sooner or later they must face the meat grinder of scale.

I hope it isn't true that processes must trend toward the absurd, that patently outlandish medical billing is the fate that awaits us all.

As a South American I respectfully disagree on this.

As you mentioned scale I will cite Brazil as a baseline since we’re talking about US here.

Even in one of the most {corrupt, poorest} places in the world the entire infrastructure of what we know as public health care was built in a timespan of 28 years, serving a population equivalent with USA with a more pulverised population (e.g. see the demographic composition of Amazonas and Pará states).

To be clear, I would prefer to be incorrect, but this is a hard-earned cynicism.

What do you mean by that? Can you formulate?

I don't think scale is really the problem here. If so it is the scale of health care costs primarily. Since those tend to explode easily, workers are expected to document every little step they take. And if a nurse spends 10 minutes to console somebody, she is in need to justify that time towards her superiors, who in turn need to justify the cost of health care to insurers, who in turn need to justify their expenses to patients or government. It is a self-reinforcing circle of hell and fully behaves that way because wrong decisions are taken at multiple steps. It isn't even that complicated, just short sighted incentives that fail to get a system under control.

The case for Google is another one. The incentives of the company slowly changed as it moved onto the path many corporations took before. It isn't a necessity, but the drive to growth put them there where they are today. They wanted money, the problems of software logistics were already solved when they were a smaller company. In fact only when they were successful more and more people wanted an in on the profit.

Of course problems are often more complicated in detail, but I don't think too many people looking at the larger picture are really the problem.

> Naive ideological purity

That is just a rant, I don't think there is any evidence of that happening anywhere.

Certainly the entire thing is a rant, but my point is not to criticize the naïveté, it’s to sympathize with it.

Unfortunately, you win the election, you run the social network, your revolution is successful, and you come face to face with the actual challenges your predecessors faced. This is the source of our continued frustration with elections doing nothing.

Without a theoretical leap in solving problems at scale, there is little to be done, and the variances between solutions are just fiddling with coefficients.

It’s easy to think that some shadowy league of dentists is preventing a cavity cure from reaching the market, or that the forces of for-profit healthcare are arrayed together in an alleyway. We love to think that a very large series of people have just not had the balls to stand up to our favorite caricature of an enemy, but the truth is usually nuanced, complex, and slow.

The asterisk at the end of simple solutions is usually “provided everyone agrees with me or bends to my will”

We’d likely both chew our own leg off to escape the meeting to eliminate this billing line item, both blaming it on some complexity upstream we can’t agree on.

> We don't seem to have any great answers.

but we do. economics, especially behavior econ is the study of understanding why large scale systems work differently than the de facto rules.

yes, every separate "systemic incentive failure" is a terribly complex pathological state in the abstract phase diagram of the dynamics, but we luckily have a higher level control, we can introspect ourselves, our systems.

yes, the resistance from the beneficiaries is always strong, but not impossible to overcome eventually.

Let’s not pretend that American problems apply to the rest of the world.

I work in Australian healthcare systems and can tell you that this situation does apply to the rest of the world, although the charging usually occurs to the public purse rather than the other way around.

The problem is that these activity/diagnostic/lifestyle codes exist for good reasons but existing incentives can push diagnostic coders toward exploitation for maximum funding rather than as a tool for capturing treatment burdens.

The US spends dramatically more than every other OECD country.

It's not even slightly comparable. They are the clear outlier on Earth.

https://data.oecd.org/healthres/health-spending.htm

Which creates the bizarre artifact that the US government spends around the same or more on healthcare than the governments of countries with socialized healthcare systems.

I think you read the chart wrong. At least you write "around the same or more" when it is much more. The amount the source lists for "government/compulsory" spending is $9054. That is a quarter more than the _total_ of the next biggest spender (Switzerland at $7138).

The rest of the western world, may be.

Reply on news.ycombinator.com