Hacker News Reader

Comment by fallingknife | Hacker News Reader

Important but not mentioned in the article is the amount that the government paid for that incompetence driven disaster. $840 million!

View on HN

Unfortunately the cost of healthcare.gov has become politicized, but it was extremely, absurdly expensive. Wikipedia says the total cost actually landed somewhere between 1.7b and 2.1b.

https://en.wikipedia.org/wiki/HealthCare.gov

I worked in the "rescue team" for a couple of years and the politics was one of the reasons I left, you never knew when the supreme court would stick their nose in and you'd be out of a job. I used to describe that job as having 100 bosses, 55 of those bosses wanted you to do your job for the least amount of money as possible and the other 45 wanted you to fail.

and also the terror that you might get hauled into congress to testify

I actually used to use that line with I'd talk to some of our vendors support people. They'd start dragging that feet, wanting to get rid of me and I'd say, "you do know what project this is right...." and then I'd say if we can't fix this really quickly it will be on the news and your company will get dragged into it and nobody wants to testify to congress that the market place site was down because of a bug in your equipment. It was pretty amazing how people went above and beyond when you put things in the right perspective.

The crazy thing is that a small team of around 10 really good engineers could probably have put together a robust reliable working Healthcare.gov-like system that could handle the potential 300 million users in 1-2 years. If you paid all 10 of them them $300k each a year, for 2 years, that's 10 * 300 * 2 = $6 million. Worst case, Healthcare.gov should have cost 5 times that, $30 million. I feel like the sort of egregious and copious waste in government should borderline be illegal.

The craziest part of it is that the 840 million isn't fraud. The contractor didn't pocket that. They actually spent that much money on legitimate development work. And if they had just delivered a working app, nobody would have ever asked any questions.

In a similar story, USCIS spent $3.1 billion on trying to build a web app to submit forms: https://www.forbes.com/sites/realspin/2017/02/15/billion-dol...

And they failed at that too. Apparently what would have been a weekend project for a high school student was so hard for a federal agency that they burned $3.1 billion trying to do it. (This level of incompetence is truly despicable.)

I wonder if there is something pathologically wrong with government and how government (at least in the U.S.) does spending.

And I hear people seriously suggest that the government can't get things done because they are "underfunded"!

Having worked on the customer-side on large IT projects, some of these large contractors establish policies and development methods that are guaranteed to produce problems. They also hire incompetent and cheap labor, which causes more work.

Why shouldn't it be politicized? It's absolutely fair game to attack the other side for inefficiency and waste. Wish we had more of that than the usual back and forth conspiracy theories.

No, that’s totally reasonable. I was and still am a supporter of the ACA and I think the price tag for this site was bonkers.

I just think it’s a shame that it’s politicized because we can’t easily discuss it as a case study without pre-existing political opinions creeping in. It would be neat if this particular series of events could be dissected in a vacuum. Government projects go sideways all the time but there’s a lot of transparency into why this one did because of the unique circumstances of its resolution.

My limited understanding of the system is that it is a feature, not a bug.

Both sides of the aisle have patrons that feed at this trough and I doubt we'll ever get consensus better operational insights and less waste and fraud.

Like so many other issues, the only thing holding us back from achieving this is will.

Introduce competition into the electoral process by doing away with First Past The Post voting in favor of Ranked Choice voting/Single Transferable vote/STAR voting/Alternative vote.

People should be free to vote for someone that best represents them, without worrying about a spoiler effect or strategic voting practices.

That would be a great start for democracy. Real change requires Congress to do its job, and it's pretty much paralyzed these days.

There's plenty of input to consider on how to improve things, we just need to find the collective will to act on it.

https://www.brookings.edu/articles/reforming-federal-procure...

That's nothing.

It's estimated that US healthcare per-capita spending is 20% more than countries who use a single-payer program (that's nearly every other developed country in the world).

Only the UK and Taiwan have a nation-wide single-payer healthcare system, although Canada comes sort of close with each state having its own separate single-payer system. Most countries use some sort of insurance model with varying levels of public/private provision and government and employer contribution. Switzerland is the only country with 100% private healthcare provision, but it offsets that with generous subsidies and having insurers offer a mandatory not-for-profit plan meeting legislated minimums.

https://www.vox.com/2020/1/13/21055327/everybody-covered

Unfortunately the federal government is allergic to doing anything in-house.

The US Digital Service does fantastic work. The first head of the project was the person involved in the rescue of Healthcare.gov.

18F too. Although their name is unfortunate. Maybe they'll open a Cali office someday.

Right? Don't they know software is free and insurance is a simple industry?

It's not free, but the initial contractor CGI definitely hosed taxpayers.

The rescue proved something everyone knows: some software engineers suck; others rule.

And that is an excuse for spending damn near a billion for this?

> Testing was largely a manual process. Releases and deployment required lengthy overnight downtimes and often failed, required equally-lengthy rollbacks. The core data layer lacked a schema. Provisioning hosting resources was slow and not automated. Critically, there was a pervasive lack of monitoring of the site. No APM, no bog-standard metrics about CPU, memory, disk, network, etc., that could be viewed by anyone on the team.

The “beauty” of contracting for the gov’t - if you know how to play the game - is that their internal management and incentive structure is usually so screwed up (and will never get better) that manipulating the customer is easy and lucrative.

If you know how to do it without being indicted anyway.

Constant back and forth change orders, pointless delays to serve constantly changing stakeholders, constant demos with no real product, no one in the customer with any power who has experience to call BS on timelines or delivery, etc.

And if done correctly, everyone in the ‘customer’ is simultaneously happy and angry and so it’s easy to deflect blame and point fingers back into their own org.

Or so I’ve heard from folks who are involved. I try to avoid it like a plague, the BS I deal with hurts my soul enough already.

Having worked for a Beltway Bandit "consulting" company in the past, you are exactly right. It's not about getting work done, it's about spending hours so you can receive more.

That was one of the takeaways from this whole episode. In its wake, the Technology Transformation Services and 18F resulted. Part of what they do is show government agencies a better way of procurement (a "modular" or "agile" approach). Performance-based contracting, and where the source code is, say, in a GitHub repo rather than on the contractor's servers. Contractor not performing? Get rid of them and get someone who will. Need to change the deliverables? No problem because you're not contracting for a deliverable anymore.

The federal government is a big beast, though, so it takes a while for the idea to spread.

What good does it do to show them a better way? They are a monopoly that takes their revenue by force. What is their incentive to improve efficiency?

That depends on how obvious it is. They can’t be clearly doing it on purpose, or they could be indicted too. If they look like they’re doing their best and following the law, but it doesn’t work? That’s perfectly fine.

Of course not, but the people getting paid the bills also understand how to say "it is a complicated and difficult regulatory environment" frequently.

If leadership isn't careful (and it seldom pays to be careful), it becomes an excuse for every real wrong.

Reply on news.ycombinator.com