Hacker News Reader

Comment by mcguire | Hacker News Reader

"There’s an argument that “if there aren’t enough ventilators and patients are being turned away, I’d rather have a slim chance with a hacked solution than no chance with no solution. The reality is closer to “if there aren’t enough ventilators, a hacked solution will most likely do more damage than good, take up too much time from already strained health care resources, and could lead to death.”"

This is a difficult point for many people to understand. Sometimes the problem really is too complicated for a simple solution.

View on HN

I really wish YC would do their part and start nuking all the DIY ventilators articles. I don't want to wait for homebuilt vents and intubation to start killing people before we put an end to this nonsense.

It's a great illustration of why disaster coordinators want real engineers, not software engineers.

Programmers, just get out your checkbooks and stop trying to help.

What’s a “real engineer”? Railroad Engineer? Civil Engineer? Aerospace Engineer?

Don't be cute. Professional Engineer / Chartered Engineer is a "real" engineer. A random coder calling themselves an "engineer" doesn't make them qualified to design and implement life-critical safety systems.

Some countries have that norm, but in the United States, PE licensing is fairly unusual, due to rules allowing corporations to have one PE sign off on the work of a lot of engineers. As a result, even in the case of critical medical equipment, most of the real engineers who scoped, designed, and tested the product are not PEs.

Often the PE is more of a lawyer than an engineer. They typically do have an engineering degree, but are years out from any hands-on engineering work. They do final legal sign-off, and are knowledgeable about statutes and insurance policies, but rely on people with more recent technical experience to do any technical work.

Sure, I could imagine that is the case in fields like nuclear engineering. All the PEs I know are smart and have technical chops. Though, most of my PE friends are fairly young (early 30s) and are maybe on their way to that destination.

"Some countries have that norm, but in the United States, PE licensing is fairly unusual, due to rules allowing corporations to have one PE sign off on the work of a lot of engineers."

Which is significantly stronger than the state of software engineering.

There is such a thing as Software Engineers, but most developers are not that.

In this case, I think it means someone with the perspective to think things through at a systems level and relevant domain expertise.

So, a medical device engineer.

Probably someone with a PE license.

Someone that has at least passed the fundamentals of engineering exam, preferably with a PE. Then we could guarantee a minimum understanding of the relevant fields.

Engineer who pAssed the FE exam here. 9 years design experience. I can do math.. what's needed?

I agree programmers are not on this level.

Why 'level'? Its a different body of knowledge.

Tons of programmers are on that level - in fact, probably more developers at 'Engineering caliber' than any other Eng. discipline.

That said - the vast majority of developers are nowhere near that.

It's a really long tail.

I assume Boeing engineers are enough

I've previously expressed my concerns that the focus on ventilators will lead to a secondary pandemic of antibiotic resistant secondary infections. But I don't think "nuking" these articles helps. I think it is much better for them to make the front page and get appropriately criticized. That pattern seems to be actively fostering better articles making it to the front page.

I also wonder what's so bad about all the simple existing designs, which are all out there, ready to be produced (and validated to work?)? I would understand if people would focus on taking these designs and adapting them for different materials/.... What I don't get is, that everyone is trying to reinvent the wheel (probably classic NIH-syndrome).

When all you have is a hammer, every problem looks like a nail. With an Arduino and the ability to program, we get ventilator designs. Elsewhere there are masks and gowns being hand sewn, hand fabricated, or 3d printed.

Maybe the existing designs are fine. They might also be proprietary though and controlled by slow moving corporations (as opposed to fast moving ones and startups). There may yet be good opportunities for innovation but the main thing is to spike the rate of production. Yes while being safe and legal, those are table stakes.

Anyway reinventing the wheel is underrated. Taking actual wheels as an example, there’s a long and impressive history there and we’re not close to being done yet. And so with most other things.

It’s a nightmare with the crowd that was so in love with Quinines the last few days. The argument is, basically, “I believe in science, but not during a crisis”. I guess that saying about foxholes and atheists isn’t entirely wrong.

At its core, these are all variations of Pascal’s wager. I believe what would be most useful is pointing out that using the scientific method does not automatically lead to multi-year 3-phase trial requirements. It’s perfectly possible to use the statistical methods we usually do, but adjust parameters such as cutoffs for significance or weighting of side-effects to better match our preferences and increased risk tolerance.

> The argument is, basically, “I believe in science, but not during a crisis”. I guess that saying about foxholes and atheists isn’t entirely wrong.

An important factor here is also the unfettered, obnoxious arrogance that seems to be ubiquitous in tech culture. It's not so much "I don't believe in science" so much as "I'm above science, and therefore already understand all of it."

It's the idea that if you have a sufficient mastery of software engineering then, by extension, you have a mastery of the technical aspects of literally every other field, whether it's medicine, science, psychology, law, or even art. (Or, at least can easily master it after reading a few Wikipedia articles on the topic.) And if only the experts in that unrelated field would listen to the ingenious ideas that you thought of just now, then the crisis would be averted.

But, I'm being uncharitable here. It also comes from a place of genuinely wanting to help during an unprecedented global crisis when we're all frightened, isolated, and feeling helpless.

On the one hand, you are imagining that arrogant thoughts you came up with are happening in the heads of these unnamed tech leaders. (Probably not the case. My guess is they do things from a humble bottom-up “let’s learn the fundamentals and go from there” kind of approach.)

On the other hand, sometimes they do upend industries by remaking things, usually by finding and banishing some devastatingly huge false assumption the legacy players have built everything around, and exploiting the opportunity that this opens up.

So yes, uncharitable.

It could lead to death, but if your alternative is dying from natural causes, I’d rather have the choice. I don’t see how it could do “more harm than good” on an individual level. On a mass distribution level, sure.

> Sometimes the problem really is too complicated for a simple solution.

I've been yelling this at the top of my lungs across a range of platforms. I understand the hubris created by the Dunning-Kruger effect. We all fall prey to that at some point or another in life.

That said, just because people don't know what they don't know it doesn't magically endow them with supernatural powers to be able to create solutions to problems that have decades of development, a complex history of evolution and are the result of intense development, testing and manufacturing by dedicated multidisciplinary teams over years of dedication.

You cant Superman ventilators.

Please stop.

Find other ways to contribute. If you must work on ventilators, call established companies making them and ask how you might be able to help. Ask to work for them for free if you are able to.

> If you must work on ventilators, call established companies making them and ask how you might be able to help. Ask to work for them for free if you are able to.

Anyone doing this should consider the possibility that the answer may be that there's nothing they can do on a useful timeline except write a check.

If you want to help, you gotta swallow your ego.

> For every complex problem there is an answer that is clear, simple, and wrong.

-H. L. Mencken

I agree with what you are saying, but to be fair it's the exact sentiment Elon Musk had to deal with in the early days of SpaceX. Thankfully he didn't listen.

That's not what Dunning-Kruger's thesis is.

I guess that makes your comment Dunning-Kruger about Dunning-Kruger.

From Wikipedia.

"In the field of psychology, the Dunning–Kruger effect is a cognitive bias in which people assess their cognitive ability as greater than it is. It is related to the cognitive bias of illusory superiority and comes from the inability of people to recognize their lack of ability."

Engineers who don't know they don't know what it takes to design, manufacture and test a life-critical medical piece of equipment at scale. Yup, very much Dunning-Kruger. They think they know far more than they actually do. There's was a link HN just a few days ago from a bunch of physicists who set out to design a ventilators using, among other things, Home Depot parts. Let me put it this way. We have someone in my family with a PhD in Physics. Super smart guy. He has never built anything and can't build anything, much less life saving devices by the thousands.

DK doesn't just apply to "dumb" people. I have worked with dozens of engineers who don't understand just how much they don't know.

Maybe not the strict intent of DK, but very much the same kind of effect.

Let's not split hairs. I think you know what I meant.

I've worked at companies that do light manufacturing. Light manufacturing because we contract out components to houses that specialize in that. I feel like a lot of software people don't get that hardware at scale requires 10X the amount of resources each step from design to production to shipping and support.

hardware also fails in the field in unpredictable ways. Especially to anyone without long practical experience.

Exactly. There's a general and very real disconnect with what it takes to make things. I watch the White House and Gov. Cuomo's press conferences every day. The repeated questions about test kits, beds, ventilators, etc. drive me up a wall.

I calm down when I realize none of the people asking the questions understand manufacturing one bit. And then I get rattled again because they are communicating falsehoods to the general population. We don't have enough masks, test kits or ventilators because you can't Superman hardware production.

It isn't Trump's fault or anyone else's. It takes time --lots of time-- to setup a manufacturing process and start to produce product at scale with consistent quality and performance. Weeks for the most trivial of products, years for more complex hardware. Yet people don't have an understanding of this.

Ventilators? Holy cow. Talk about not being able to Superman something into existence.

If people understood they would realize that we made a massive mistake over the last 50 years by allowing our industrial base to evaporate and migrate to China to the extent it has. Consumers are as much at fault as the decades of incompetent politicians who allowed this to happen. While we bothered with nonsense (wars, etc.) China quietly went from an agrarian society to the second economy in the world. And, even worse, the world's factory. We, quite literally, cannot make anything in the US or Europe without Chinese components and, in some cases, all we are able to do is design in the US or Europe and have no choice but to manufacture in China.

I suspect things might change over the coming years. The money we wasted in wars would have been far better spent keeping industry alive locally. Yes, subsidies, no taxes for manufacturers, relaxed regulatory burden, etc. Either we do those things or the next pandemic will, once again, find us not being able to even manufacture face masks at scale. How laughable is it that people have to sew face masks at home in the US? Sad.

BTW: In Spain they just woke up to the reality that they received 340,000 bad test kits from China.

I'm not a manufacturing expert, but if came down to it and I was sick and only had the choice between "drowning alone as my lungs filled up with mucus" and "trying a ventilator that probably doesn't work and was built by an idiot using Home Depot parts and a Raspberry Pi" I'd still go with the idiot.

Sure, ideally we'd want the certified equipment, built by the professional knighted safety engineers, stamped by the right bureaucrat, with 1000 pages of documentation proving how thoroughly expert-approved they are, but when you run out of these, what do you do?

The problem with that idea is that the application of a ventilator is far more nuanced than inflating a car tire. A home-brew ventilator could kill you faster than the mucus (and likely in a horribly painful way, for example, causing your lungs to bleed internally).

Reply on news.ycombinator.com