Nvidia Releases Low-Cost, Open-Source Ventilator Design
blogs.nvidia.com
blogs.nvidia.com
I have spent basically every waking hour of the last month and a bit working on a ventilator for my employer (part of Ventilator Challenge UK - project recently suspended). Making something that delivers air in the right duty cycle for breathing is easy. There's a few more things you want to do:
Blend O2 and Air ("FiO2"). Ventilators need to deliver Oxygen rich air to patients, but not usually 100%. Blenders are harder than you'd think to make. Note also that having pure O2 in your system means that there's a load of stuff you have to get right to prevent fire.
Alarms. This is the big one. most of the implementation (even taking account the next feature), was about checks and associated alarms, for blockages, failures etc. Making sure that we beeped loudly if something looked wrong. Note that 'not beeping' is a much WORSE failure than 'not ventilating'. If something goes wrong (including complete hardware failure of the ventilator), as long as a healthcare practitioner is aware and responds, then there isn't an issue. If the ventilator is happily ventilating away but something is obstructed then the patient might die (there are backup checks like pulse oximetry on the patient, but they're slow).
Assisted breathing. Driving a solenoid on and off at a particular rate and duty cycle gets you 'mandatory breathing'. That is great for keeping someone alive. If you're not unconscious though, it is both incredibly unpleasant and doesn't help you get weaned off and get better. What you want is a device capable of sensing attempted breaths and using the ventilation to 'help' them. This feature isn't 100% necessary - the early plans in the UK were for a massive shortage and for the simplest things we could get in a hurry. It quickly became clear that actually we needed ventilators to help people recover.
Even just our alarm system was very complicated, with cross-checking alarms and even cross-checking that the audio was indeed playing back. Things like playback tones and frequencies and alarm indicator colors and flash frequencies are highly regulated with relevant standards. And nuisance alarms are almost more important to avoid than having enough actual alarms. No alarm is as useless as the one the clinician turns off because it's the 4th time this hour it's gone off and the patient has been fine every time.
These are not easy devices to build, and selling something that doesn't meet the applicable standards and isn't substantially equivalent to something marketed today is unlikely to be particularly safe. You can easily do more harm than good by giving someone an untested ventilator.
Alarms are critical, if you don't know that everything is working PERFECTLY then you don't know if one of your 20 patients is dying because of a failing $400 machine.
Also so is lifespam. It would suck to get all these ventilators that break down in a few days.
I've seen medical machine designs, and they have to check by the second that everything is working, and redundant systems so that if one system thinks all is well and the other doesn't big alarms start going on. And redundancy so that if part of the machine is failing the patient is kept alive while another machine is gathered. The quality is just as important as the function.
You just invented an awesome word
From what I've seen doctors actually talk about, is that there's ways to hook up even four people on one of those expensive type ventilators, and doctors still don't do it, because just giving people air isn't good enough.
The thing that makes me feel bad is this feeling it's some sort of off key joke. Maybe it has a use, but to me it seems like why is a big serious company like nvidia, making noise about constructing a ventilator that no doctor would use unless in an absolute emergency where they couldn't use a serious ventilator?
I wonder, what is being done about that?
My understanding is that these aren't standrards you must adhere to, but because they're 'accepted' using them makes your life much easier with the regulators so everyone does unless they've got a good reason not to.
https://www.youtube.com/watch?v=Elgct0nOcKY
https://www.reuters.com/article/us-health-coronavirus-ventil...
https://www.cnn.com/2020/04/22/health/coronavirus-ventilator...
The video in particular is telling, highlighting the difficulty in getting a hospital to change what is considered the standard procedure in the face of new conditions (told by a frontline doc).
We are going to be up to our eyeballs in excess ventilators. Maybe they'll be handy for the next pandemeic ...
Do we have sufficient ventilators for everyone who would need it? Probably. Are there not enough in practice? Yes. Why? We don't know how to differentiate between who needs it and who doesn't effectively. So people are getting put on them excessively to try to save the ones for whom it would help and we have no way currently to distinguish those people better. Should we fill this knowledge gap? Absolutely. I'm sure doctors & researchers are scrambling to do that.
https://time.com/5820556/ventilators-covid-19/
> But Dr. David Hill, a pulmonary and critical care physician who treats COVID-19 patients in Waterbury, Conn. and serves as a volunteer medical spokesperson for the American Lung Association, says arguments against COVID-19 ventilation have been over-simplified. It may be less that ventilators aren’t the proper treatment for coronavirus, and more that they’re not a panacea for a pandemic that has pushed the health care system to its breaking point, Hill argues.
> “You have really sick people, [while] the people who have the best training are in short supply and ventilator management is not simple,” Hill says. If a dedicated lung specialist were available for each patient, he believes, outcomes would probably be better. They could make the subtle adjustments required for effective long-term ventilation, or try less-invasive options and only move to intubation when absolutely necessary. But with many hospitals nearly at capacity, last resorts can become first resorts.
> Few doctors are saying COVID-19 patients should never be ventilated, but there is a growing subset that thinks it’s happening too quickly.
<snip>
> Dr. Ken Lyn-Kew, a pulmonologist at National Jewish Health in Colorado, agrees that there are some differences between classic ARDS and COVID-19, but he emphasizes that there’s a lot of variation among COVID-19 patients he’s treated. He says most still meet the criteria for an ARDS diagnosis. In his view, coronavirus patients likely have ARDS plus other issues, but they still have ARDS. With so much unknown, and with treatment protocols being updated on the fly, he thinks it’s too soon for doctors to go off-book and avoid conventional protocols like mechanical ventilation.
> “The world is not a dichotomous, black-and-white place, but a lot of people are having trouble with that,” Lyn-Kew says. “We might be able to do better, but in the absence of data on the way to do that, we need to follow our societal guidelines and 25 years of research.”
A better reason is because we beat the outbreak. A few areas like Milan/Madrid/Brussels went past their health care capacity and could have used more hardware. But they're all past peak now, and everywhere else (almost -- there are a few worrisome spots in the US still growing, and the developing world is still a big question mark due to lack of testing) managed to reach peak without hitting their limit.
That's good news. We won. It's certainly not a problem that some of the efforts turned out not to be needed. It's like a war effort: everyone had a part to play, we needed to be conservative with our risk analysis and try different things.
What a growing number of doctors are saying is that ventilators often turn out to do more harm than good in treating COVID, ventilators have been used too soon and too often in treating COVID, and ventilators have probably resulted in excess deaths over alternative treatments.
As far as whether ventilators have "probably resulted in excess deaths"... that's just not justified by any science yet. We don't know. There's some suspicion in that direction, based mostly on New York ICU fatality rates (which are quite high despite otherwise good numbers through the rest of the process). But there's equally good evidence from e.g. Milan, which experienced very high death rates due presumptively to lack of treatment hardware.
It's widely accepted and known that fatality rates once you're put on a ventilator (due to Coronavirus) is ridiculously high compared to when you're put on a ventilator for other illnesses (something like 80% vs 20%). It suggests that ventilators might not be the right solution to what's happening to these people. So some doctors have been questioning the wisdom of continuing to use ventilators, and some doctors have been looking to alternative solutions that might have more success and better outcomes. One idea that was floated was CPAP (though I don't know what happened to that).
Nobody here is saying "ah, ventilators aren't working, fuck it, let everybody die". That's offensive to everybody here and a bad faith argument that isn't constructive in any way.
https://www.cnn.com/2020/04/14/health/coronavirus-prone-posi...
As for offense: I apologize again. But the "ventilators don't work" take is getting too much traction among people who do make that argument, and IMHO it needs to be shut down. It's not good science. Not yet. Let the doctors do their work.
Second, at least in the US, it appears things are slowly gaining steam in more rural areas where healthcare is much more thin on the ground. At the same time, there's a push to reopen things, especially in states the skew more towards rural. It might not end up amounting to much, or it could be a disaster in the making, but it's not quite time to say "we won".
That's good news, and needs to be celebrated, so that we can continue the work. Instead, people on the left think "It's still serious!" and refuse to celebrate what we've achieved. And people on the right, not invited to the party, are being led to believe that the lockdowns somehow weren't the victory they actually were.
And that's bad. We won the war. The battles aren't over, but we know we can beat this.
It's problematic because it also didn't seem an accurate characterization of of opinion trends, which shouldn't have mattered one way or another in the conversation.
But my opinion that it wasn't an accurate characterization should not have been an issue in this thread at all: The introduction of politics derailed the discussion from one about progress against the virus to political rather than factual motivation for our opinions. Had the thread started out differently I wouldn't see that as much of a problem, though I think the introduction of political ideology into conversation on HN is often counter productive in general.
I agree things looked terrible then. I think a big part of the conversation now is "how much was lockdown, and how much was bad projections?"
Given what we knew at the time, it was the right choice. Now we need to decide how to move forward.
With all due respect: That's not a conversation among any serious academics. It's a conversation happening among political actors in political contexts (c.f. Tucker Carlson floating this two nights ago).
It's just not correct. And as you phrase it, it doesn't even make sense. "Projections" don't cause a change in R0. If you want to argue that this happened for some other reason than lockdowns you need to have a reason. And there are none that make sense.
(I mean, maybe it's herd immunity. In that case, why did the infection curves in different US states at very different death counts all bend the same way at roughly the same time after lockdown? Maybe it's weather: likewise, also the fact that the early China outbreak with similar behavior happened in the middle of winter. Maybe it's a mutant strain: nope, no particular evidence for specific ones having different behavior.)
If you want to have a conversation, you need to come to the table with a working theory. Social isolation has been known for centuries to work exactly the way it has here. It's a very attractive hypothesis, and is going to be difficult to disprove.
Yes, things are terrible in New York City.
The question, maybe not debated among "academics" but debated among others, is whether spread rates ever would have gotten as bad in suburban and rural areas as they did in densely populated cities.
Models that used the same variables for NYC with population density of 30k people / sq mile as for rural states with tens or hundreds of people per square mile are just not realistic.
The virus is a serious threat. It's good that we took the actions we did. That is in the past. The question now is how we move forward. Part of that is gaining a better understanding of how bad the conditions on the ground actually are. All indications are that we can relax restrictions to some extent and still not overwhelm the hospital system. Indications also point that our current restrictions are absolutely catastrophic for the economy.
So we have to find the balance.
The idea that rural areas don't see rapid growth isn't really well-founded. What happened was that large international cities experienced their outbreaks earlier. Nebraska may be small still, but at their current growth rate they'll match New York's all time high new infection rate (per capita) in just two weeks.
Go browse the charts at https://91-divoc.com/pages/covid-visualization/ and note how similar all these curves look. No one is special. Everyone grows rapidly without lockdown. Everyone levels off with it.
Only a few states have actually seen a peak and a significant reduction in new infections (c.f. Vermont, Montana, Hawaii). Those are the ones that might consider opening up right now. Everyone else just isn't safe yet.
I live less than an hour away from the capital, Des Moines. We have ~30 confirmed cases total in my county. My mom lives near the northern border. They have 2 confirmed cases total.
We never went to strict "shelter in place". Today the majority of the state is relaxing restrictions and restaurants can allow limited dine-in options if they wish.
Maybe it will go terrible. Maybe it won't. We will see.
Would you have recommended that strategy to Lombardy in February?
I mean, you're citing current absolute statistics at a time when your infection count is growing rapidly. Your per capita new infection rate now is the same as New York's was on March 22. You understand that's not the right way to think about an epidemic, right? If you aren't flat then you're in the process of growing exponentially, and that doesn't just fix itself.
It seems you are willfully misinterpreting me. As you will see above, I said:
> The virus is a serious threat. It's good that we took the actions we did.
Things were much less known then. Now we have better ground truth and are less likely to be blindsided. I stand by my state's actions and decisions to date.
I have been exercising caution and following the recommended guidelines. I will continue to. I would guess most people here will as well.
We can't be in lockdown forever. The prudent thing to do, in my opinion, is to loosen restrictions slowly, being fully aware that we may have to tighten them again.
Your state's decisions and policies are producing rapid outbreak growth, though. Why do you expect that growth to change if the policies don't?
Going back to the link above, I see that Iowa's new infection rate (I'm looking at the 7-day average because per-day numbers are jumpy) has grown by a factor of 3.78 in the past 14 days. That's a daily growth rate of 10%, or a doubling period of a tiny bit over one week. At this rate, you will reach the same per capita rate that New York experienced at its absolute peak (April 9th) in just 13 days.
So... what's going to change in two weeks? Your curve isn't bending, your policies aren't changing. What's going to save you?
Here's what won't save you: locking down once things "get bad". Go back to the same site and see how long it takes from a lockdown until a state reaches peak. It's 10-20 days.
I mean, maybe I'm wrong. Maybe there really is some as-yet uncharacterized effect separate from lockdowns that steps in at the last minute just before states reach capacity to halt the outbreak. I just don't think Iowa should be betting on that, and I'm very scared about what we're going to see in the north central plains states in the comming month.
You... genuinely don't think this is going to happen, do you? Even with the numbers above, you just figure something's going to change and you won't need to do anything?
You seem to be going out of your way to misunderstand what I am saying.
Iowa is highly locked down. We simply have not been ordered to "shelter in place".
Most of these restrictions remain in effect. The twenty counties that are experiencing outbreak growth remain completely locked down. The plan is to re-evalute and dynamically change restrictions as time goes on.
The official Coronavirus info page for Iowa is at [1].
Every county that has had restrictions lifted has experienced multiple weeks of declining case growth. Supposing they see an uptick in cases, measures will be re-instated.
It makes no sense to lock down rural counties that have seen zero or next to zero cases. Doing so will not ease the case load in the metros a hundred miles away.
The models probably weren't perfect, but I don't think they were wrong in their general trends, even if the magnitude of those trends would have been slightly different in reality.
In any case, we'll know more in 2-3 weeks if things start to tick back upwards in areas that ease restrictions. Though even then, my concern is that the effect of lifting restrictions will occur too slowly and subtly, and so by the time we know there's a problem it will already be too late to easily contain.
Whether that is true in other countries is a different matter.
That said, I hope the most promising of these designs get carried through to 'production-ready', in case we get hit with something even worse than COVID in the future, which is easy to imagine.
Basically the US govt attempt detailed in this story, but with a sucessful conclusion instead of a failure. https://www.nytimes.com/2020/03/29/business/coronavirus-us-v...
Our experience at MGH also puts severe COVID lung disease squarely into the ARDS category, and it responds to standard ARDS ventilation protocols[1]. Our mortality to date is shy of 20%.
Data over anecdote.
1 = https://www.atsjournals.org/doi/abs/10.1164/rccm.202004-1163...
Is that across all Covid patients? Or ventilated Covid patients?
In the paper I see no direct comparison between COVID response to ventilation vs ARDS response to ventilation, but I think I understand that you say the responses are similar, based on your experience. Am I getting that right?
Also, of course, there is nothing in the paper comparing COVID invasive ventilation outcomes with, for example, oxygen and positioning treatment outcomes. So, interesting, but I don't think this qualifies as "data over anecdote".
I take it you are a pulmonary or ER doctor. If so, can you comment on the widely circulated claim that high death rates on ventilators in NYC are likely in part attributable to very aggressive ventilation and/or reduced attention to ventilation settings due to extreme fear of exposure to the virus (tubes leading from patients to other rooms and the like)?
The high death rates in NYC that were published were attributable to a statistical error (failing to use adequate censoring of people still on ventilators). Note that the article has been updated with correctly censored numbers (hence 25%): https://jamanetwork.com/journals/jama/fullarticle/2765184
Finally, we only intubate patients after maneuvers like awake proning fail, so it’s not something that is comparable.
The original paper is: https://jamanetwork.com/journals/jama/fullarticle/2765184
My guess is that in the final analysis, NYC death figures for those on ventilators will be higher than they should have been (after adjusting for age etc.) The question then will be why.
But the real question, I think, is whether the death rates were higher than they would have been with less aggressive ventilator use.
Even if ventilator supply was not a constraint, would there be other constraints caused by staffing that would limit effectiveness of ventilators?
Note: I am NOT a doctor or involved in health or medicine in any way. My source for this information is this MedCram lecture on Youtube: https://www.youtube.com/watch?v=okg7uq_HrhQ
That sounds more like the typical result for ventilators.[1]
I think it's more like 80% fatality on ventilators for this disease[2][3][4]. According to the published literature of course; I ain't that kind of doctor, but based on what is known, if someone comes near me with one of those things, well, they won't be a doctor for long either.
[1] https://www.sciencedirect.com/science/article/abs/pii/S05315...
[2] https://www.thelancet.com/action/showPdf?pii=S0140-6736(20)3...
[3] https://jamanetwork.com/journals/jama/fullarticle/2765184
[4] https://www.physiciansweekly.com/mortality-rate-of-covid-19-...
Theirs and ours may end up higher than they are now once the final patient is discharged, but the current US numbers are more like 20-25% mortality on ventilators at this point rather than 85%.
We've been a month+ behind on a lot of the response. Now's not the time to talk about ventilators, it's the time to start working out logistics of vaccine production and distribution.
Traditional ventilators (ventilator) are very different from bag valve masks, resuscitators, PAP and CPAP, BiPAP machines.
This Nvidia design is not alternative to that $20,000 hospital ventilator. It's emergency-response ventilator alternative these other cheap designs you can already order from Amazon. You can't keep people long time in this kind of device without serious risk of damage to the lungs.
Not a lot.
Only 4 countries on record now managed to start mass production of designed from scratch emergency ventilators.
Expanding existing capacity would've been an incomparably better option.
Are your criticisms based on an understanding of how this specific device actually works? Sorry to be blunt, but by suggesting it is comparable to a bag valve mask, it sounds like you did not learn a lot about the device before making up your mind about it.
Noting wrong with being blunt. But please, be blunt only after you read whole sentences and trying to decipher the meaning. I listed several different classes of devices of varying complexity that are not equal to those $20,000 ventilators used in ICU.
I have experience in design of electronics that becomes part of ventilators and other medical equipment. GE healthcare and other firms.
>Are your criticisms based on an understanding of how this specific device actually works?
Yes. The system in question is very low-end emergency-response ventilator (explained in documentation). It's simplified design from those you see used in emergency response and patient transport. The operating mode seems to be basic pressure-controlled ventilation PCV.It can also measure tidal and minute volumes.
ICU ventilator is device is very different from everything else. It's connected to the patients trachea. You must sedate patients because it's super uncomfortable to use. Patients are put into those ventilators when they are really sick. Removing carbon dioxide is an issue. Long term use can damage lungs so the system monitors large number of variables and it has multiple operating modes.
For what it's worth, I did read your comment carefully. The reason I took exception (and reacted strongly, I fully admit) is that since the article addressed the bag thing, that makes it a red herring, so I would have expected you to filter it out of your list. Maybe that's overly rigid or overly literal of me, but that is why I reacted how I did.
This device definitely intends to be less than a full-featured ventilator. On the Pneumatics page (http://op-vent.stanford.edu/docs/pneumatics.html) it explains their goal is to produce kind of a stripped down "simple ventilator with the features needed to treat COVID-19".
So they believe it can be useful even though they know it lacks features. Whether they're right is certainly up for debate. They mention several optional features, which is interesting because maybe it means they're just trying to be flexible or maybe it means they don't really know what is the minimum necessary to have a net benefit (save more people than you kill) if we run out of full-featured ventilators.
There are plenty of proven designs that can be manufactured cheaply, the price of the end units especially for things like ventilators isn't often due to complexity and manufacturing costs but due to the costs of maintaining a supply chain, manufacturing and providing support (including liability) for the healthcare market.
None of these things seem to solve these issues other than potentially removing liability to who ever decides to manufacture those units.
Any thing can be built with off the shelf parts but the reason why most medical devices aren't it's because you need to take special care in your supply chain.
Ventilators use valves that aren't that different and often the same valves that are available for a plethora of other applications like industry and agriculture but the fact that it's the same valve doesn't mean that it's the same supply chain.
The suppliers for medical equipment make sure that say the lubricants that are used are safe and that no one can say replace o-ring used in the valve from one supplier to another without recertifying the entire thing making sure that the new o-ring meets the exact same specs.
For things like electronics then you have a whole other world of safety from things like resistance to liquids (anything from blood splatter to leaking IV bags), EMI to make sure it won't interfere with any other devices or be susceptible to interference on it's own and ESD to make sure there could no chance of sparks especially in devices that pure or high concentration oxygen can flow through or devices that will be near oxygen feeds.
And heck even if all that isn't the reason for ventilators being expensive and the true reason is IP Licensing/Patents then the US government and every other government can easily either compensate the IP holders directly or simply withdraw protection on the IP in question.
Many countries have already had legislation in place to allow governments to suspend IP protection or grant protection to violators during an emergency.
There will be lots of wasted effort during the sprint to plug things in a hurry. But it may not be really wasted effort, given that projects like these can be fountainheads to unknown future innovations or relations.
Other countries simply took existing designs for ventilators and shifted manufacturing capacity e.g.:
https://www.timesofisrael.com/iai-defense-ministry-inovytec-...
So overall I don't see why the "design" was ever the bottleneck.
I agree with this comment 100%. What the world needs now is more of the ventilators we already have designed, not new designs. With any new medical device comes a bunch of teething pains, finding and fixing bugs in the hardware and software design, training, testing andd iterative improvement of all these things as new issues are identified. All of this requires significant clinical support and companies that do this stuff typically employ dozens of clinically trained folks to aid in the introduction and development of the clinical aspect of the device. All of this time and effort is just not available now, any and all clinical resources need to be focused on treating patients, not helping engineers find bugs in their systems.
I believe this because an alternative exists, we do not need to reinvent the wheel, we just need more of the wheels we've already invented.
To be clear, I do however believe there is a lot of value in investing heavily and exploring all potential novel treatments and devices that could help fight this. I just don't see a value add to designing new respirators vs. Just making more of the designs we have now. I could be wrong, there could be significant manufacturing or other clinical advantages which i am not aware of.
As we learn more we might discover things that make the current ventilators not optimal at which point it is worthwhile to design something new.
The ventilators we have aren’t the problem and their cost and design also don’t impact the production capacity that much what impacts it is the fact that most medical manufacturers never had to operate on such scales and those that even come close tend to big the big international players with global manufacturing that is heavily reliant on China.
There is a plethora of designs for ventilators of all types from emergency CPAP machines to multi-mode life support for ICU’s.
Those designs have already been validated and there is a huge bank of knowledge to support them.
All these pump/turbine based respirator not to mention the pneumatic Ambu bag auto-squeezers don’t really bring anything new to the table they just introduce an untested hardware and software.
Like seriously half the ventilators that were shown were an ambu bag and actuator and a micro controller attached to some pressure sensor yes it’s a very simple respirator but do we really need 50 university teams coming up with a design variation?
And as far as discovering something new the only thing we discovered so far is that people were put on a ventilator and even more so intubation way too early and that has negatively impacted their survival rate which is one of the reasons why the demand for ventilators has dropped.
Everyone of these ventilators is basically an existing design just with different off the shelf components.
None of these projects created a new design or a new ventilation mode.
The vast majority of these are CPAP machines some maybe able to provide APAP in firmware.
If you look at the designs they are basically split into two categories the ambu which just basically compress an off the shelf ambu bag with an actuator or pump/turbine based machines that are essentially very similar to those which are given to people suffering form sleep apnea.
I don’t think there is even a ventilation mode we haven’t thought of yet even the most advanced ventilator on the market are mechanically very simply machines the cost is usually based on the brand and which modes they support.
This isn’t a field in which you can have a major breakthrough and not to dismiss this effort it’s not a field in which a breakthrough is needed.
Ventilators are cheap by medical standards even at $20,000 a pop the cost isn’t the issue the issue is that simply we never needed to produce them at such quantities in such a short time frame and for that you don’t need a new design you just need to accept that you aren’t going to reinvent the wheel and just take a proven design and make more of them and sort the licensing after this.
The fact that some teams had a working prototype within a day should be a very strong indication that there isn’t much here.
Having a turbine and a microcontroller to set the positive pressure on the outlet or having it set up how many times per minute an ambu bag should be compressed and at which rate isn’t a particularly difficult engineering challenge to conquer.
This is why there are now 100’s if not 1000’s of virtually identical designs out there from various teams across the world.
Yes anyone can build them in their garage and yes if the world have ended it would’ve been better than nothing but we have real options out there that can be just as easily manufactured if we actually had the will.
Forget about that, I'd expect something from Nvidia to at least be able to run Doom.
It's true that the durable medical equipment (DME) industry is ripe for disruption. But that disruption isn't going to come from cheap stripped-down equipment, even if that equipment has stunningly wonderful software in it.
DME industry disruption will come from figuring out how to break the innovation-restricting stranglehold of Group Purchasing Organizations on health care supply chains. That's harder to pull off than quick-turn prototypes. What's needed is market disintermediation. That's a political and financial challenge, not a gadget-creation challenge.
Somebody once said to me, "software is the most complex thing ever invented." But that can't be right: look at a modern airplane. It has software in it, but it has plenty of other complexity. DME is the same way.
Creativity like Dr. Daily's is great. Really great. But it alone doesn't solve hard problems. If we software types claim it does, we're setting people up for disappointment.
[0] https://www.global-medical-solutions.com/Allied-Healthcare-M...
This is the kind of work I would like to see the best minds of my generation to work on.
I am not saying that I am one of the best minds of my generation, but I couldn't work for ad-network companies. I took consulting gigs only when I was trying to gain some big data exposure, and was in that field only for three years over three different occasions.
I would like to see more and more people to stop working for ad-tech companies.
namely lack of competition, regulatory capture, and price not paid directly by end users, for the most part.
This isn't really relevant here, unless you are trying to create a strawman or frame this as a false dilemma of extremes. You can have a sense of humor while also keeping the conversation substantive.
[Edit: Assuming solar is correct: ]
Comments like the above are cancer when they inspire more of the same, and not long ago would have been appropriately dis-incentivized by the community.
[And if solar turns out to be wrong, then replying in earnest is still a good response.]