Hacker News new | comments | ask | show | jobs | submit login
The Colgan Air disaster was a milestone in aviation safety (bloomberg.com)
231 points by jaredwiener 6 days ago | hide | past | web | favorite | 159 comments





One of the biggest changes that came out of the Colgan crash is the so-called "1500 hour rule." Before Colgan, only the captain needed to have an airline transport pilot (ATP) certificate, the highest level of pilot certification. First officers (copilots) could fly with a commercial pilot certificate, a lower grade of license. Now, both pilots are required to hold ATP certificates.

ATP applicants need to have 1500 hours of flight experience—hence "1500 hour rule"—though there are exceptions if you've attended specific approved aeronautical schools or have military experience. Commercial applicants need only 250 hours. (Here too there are exceptions for certain training programs.) The standards on the flight tests are tighter for ATPs and the written exams are worlds apart.

As an instructor I can tell you the difference between a 250-hour pilot and a 1500-hour pilot can be enormous. Not always—some pilots have 1500 hours, and others have the same hour, 1500 times in a row—but there can be a huge difference.

This rule has been controversial though, not least of all because both pilots in the Colgan crash already had more than 1500 hours. And the FAA made so many changes after Colgan that it's been hard to tell which of those changes actually made the difference, and which, if any, were simply no-ops. Critics of the rule have especially focused on how many flight tests the Colgan captain failed, and how the response by the pilots to an aerodynamic stall was completely backwards. Aerodynamic stalls are something you learn in your first ten hours of training as a private pilot and drill constantly throughout your career, so to get that wrong at the air carrier level speaks to an incredible failure in training and evaluation.


It also isn't clear (aside from military) how people are meant to fund/attain ATP status.

There's limited jobs with which a commercial licensed pilot can gain hours (e.g. crop dusting, sky jumpers, instructors, etc) but the number of people trying to attain ATP/jobs requiring ATPs well outstrip the number of jobs to support those people getting to 1500 hours.

As you said, previously some airlines would take the pressure off by hiring commercial pilots and getting them hours as copilots. With the new rules they need 1500 before day one, so that has to come from somewhere and it isn't clear where.

This is one reason airlines are struggling to find pilots.


Yes, exactly. The 1500 hour rule is a factor in the pilot shortage. Not the only factor—low pay at regionals, generational demographics, and changing perceptions of the career also play a part—but it's definitely a big factor.

How can there both be a shortage and low pay? Shouldn't the need for pilots drive up pay as demand for air travel itself isn't going down.

That's paradoxical seeming isn't it?

But low-pay + labor shortage is seems to be a common thing.

I would speculate it's driven by the production of a very price sensitive commodity. Lots of people choose the cheapest flights possible regardless of other considerations so a significant rise in pilot pay might make a given airline unprofitable since they can't raise prices to compensate.

So airlines feel strong pressure to keep wages down.

And really, you could say a "labor shortage" is what every high volume, low margin operation complains about but the reality is basically "we pay low wages so it's hard to get enough people to work for our low wages". (Obviously, with airline pilots, it's low wage relative to the cost of gaining the skills etc). We can see similar language with H1b visa proponents and even fast food restaurant operators use it periodically.


It doesn't matter how price sensitive your product is, you can raise prices a little and pay people a little more to help with hiring. There's a point of equilibrium somewhere.

"Labor shortage" is just a tactic to get someone else to help with the problem in a way that doesn't require increasing pay. It's the hiring equivalent of "don't miss this exclusive sale!"


I'm not sure if the cost of flights would really go up that much.

Just to play with some numbers, let's consider a pilot with a salary of an astronomical $1,000,000/yr who flies transatlantic flights on wide-body jets (300 or so passengers). She is limited to 1,000 flight hours/yr, so each hour costs the airline $1,000. For an 8 hour flight, $8,000 split between 300 passengers is about $26 per person. Given that these flights come in around $150 at the cheapest, often going up to $1,000 at peak times, this (quite high) wage estimate doesn't translate into a huge price for the consumer (especially since you could shift more of the cost towards the first- and business-class passengers).

I think the numbers work out similarly for regional flights which have fewer passengers but shorter time in the air. These pilots start with base salaries in the $30,000 range; if that were raised to $100,000 for flying (for example) 3 hour flights with 50 passengers, that's only ~$4 extra per passenger per flight.

There might be other factors I'm missing - do these pilots actually fly significantly less than 1,000 hrs/yr?


You should check this video out: https://www.youtube.com/watch?v=6Oe8T3AvydU

You would be pretty surprised how many costs airlines have to pay. An extra $4/passenger/flight would eat significantly into their bottom line. This is especially true for budget airlines, which often already make less than this per passenger already (source: https://www.youtube.com/watch?v=069y1MpOkQY)


Your $4/pilot/flight misses the overall point. A company in a highly competitive industry is going to be minimizing costs across the board. To keep a plane in the air, you need mechanics, pilots, flight crew and so-forth. You could be completely spendthrift with just pilots and hold the line on everything else but then the potential for the rest of your personnel to be jealous is high. Moreover, what's the motivation?

"You cheap bastards could easily afford XXX" misses the point. A large company runs on general principle, and company that's producing at minimal cost is going to run on the principle of saving money everywhere.


> Lots of people choose the cheapest flights possible regardless of other considerations so a significant rise in pilot pay might make a given airline unprofitable since they can't raise prices to compensate.

In addition, certain airlines have the backing of massive nation state funds or are otherwise simply enormous - just like cheap venture capital, they can simply drown smaller competitors, buy their slots and airliners for a penny on the dollar and then jack up the prices. Lufthansa and AirBerlin are a perfect example.


It's very simple. There's not actually a labor shortage. If there were, the flights would constantly be sold out and there would be empty planes on the ground due to no pilots.

" Lots of people choose the cheapest flights possible regardless of other considerations ". so true, and then they complain when there is extra cost for bags etc. it's the customers that drive the airlines to eliminate all frills or things that use to be 'free'.

Probably one factor is the extreme time lag between someone becoming interested in becoming a commercial pilot and actually becoming a commercial pilot (for example, because of the need to get so much experience). It might be similar with doctors, where the required credentials take years to achieve—so if people say "hmmm, I might want to go into medicine, I think I'll do pre-med" they could still be more than a decade away from practicing medicine.

It seems from upthread comments that the institutions to train prospective doctors are somewhat better-developed than the institutions to train prospective commercial pilots. (My understanding is that this used to be primarily military, but now pilots are coming in through a greater range of paths.) If people think they want to be doctors, they can apply to certain schools, get certain loans, participate in an established process to find internships (which their schools will guide them through), and so on. I don't know that there are really institutions to help the pilots this way.

That doesn't mean that supply and demand somehow don't apply or don't influence people's behavior, but a need for massive amounts of training and credentialing (literally years' worth that themselves come at a high financial cost) surely slow down the response.


If it takes a long time for additional supply to become available the price spike in response to demand changes should be even more pronounced.

This is the expected market reaction when the market is free, but if there is a oligopoly effect in play it can distort this and defer that response.

The average age of airline pilots is going up very quickly though, so these terrible practices will eventually illicit a market response, since nobody (or very few people) actually wants to do this.


> will eventually illicit a market response

Completely off-topic and unrelated, but on the small chance that this helps you (or anyone else reading): illicit is something forbidden or illegal, while elicit is to evoke a particular response.


I think a large portion of the issues surrounding pilot pay are

1. Airline companies are few and far between, it's basically an oligopoly and getting black listed by one company might impact your job prospects in another

2. Pilots invest a lot of time into qualifying for their jobs and the sunk cost fallacy traps them into looking for jobs flying planes.

Also, there might be some strong self-motivated reasons, not wanting airplanes to go without pilots due to altruism, just generally enjoying flying... These aren't IMO strong financial drivers so they seem like weaker points, but they may provide some influence.


Something I haven't seen brought up: if pilots are passionate enough about flying, then that's what they're going to do no matter how little they get paid. Companies don't have an incentive to give a raise to people who are not considering a different job.

Two words: unions and seniority. Here's a good intro:

http://philip.greenspun.com/flying/unions-and-airlines

It's also worth noting that this is largely a US-only phenomenon. It has led to a lot of junior US pilots heading overseas to fly with other carriers, most notably in the Gulf, although that party is slowly starting to wind down.


Please have a read of the article before downvoting: it's not some random neoliberal "unions are evil" rant, but a very coherent explanation of how seniority-based pay works and how the system is set up to ensure senior pilots get paid $300,000 while juniors get $16,000.

Luckily a pilot license can be used anywhere in the world so people can find a job abroad. At some point when US airliners can't find enough pilots something will have to change.

Given that the author's experience (as referenced in the article) covers a period of time before the Colgan reforms, I'd expect that a decent amount of what he's written is outdated.

There is arguably a shortage of pilots with more than 1500 hours and good salaries for them. There is a glut of pilots with less than 1500 hours, and bad salaries for them.

Pilot unions still exist even for these regional carriers. How has collective bargaining worked so well for 737 pilots but not for a regional pilot?

Same way that there's a shortage and low pay for emergency medical services staffing, unfortunately.

Market response of raising pay is always slower to respond if there are other mechanisms to try first.

> This is one reason airlines are struggling to find pilots.

There isn't really a pilot shortage. Merely a shortage of pilots willing to work for slave wages. There are US pilots flying all over the world because they don't get paid what they're worth in the US.


Agreed. There are pilot jobs in the US that pay great and offer a very desirable lifestyle. However, these jobs account for what, 30-40% of US pilot jobs? The rest are fighting for scraps making $50-60k/year while paying back $100k+ in student loans and never seeing their homes. There’s a reason the FO on Colgan 3407 commuted from her parents home on the other side of the country and the CA slept in the break room the night before the trip started. They didn’t make squat and couldn’t afford anything else.

Quick reminder: "slave wages" are exactly $0, and also you can't leave the job due to threat of violence. (And also there's a lot of violence anyway, and also bad housing and bad food and complete lack of personal time or space or privacy or any freedom of choice, etc.)

It sounds like regional airline pilots may have surprisingly low wages given how highly-skilled their jobs are. "Low wages" is a good term for this.


It isn’t slavery, indeed, but waged can get lower than low. Pilots can be pressurized in accepting very low wages for the right to fly an airplane.

Reason? Pilots have to make flight hours in the plane they’re licensed for to keep their license.

So, you are, say, $100,000 in debt to get a license to fly a 737, and you need a few hours in it this month to keep that license. Renting said plane for a few hours is very expensive. That makes an $0 an hour ‘job’ flying for a few hours look mightily attractive.


"Slave wages" relates to wage slavery, not regular 18th century America slavery. Slave wages are so low they barely let you survive, and give you no way for saving up - essentially enslaving you in the regular sense, with threats of violence being fulfilled by the economy at large, instead of the employer.

> As you said, previously some airlines would take the pressure off by hiring commercial pilots and getting them hours as copilots.

Do copilot hours really count? I thought you had to be pilot in command?


> It also isn't clear (aside from military) how people are meant to fund/attain ATP status.

There is a boon to unemployed, retired US military pilots. There are tons of them, even here in China.

Feels surreal.


This. My son dreams of becoming an airline pilot. But I find it difficult to rationalize investing more than $100k to satisfy the 1500 rule just to see many pilot jobs in the future replaced by automation.

Buy an arplane. This is what I did: bought a simple Cherokee for $22,000, got my ratings (Private, Instrument, Commercial, CFI, CFII), then went to work to Alaska, sold the Cherokee (for $21,000, airplanes do not depreciate that much), which paid for all my training, got a lot of hours, got an ATP, floatplane, and commercial helicopter.

On the other hand, I usually try to talk people out of becoming a professional pilot: one health hiccup and your medical certificate is in jeopardy.


> On the other hand, I usually try to talk people out of becoming a professional pilot: one health hiccup and your medical certificate is in jeopardy.

Is there insurance coverage for this? Something akin to disability insurance for this specific scenario?


Not really. At major airlines there are special programs that help with this (even with DUI's surprisingly) mainly because pilot unions make it part of the contract. For a corporate pilot, there is rarely any recourse. one other thing to remember is that the environment in the aircraft is not very friendly: very low humidity, constant changes in pressure (even if the plane is pressurized). Flying a lot of passengers also means you are exposed to a lot of germs. On top of it all, add constant lack of sleep, sedentary lifestyle, etc. You have to be very proactive to stay healthy.

The ‘legacy’ airlines like American, Delta, United , Alaska, Hawaiian etc provide quite good disability programs. Additionally, ALPA, the pilots’ union, offers a loss of medical insurance. Often times a loss of medical is not something that was also debilitating. There are plenty of other jobs for airline pilots to do in training departments and other areas in flight operation requiring the same knowledge base.

Edit: what I meant is that working in Alaska paid for my training, not selling the airplane.

What kind of work in Alaska?

Flying freight and surveys. C206 at first, then the usual progression to Navajo, Metro ...

Do you mean what kind of flying work?

"Not always—some pilots have 1500 hours, and others have the same hour, 1500 times in a row—but there can be a huge difference."

This is an excellent statement emphasizing the importance of concepts like deliberate practice and a varied curriculum/experiences. I think this idea applies to all learning.


Definitely relates to software dev where some developers have 15 years experience and some have 15x 1 year of the same thing.

In reality I think most developers have careers that look like:

* 2-3 years of explosive growth and exposure * 5-10 years of slower growth and increased mastery * then diverge to shorter periods of stair-stepping (new technologies / faster mastery), coast on what they already know well or move out of active development


Teachers can fall into the same category. A teacher for 15 years or a teacher for 1 year 15 times.

By the nature of their jobs, teachers often teach the same 1 year 15 times...

Also the importance of regular evaluation. You can develop a lot of bad habits in 1500 hours.

As my band conductor drilled into us:

> Practice makes permanent. Perfect practice makes perfect.

Of course "perfect practice" is an ideal, and you can't achieve it yourself without having someone or something assist to evaluate in most cases. But repeating the same mistakes for 1500 hours just makes you really good at making those mistakes.


I don't think this is a good general saying, most activities that we're performing have discoverable local maxima we can achieve... or most people will achieve, by varying their technique to observe how much better/worse easier/harder the task becomes. That said something like an instrument that is very highly tied to developing memory muscle is definitely something you can learn _wrong_. Similarly any activity where actual execution of the task is expensive/infeasible (combat for instance) also benefits a lot more from instruction.

In their defence, on a large turboprop like the one they were flying (Q400?) an iced over horizontal stabilizer feels similar to a stall but the response is entirely different.

They were hopelessly tired and distracted. The captain apparently ignored both the stick shaker and the stick pusher (when the aircraft is trying to get the pilots attention). Sometimes experience helps with self-evaluation but more often it does not, depends on the person. They should not have been flying that night (the first officer had not slept for 36 hours).


> This rule has been controversial though, not least of all because both pilots in the Colgan crash already had more than 1500 hours.

Is there a term for this: reactive legislation that wouldn't have covered the incident it was in response to?


Interesting question! There is probably a name for that kind of fallacy. The regression fallacy is close: "Ascribing a cause where none exists in situations where natural fluctuations exist while failing to account for these natural fluctuations."

https://www.logicallyfallacious.com/tools/lp/Bo/LogicalFalla...

The term "magical thinking" is a broad label that could apply.

https://www.logicallyfallacious.com/tools/lp/Bo/LogicalFalla...


Typically when you use a crisis to lobby for legislation, you come with a whole bag of items that are useful to the general purpose, but are perhaps not directly relevant to the incident the response is directed at.

That's a good point. In my head the "1500 hour rule" is a distinct thing, but it actually came in as a whole set.

Never waste a crisis.

"Shotgun approach"? Where you react not only to the proximate cause of the incident, but to all possible proximate causes?

How does a pilot get 1500 hours if [s]he can't get an ATPL without 1500 hours? Non-passenger/cargo commercial flying?

The 1500 hour rule only applies to air carrier operations—scheduled airlines. A commercial pilot can fly charters, survey, banner-towing, skydiving, sightseeing, demos for sales, even corporate (business jets). Add a flight instructor certificate and they can teach, logging all their teaching time.

>skydiving

You misspelled "elevator operator"

I used to work at an airport that had a skydive operation, we jokingly called them elevator pilots (because they just go up and down all day). They used to only pump a the required amount of fuel to get the aircraft to altitude and idle back because carrying more fuel up and down is a waste of fuel. Occasionally they would run out of fuel and glide in. One time they ran out of fuel really short and landed in the pond.


I used to fly out of an airport that had the same frequency as another airport that was a very popular skydiving destination (Lodi, Ca). One of the pilots for that outfit was on frequently; weekday, weekend, it didn't matter. Same radio calls, same slight drawl. Day in, day out. He sounded completely bored out of his mind. Even the announcement that he was dropping meat bombs ... "Skiy-diivers in the skieee over Low-die Airport, 18 thousand feeet.." like he'd said it ten thousand times. And I think he may have!

I find that if the airport is quiet you can hear the skydivers screaming, then opening their chutes, then more screaming.

I certainly yelled as loud as I could on the way down.

Were they actually at 18000 feet? I'm not a skydiver myself but I thought they usually jumped from lower to avoid the need for supplemental oxygen.

I think you're right. 12 or 14,000 feet perhaps.

This sounds like excellent practice compared to something where much more of your time is in level flight.

I love them. Without exception they are absolutely nuts. :)

Is this why commercial carriers pay so much better than all the rest? Or is this no longer as true?

Also, is the influx of commercial pilots into the western markets from other countries analogous to the software development "We can't find enough good developers / You don't want to pay what they demand" trends we see in tech?


> is the influx of commercial pilots into the western markets from other countries analogous to the software development "We can't find enough good developers / You don't want to pay what they demand" trends we see in tech?

If by "western markets" you mean the US, it is probably the other way around. There is a shortage of pilots in the US because their citizens are flying in other parts of the world where they get paid much more than they would in the US. And unlike in tech, when a pilot changes employers he starts again at the bottom of the seniority list which means lower pay, the worst routes, last pick of vacation days, etc.


You can fly with passengers on a commercial license. There’s lots of non-scheduled and non transport category aviation in part 91 and part 135 including private jet flights, various kinds of charters, deadheading, flight instruction, and so on.

Being a military pilot is one of the most common ways. Although being a military pilot is such good training in general that you could probably get a much better-paying job doing almost anything other than an ATP.

Such as?

Do you have an opinion on to what extent simulators can help? I've heard a range of opinions on this, and while there is no substitute for the real thing, I can't help but notice that full flight simulators seem freakishly good these days to the point where many pilots say their minds are often tricked while in the simulator. How much can commercial airlines rely on increased simulator time to help safety and apply hours to regulatory requirements?

I've not had the pleasure of flying "full flight simulators" (the kind used by airlines) so I can't comment on those. The lower grades of simulators are good for drilling procedures but are quite bad at modeling flight dynamics.

Tons. Commercial flight simulators are very good. They help with mundane stuff like the layout of the cockpit or familiarization with different airports (hong kong is a must) And the harder stuff pilots could not train for in real planes like equipment failures, maybe all of them. And yes, their minds can be "tricked", meaning pilots do come out of simulated emergencies sweating.

As a PPL-IR ASEL, I've wondered a lot about this. Does an extra 1250 hours of watching student pilots put flat spots on tires and inadequately use right rudder to counter act undesired yaw really make someone a safer pilot of a large, multiengine turbo{prop,fan}?

Also, as others have noted, both pilots of 3407 we well above 1500 hours


> to get that wrong at the air carrier level speaks to an incredible failure in training and evaluation

The problem is pulling up in case of a stall is the natural, intuitive thing to do. In a high stress situation, one is liable to revert to it. This behavior is the cause of many, many crashes.


Re flight hours, I’m reminded of this post from Phillip Greenspan:

http://philip.greenspun.com/flying/foreign-airline-safety

“Foreign Airline Safety versus U.S. Major Airlines” (2009, updated 2011.)


> Aerodynamic stalls are something you learn in your first ten hours of training as a private pilot...

In fact the CFI who taught my introductory flight demonstrated a stall and recovery.


This is an excellent example of how regulation should work as opposed to how it works in regulatory capture situations.

In a prisoners dilemma like set up, if any airline unilaterally changed their rules for pilots and crews, they would "lose out" when they didn't have crews to fly, while other airlines would "win more" by overworking their flight crews. The only "win" was for all airlines to not implement these changes.

But when the changes were forced on everyone, it removed the advantage of not following the guidelines (well it added a criminal or civil prosecution risk) and so all airlines have at least minimally rested crews.

That message seems to get lost sometimes.


I would say this is an excellent example of how regulation shouldn't work.

Even before the Colgan Air crash, it was widely understood that regional airlines' overworking and undertraining of their pilots was risky. But the regulators either weren't able to do anything about it in the face of industry opposition, or didn't care enough to do so. It took a plane actually crashing, 50 people dying and Congress, which was lobbied hard by the families of the victims, passing new legislation ordering them to do something about the problem to get them to take action.

It's great that we have regulations protecting passengers from these problems now. But it would have been much better if it hadn't taken the deaths of those 50 people to create an environment where it was possible for such regulations to be enacted. Good regulators protect people before there's blood on the floor, not after.


Though I do agree with the a lot of what you're saying it does seems like what you are suggesting is that had the regulations worked "as they should" then we could have always prevented accidents like this. I'm not sure that's realistic.

> Good regulators protect people before there's blood on the floor, not after.

So to be good regulators, they have to be essentially omniscient?


No. Post-mortems of these types of disasters usually reveal that warning signs were visible if anyone cared to look or act.

One example happening now: extreme, pervasive sleep deprivation in the US Navy. We've already had disasters that could've been prevented if someone talked to even a single sailor and realized how dangerous that is.

Another example is self-driving cars. We just had a Tesla crash, and yet Tesla will not be regulated properly and will likely kill someone soon. Arguably, they already have.


If the regulations had been enacted before the incident, how would anyone know whether or not they even worked?

On youtube now you can find flight simulator reenactments of most every crash and air incident. They are quite fascinating and much much better than TV style dramatizations.

Here is this incident: https://www.youtube.com/watch?v=LzY-hzxlqig


The air traffic control recordings alone are really great. When the guy stole a plane from Seattle airport to take it for a joyride, i got stuck and clicked myself through what felt like half the ATC recordings on youtube. Can only recommend it, the level of international communication is rather astonishing.

https://www.youtube.com/channel/UCuedf_fJVrOppky5gl3U6QQ


The most significant regulatory change following the Colgan crash was the added requirement that all airline (Part 21) first officers possess an Airline Transport Pilot (ATP) certificate prior to hiring. Previously, only captains (as pilots-in-command) were required to hold that certificate, which requires a minimum 1500 hours of flight experience. [1]

In practical terms, this wasn't a big change (insurance underwriters already required it), but 1500 hours is a significant hurdle at at time when the US is facing a shortage of pilots.

[1] https://www.flyingmag.com/news/faa-finalizes-atp-rule-first-...


It's not required prior to hiring; it's prior to flying your first revenue flight. The distinction is somewhat important as it's not as easy to get an ATP (any more) with the introduction of the ATP-CTP requirement.

Though I have no need for an ATP, I considered getting one under the old (quite easy) rules.


It was most definitely a huge change in practical terms. When regional airlines began their boom cycles they frequently lowered their minimum requirements to essentially only requiring a brand new commercial pilot certificate. In theory that could be as low as 275 hours or so. I know many people were hired in the 300-500 hour range.

Raising the bar to 1,500 hours drastically changed the hiring landscape and coupled with mass retirements due to the age 65 rule (retirement is required at age 65 for 121 airline pilots), the airlines have had to improve working conditions to continue attracting qualified candidates.

All you have to do is follow the money. The regional airline association of America lobbies extensively for lowering the bar again to save regional airlines money.


Compare this to cars. Every year in the US there are about 6 million car crashes, 3 million injured people, and about 33,000 dead people (so, imagine a city like Dover, Delaware being wiped off the face of the earth every year).

The most typical causes are alcohol, speeding, and reckless driving. In addition, seat belts cut the risk of death by 45%. So, if we wanted to keep people safe the way we do for airlines, we would just attack these four problems.

There are many solutions, but the simplest ones would involve 1) a mandatory breathalyzer, 2) speed limiters, 3) sensors that shut down the car when reckless driving is detected, and 4) shutting down the car if seatbelts are not used. These all exist today, and would save tens of thousands of lives a year, and prevent millions of casualties, lawsuits, traffic jams, etc.

Why don't we do these things? My theory is the illusion of safety. In a car, you're wrapped inside 2 tons of steel and plastic, and you feel safe. Even if you know other people are dying inside, you feel like it won't happen to you. So we don't worry, so we don't care about changing things to save lives, because it'd be an inconvenience. But in an airplane, you're not in control; some pilot is. And you're hurtling along at at 500 miles per hour, 40,000 feet above the sky. That's scary. We better make sure those planes are safe.


I think another aspect of it is the auto industry's reluctance to admit that cars are dangerous. Before seatbelts were introduced, manufacturers didn't want to include any overt safety systems, because they introduced the idea that we needed protection from the cars. There were no dangerous cars, just bad drivers. There's a great 99% invisible episode about it: https://99percentinvisible.org/episode/nut-behind-wheel/

> There were no dangerous cars, just bad drivers.

i remember reading about airplane accidents almost always being labeled as "human error", as it's cheaper to replace pilots than it's to replace whole airplanes.


Interestingly enough, as you'll learn in that episode, the push for automobile safety was inspired by innovations in airplane safety!

> Why don't we do these things? My theory is the illusion of safety.

Because all those cars aren't carrying paying passengers. The rules are much stricter for vehicles that do. Similarly, the rules for private aircraft that don't carry paying passengers are much looser than those for commmercial aircraft that do. This has nothing to do with an "illusion of safety"; it's a simple principal-agent problem.

Are you advocating that private vehicles (cars, airplanes, whatever) that don't carry paying passengers should be held to the same rules as commercial vehicles that do?

> we don't care about changing things to save lives, because it'd be an inconvenience

No, it would be much more than an inconvenience: it would be a drastic increase in the power of governments that are already too powerful. Who is going to mandate all these things? Who is going to decide what counts as a proper breathalyzer, proper reckless driving detector, etc.? If you think a government regulatory agency can be trusted to make those decisions responsibly and not get captured by the industry, I have some oceanfront property in Montana I'd like to sell you. I don't want some regulator mandating such "features" in my own car. I'll take responsibility for safe driving on myself, thank you.

The question that should be asked is, why do we allow such power to regulatory agencies for commercial vehicles? After all, as someone else commented upthread, a regulatory agency that was protecting people's interests shouldn't have needed a fatal accident to convince them to impose stricter requirements; so the regulatory agency is not a perfect solution here either. And the answer is, as I said above, that there's a principal-agent problem in that case that doesn't exist for personal vehicles, and nobody has come up with a better way to address it. But there might well be one that we just haven't figured out yet.


> I don't want some regulator mandating such "features" in my own car.

I hate to tell you this, but unless you bought a pre-1966 vehicle, there are already regulators mandating safety features on your car.

> I'll take responsibility for safe driving on myself, thank you.

As long as someone else is out there who can run into me through no fault of my own, I'd like some assurances that their vehicle has safety features that will protect me.


> there are already regulators mandating safety features on your car.

Yes, I know that. That doesn't mean I agree with it.

> As long as someone else is out there who can run into me through no fault of my own, I'd like some assurances that their vehicle has safety features that will protect me.

They won't. Even with all the safety features that are mandated, someone else can still injure or kill you in a collision through no fault of your own. The main things protecting you are safety features in your car, but if they're actually going to protect you, why do you need regulations to get you to buy them? Wouldn't you be willing to pay for them anyway?


> the simplest ones would involve 1) a mandatory breathalyzer, 2) speed limiters, 3) sensors that shut down the car when reckless driving is detected, and 4) shutting down the car if seatbelts are not used.

1) Some (many?) states require drivers with a history of convinced DUI/DWI to get these systems installed. From articles I've read, they're rather a PITA -- it takes a lot of breath pressure, and you may be prompted at random times as you drive down the road. This seems reasonable for someone with a history of drunk driving, but for the majority of drivers who don't drive drunk, it would be a major turnoff; people would likely just drive pre-mandate cars for as long it takes to get the mandate overturned.

2) Let me know when you have a system that can determine what speed is "reasonable or prudent having due regard for weather, visibility, the traffic on, and the surface and width of, the highway". If you want to put a limiter at 100 mph, that's OK with me -- setting up the gearing so above say 80mph is unattainable seems reasonable too. It's very rarely reasonable or prudent to travel at that speed on public highways. The posted speed limit may offer a suggestion of what's reasonable or prudent, but it's very often not the case.

3) What do you propose here, that doesn't shut down my car when I'm avoiding a road hazard? Hooning feels like something where you can know it when you see it, but may be hard for a sensor system to detect accurately. Doing donuts in a field with no one around isn't unsafe. Doing it in the middle of an open freeway is. If we were able to compute this type of decision, we wouldn't have human drivers, and hooning would be dead. It also presupposed that there's a reasonable way to shutdown a car without consent of the driver that's safer than whatever the driver is currently doing.

4) I can't believe people would drive very far with all the godawful alerts that go on when you drive any distance with the seatbelts off. Again though, it's a question of is it safer to stop the vehicle without the consent of the driver or to let them continue to drive and annoy them into compliance.


some (many?) states require drivers with a history of convinced DUI/DWI to get these systems installed

But mandatory breathalyzers in all cars would prevent the initial DUI offense. If they were mandatory in all cars, car makers would make them more convenient, so maybe instead of blowing into a tube, you just have to blow at a sensor in the sunvisor.

The posted speed limit may offer a suggestion of what's reasonable or prudent, but it's very often not the case

But it's a good first step - set the max speed for cars at the posted speed limit, subtract 10% at night, 10% more for rain, 20% more for snow (so for example, a 60mph road would have a max speed of 55mph at night, 50mph during rain, and 40mph during snow)

Note that these are still just "maximums" and the onus would be on drivers to go slower when conditions warrant.

What do you propose here, that doesn't shut down my car when I'm avoiding a road hazard?

Cars are getting more autonomous, so having it drive to the side of the road or to the nearest exist if on a freeway seems like the obvious choice.

As for seatbelts, I think the driver should be free to choose to use them or not, but only as long as they get signoff from their insurance company. Whether or not a driver uses a seatbelt only affects himself, unlike DUI, speeding or wreckless driving.


> it's a good first step

No, it isn't; it's whatever the locality decides is a good setpoint for a revenue source. Most "speed limits" have nothing to do with safety.

Even on interstate highways, where the "revenue source" motivation is the least, many speed limits are still too low given the improvements in vehicle handling and performance since the 55 mph rule was arbitrarily selected in the early 1970s. Of course, that argument assumes that 55 mph was reasonable for early 1970s cars--but if the primary purpose is safety, well, there are still traffic deaths at the current speed limits, so why not lower them further? Why not stop people from driving at all, if nothing short of that will eliminate traffic deaths?

That kind of argument leads down a rabbit hole. I think a much better way of viewing things, as I said in another comment upthread, is in terms of a principal-agent problem. If you are driving yourself (or flying yourself, for that matter), you can make your own judgments about safety (and take the consequences). But if you are flying (or driving) other people for money, then you have to abide by an explicit set of standards because it's no longer just you who takes the consequences if you make an error of judgment. And the more people who are endangered if you make an error, the more stringent the standards should be.


> since the 55 mph rule was arbitrarily selected in the early 1970s

It was selected in 1974 in response to the 1973 oil crisis in the hope of reducing oil consumption. 55 was intended as a speed that delivers best mpg for typical 1970s vehicles.


Seems like your objection would be solved by the proposal.

Towns would no longer have this bullshit revenue source, and might actually set the limits to a reasonable level instead of one intended to make you screw up.


> Seems like your objection would be solved by the proposal.

No, it wouldn't. In fact the proposal would make things worse, because it would give an even greater level of control to governments than they have now. Right now, all they can do is give me a ticket if I'm speeding. Under this proposal, they would be able to remove control of my car from me altogether based on some arbitrary limit. That is thousands of disasters waiting to happen.

> Towns would no longer have this bullshit revenue source, and might actually set the limits to a reasonable level instead of one intended to make you screw up.

You have much more faith than I do in the ability of the government to know better what is reasonable and prudent under a given set of conditions than the person actually driving the car in those conditions. Not to mention their ability to ensure a safe implementation of the actual mechanics of the proposal.


I have a lot more faith in road engineers than individual drivers on determining safe limits. Untrained individuals are pretty crap at this, especially in areas they're unfamiliar with - an out-of-town visitor can hardly know that there's a low-visibility hairpin turn coming up.

Removing the perverse revenue incentives of artificially low limits seems like a good step.


> I have a lot more faith in road engineers than individual drivers on determining safe limits.

The road engineers aren't there. The driver is. Also, the road engineers suffer no consequences if the limits are wrong.

> Untrained individuals are pretty crap at this

Perhaps. If so, I think it's because individuals have no voice in determining the limits, and they know the limits are bogus anyway, so they don't view exceeding them as an error, they just view it as a game they're playing vs. the police. In other words, since there is no reward for exercising better individual judgment, and the penalties are unrelated to the quality of an individual's judgment, there is no incentive for individuals to develop better judgment.

> an out-of-town visitor can hardly know that there's a low-visibility hairpin turn coming up.

That's what warning signs are for. There's no need for it to be a speed limit; just a "caution, hairpin turn" sign is enough.

> Removing the perverse revenue incentives of artificially low limits seems like a good step.

The way to remove those incentives is to stop allowing governments to penalize people who have caused no harm. Sure, put up a sign that says that, according to our best qualified road engineers, the maximum safe speed for this road is x. (For bonus points, make the sign programmable so the posted limit can vary with weather conditions, day vs. night, etc.) But don't allow the government to give me a ticket and make me pay a fine just because I exceed speed x. If I cause an accident and it's found that I exceeded speed x, then the government can penalize me--but penalize me because I caused harm (and, if you like, tack on an extra penalty because I ignored the advisory sign and my error of judgment contributed to the harm).


> The road engineers aren't there.

They were there when the road was designed and built, and they did various calculations based on actual physics, road width, medians, upcoming intersections, etc. to determine a limit. "Eh, this feels like a 75" from an untrained driver with three big accidents on their record shouldn't be of equal value.

edit: Here's the US DOT's flowchart for picking speed limits: https://safety.fhwa.dot.gov/uslimits/documents/appendixk.pdf - they even have a handy web tool for putting in the variables. https://safety.fhwa.dot.gov/uslimits/

> Also, the road engineers suffer no consequences if the limits are wrong.

Real engineers have liability and licensing implications for fucking up.

> If so, I think it's because individuals have no voice in determining the limits...

Nor should they, really, as they're supposed to be based on empirical data about the nature of the road. Your attitude reminds me of this New Yorker comic: https://www.newyorker.com/cartoon/a20630

> That's what warning signs are for. There's no need for it to be a speed limit; just a "caution, hairpin turn" sign is enough.

There are several tight turns near my house. One's a 25 mph one, another is a 15 mph one. Even as a local, I find the numbers a handy reminder; there's absolutely no way a non-local would be able to figure out which turn should be taken at which speed.


> "Eh, this feels like a 75" from an untrained driver with three big accidents on their record shouldn't be of equal value.

Any such driver should have already suffered enough consequences from three big accidents to have changed their behavior. And if they haven't changed their behavior, how is a speeding ticket (which costs a lot less than three big accidents) supposed to?

> Real engineers have liability and licensing implications for fucking up.

For things like bridges and buildings, yes. And for consequences that are easily seen, yes. But what consequences do road engineers suffer if they set the speed limit on a road too low, forcing people to take more time and burn more gas?

> There are several tight turns near my house. One's a 25 mph one, another is a 15 mph one.

Yes, turn warning signs with a safe speed recommendation included. No problem there. But can you get a ticket if you go around the turn at a higher speed than posted, but don't cause any harm? (As I understand it, you can't; the speeds on those yellow warning signs, unlike the ones on the white speed limit signs, are advisory only and you can't be ticketed just for exceeding them. Which is exactly the kind of thing I'm advocating, just for all limits.)


It's hard to tell from a posted limit sign if an engineer was involved. Or, if the engineer was directed to a target limit rather than allowed to find the appropriate limit based only on the road design and such.

Most freeways had a design speed of 75, for vehicles manufactured in the year they were designed.

When it's enforced by cars, it's no longer a revenue source.

No, just an even worse power put in the hands of governments.

Try this thought experiment: suppose we are at some point in the future, when self-driving cars are ubiquitous and are proven to be safer than human-driven cars. Do you want governments to have the ability to force your self-driving car to do something based on some arbitrary setting of a limit, controlled by government bureaucrats? Or do you want to have control over what algorithms your self-driving car runs and what actions it takes if it detects a problem?


> Or do you want to have control over what algorithms your self-driving car runs and what actions it takes if it detects a problem?

I'd rather not be driving in a world where the other drivers get to pick the "run a family off a cliff if it prevents a bumper dint on my vehicle" algorithm.


See my response to Johnny555 just now.

I want everyone's cars to follow the same rules, I don't want individual people modding their cars to obey only the laws they like or to behave differently than everyone else's cars.

> I want everyone's cars to follow the same rules

How would the rules be determined?

> I don't want individual people modding their cars to obey only the laws they like or to behave differently than everyone else's cars.

So basically, you don't trust the judgment of individual people. But you trust the judgment of a government that is composed of individual people?


Sounds like you have a hard-on for the government.

The federal government already regulates motor carriers, including inspections by federal and state regulatory bodies. They understand how to predict risk for vehicle operations, and can scale that assessment up to massive fleets like UPS or USPS or down to an individual vehicle operator.

End of the day, individuals on average are awful at risk assessment and consistent behavior. To borrow another phrase, brakes are what allow cars to drive faster.


How would the rules be determined?

USDOT/FHWA

So basically, you don't trust the judgment of individual people. But you trust the judgment of a government that is composed of individual people?

Yes.


Ok, then I guess we'll just have to disagree.

yeah, I think it's the difference in opinion between those that see cars only as transportation, and those that see driving as an experience itself.

I'm happy to give up control of driving if it means car travel is safer and more convenient, just like I'm happy to get on a train, bus, or airplane and let someone else drive (under high regulation - I don't want my pilot to decide on his own that ATC is being ridiculously conservative and there's plenty of room for him to land between two A380's).


The software author will ultimately be liable for accidents, so they will govern how the car behaves.

Government bureaucrats and corporate bean counters will determine how you get around.


The posted speed limit is some information, but more important are things like size of the lanes, visibility, amount of other traffic, separation from the other direction of traffic, separation from non motorized road users, curves in the road, pavement condition, etc.

If US-101 in Sunnyvale were posted for 45 mph, that wouldn't be a reasonable and prudent limit; it's got lots of separation and wide lanes, and the pavement is ok -- traffic permitting, you can travel at high speeds. If it were posted at 80 mph, it wouldn't be reasonable and prudent either -- the pavement isn't that great (unless it was repaved recently), most cars would be bumping around too much at that speed.

Moreover, I don't know who would purchase a car that promised to strictly follow posted speed limits, when they have alternatives.


Unlike cars, all commercial airliner ops are under instrument flight rules -- meaning positive control by ATC at all times.

Speed and heading are all known. Mandatory spacing and separation minimums are enforced, ATC calls out traffic and any re-routing from the flight plan requires clearance. While in cruise, autopilot use is generally required.

If we had to file a driving plan, get clearance, wait for a departure slot, and use cruise control/autosteer to maintain a set speed and routing every single time, it probably cut down on car accidents as well.


It's harder to implement on cars.

1) People would (and do) cheat at mandatory breathalyzers.

2) It's not speeding that injures people, it's the speed differences that cause issues. Limiting a car to 70mph won't help when it's icy, and they should be driving 40mph.

3) How exactly would a car detect reckless driving? How would it know that it wasn't warranted (ie swerving to avoid a kid)?


> It's not speeding that injures people, it's the speed differences that cause issues.

When two car are at a speed of 70mph, the maximum speed difference between them is 140mph, when they are at 50mph, it's only 100mph. So the speed of a car has a direct impact on the speed difference between cars.


Reckless driving is rarely a single event but a behavior over time. I know a couple of folks that are pretty reckless, doing 35 through neighborhoods, tailgating, abrupt stops. Hell just monitoring brake rotor temperature is probably a pretty good indicator.

A lot of things are hard. Do we just give up because things are hard, or because we don't know how to do them yet?

We have created nuclear power plants, airplanes, giantic ships, giant dams, power grids, 2,700 foot skyscrapers, and so on. I think we can design a computer with sensors to detect unsafe driving conditions, and at least get a minimum level of operational safety.

I don't have all the answers for you. But there probably are answers, and they're probably not all that difficult, considering everything we've accomplished before. Maybe it's worth the tens of thousands of human lives to at least try, even if we don't know exactly how ahead of time.


A lot could be solved by rigourous enforcement and very long driving bans. Many people lack the mental capacity to drive safely and are not actually capable of doing so. They endlessly provide evidence of this by not following procedures and rules.

But this is all very authoritarian. People quite understandably resent this kind of overwheing regulation. They don't see the difference between driving to the shops or walking. Why should the latter attract so many rules?

The hope of technology is that we can help incapable drivers be safe. But they are going to resent the imposition. It is just another kind of authoritarian control. But that is perfectly reasonable when a machine has so much kinetic energy.


That and the ubiquity of cars. If everyone flew their own plane to work every day, it would probably be different....

I flew into buffalo that night about an hour before the crash. I remember thinking when we were landing that these conditions were some of the worst, and I fly a lot into Buffalo.

My brother is a commercial airline pilot who flies world wide and pilot fatigue is real and very common.

Airlines have been making bank for the last few years and all they need to do to prevent pilot fatigue is reduce hours for pilots and increase the number of pilots/flight, especially for international flights.

These simple measures (albeit costly), will overall help reduce commercial flight fatalities across the globe.


My father was an airline pilot for years (American, US Airways, and then Etihad) flying domestically and then internationally. I'd guess that if the airlines wanted to reduce pilot hours the pilot unions would have a lot to say about that.

Are they paid by the hour? I'd assume that pilots would earn an annual salary.

Most pilots do earn an annual salary. I'm not sure what the previous comment meant. Pilots simply have to work minimum number of hours per month, but the issue is that most pilots are now working up to the maximum number of hours therefore there are actually pilot union striking because of overwork!

Based entirely on the headline: doesn’t every crash prevent most future crashes that would have had the same cause?

Yep, I always loved this line from Antifragile:

> “But recall that this chapter is about layering, units, hierarchies, fractal structure, and the difference between the interest of a unit and those of its subunits. So it is often the mistakes of others that benefit the rest of us—and, sadly, not them. We saw that stressors are information, in the right context. For the antifragile, harm from errors should be less than the benefits. We are talking about some, not all, errors, of course; those that do not destroy a system help prevent larger calamities. The engineer and historian of engineering Henry Petroski presents a very elegant point. Had the Titanic not had that famous accident, as fatal as it was, we would have kept building larger and larger ocean liners and the next disaster would have been even more tragic. So the people who perished were sacrificed for the greater good; they unarguably saved more lives than were lost. The story of the Titanic illustrates the difference between gains for the system and harm to some of its individual parts. The same can be said of the debacle of Fukushima: one can safely say that it made us aware of the problem with nuclear reactors (and small probabilities) and prevented larger catastrophes. (Note that the errors of naive stress testing and reliance on risk models were quite obvious at the time; as with the economic crisis, nobody wanted to listen.)”


The engineer and historian of engineering Henry Petroski presents a very elegant point. Had the Titanic not had that famous accident, as fatal as it was, we would have kept building larger and larger ocean liners and the next disaster would have been even more tragic. So the people who perished were sacrificed for the greater good; they unarguably saved more lives than were lost. The story of the Titanic illustrates the difference between gains for the system and harm to some of its individual parts. The same can be said of the debacle of Fukushima: one can safely say that it made us aware of the problem with nuclear reactors (and small probabilities) and prevented larger catastrophes.

Another interesting argument along the same lines is that if we hadn't bombed Hiroshima and Nagasaki, there would have been no reason for Truman to stop MacArthur from using bombs a hundred times worse in the Korean conflict.

It's a sobering thought regardless of one's opinion on the atomic bombings in Japan. The lesson was going to be learned one way or the other, and arguably humanity got off easy.


I have never heard that one -- but it makes sense.

Only if they’re actually investigated and the lessons learned are translated into real changes. Fortunately, commercial aviation is really good about that.

In this case, however, the problems were already recognized. What this crash did was to provide the impetus to stop arguing and finally do something effective about them.

In "Fate is the Hunter", Gann lamented that the airlines had to be forced by regulation to adopt as straightforward a safety measure as the rotating beacon (flashing red light), a change that pilots of the day had to lobby for.


Earnest Gann's Fate is the Hunter is a must-read for aviation buffs which flies under the radar in discussions on the great books on flying.

I worked in the business and I'd never heard of it until a hardware engineer (with WDC) told me I had to read it. He was right.

It's a nail-biting, seat-of-the-pants diary of the author's career during the early days of commercial aviation. So good.

Highly Recommended.

(me: aircraft dispatcher in my previous life)


The article opens by explaining that the cause of this crash wasn't ever determined- rather, it's the changes in safety regulations that resulted.

This reminds me of the book "Black Box Thinking" that basically says it's more important to learn from mistakes so we don't repeat them than to assign blame. The commercial aviation industry does this better than any other industry, even medicine.

Yes. Blame is poisonous. Humans hate blame, they'll deflect blame onto the innocent, they'll lie and cheat, to avoid blame. The correct focus is prevention of future harm. Not "who is a bad person?" but "what will we do differently next time to avoid this outcome?".

In that frame suddenly the driver stops saying they weren't too drunk to drive and says next time they'll take a cab. The policeman stops saying the suspect "wasn't complying" and agrees that they need training on how to de-escalate situations.

Medicine does have tools for this, the M&M conference (medics discussing why somebody died or had a bad outcome and how to do better next time) is under-used.

Humans are fallible, blaming a specific human makes us feel better but does not prevent the same thing happening again.


I’m completely with you on this and have been implementing this in practice. Sometimes, however, I still meet resistance. Do you have any source/literature that underpins this method?

A nice book explaining some techniques used in aviation and how to transfer them eg to medicine is Atul Gawande's The Checklist Manifesto.

... Black Box Thinking, by Matthew Syed.

Blame isn't the only problem. Humans also hate change. Eliminating blame reduces the resistance to implementing better policies but does not completely resolve it.

Aviation accident investigations always assign blame when it can be determined. In this particular case [1]:

> The National Transportation Safety Board determines that the probable cause of this accident was the captain’s inappropriate response to the activation of the stick shaker, which led to an aerodynamic stall from which the airplane did not recover. Contributing to the accident were (1) the flight crew’s failure to monitor airspeed in relation to the rising position of the lowspeed cue, (2) the flight crew’s failure to adhere to sterile cockpit procedures, (3) the captain’s failure to effectively manage the flight, and (4) Colgan Air’s inadequate procedures for airspeed selection and management during approaches in icing conditions.

[1] https://www.ntsb.gov/investigations/AccidentReports/Reports/...


The primary purpose of the NTSB is to prevent future accidents. Any "blame" they assign is in furtherance of that purpose, not punishment. So critical is this to the NTSB's mission that any conclusion the agency draws as to the cause of an accident cannot be submitted as evidence of civil liability[1] in court and every NTSB report includes a footnote to that effect. You'll find it on page 4 of the document you've linked.

[1] The report would be hearsay in a criminal case and admissible only if the government demonstrates an exception to the rule against hearsay evidence.


Woah, slow there.

The NTSB tries to determine "Probably cause". That's a far cry from assigning blame. In particular, on every NTSB report there's a big box specifically saying (my emphasis):

> The NTSB does not assign fault or blame for an accident or incident; rather, as specified by NTSB regulation, “accident/incident investigations are fact-finding proceedings with no formal issues and no adverse parties ... and are not conducted for the purpose of determining the rights or liabilities of any person.” 49 C.F.R. § 831.4. Assignment of fault or legal liability is not relevant to the NTSB’s statutory mission to improve transportation safety by investigating accidents and incidents and issuing safety recommendations.

Furthermore, to "ensure that Safety Board investigations focus only on improving transportation safety, the Board's analysis of factual information and its determination of probable cause cannot be entered as evidence in a court of law."


from the article: "By any measure, the safety record since the crash is unprecedented. Out of more than 90 million flight departures on U.S. airlines carrying billions of passengers since then, there has been just the single death: A woman died on April 17 on a Southwest Airlines Co. flight near Philadelphia when an engine failed, sending debris into a window next to where she was sitting."

Like our chief pilot used to say: "All regulations are written by dead people".

"Written in blood" is the way I heard it, but yes. Agreed.

He would often add: "So do not rush to be the reason for the next one ...".

Meanwhile in Canada regulations are still appalling and some of the worst in the world: https://www.cbc.ca/news/politics/garneau-pilot-safety-airlin...

How catastrophic must a software disaster be to receive the benefits of such safety regulation?

There are software glitches responsible for the deaths of on the order of 100 people, which has resulted in industry-localized development and deployment standards. When we see tragedies where 1000 people die because of a software glitch, or more likely, a systems or security failure, then I think we'll see legislators start coming for the software industry in general. There are already critical systems development and systems requirements; likely legislators will expand and formalize those kinds of programs.

All we need to do is have a public tragedy that can be pinned on "software".


This is real engineering. And this is what I'm most excited about with autonomous cars- black boxes allow the NTSB to get involved and work accidents out of the process.

I think something we haven’t learned from the Colgan disaster is how regional airlines are configuring these smaller jet and turboprop planes to fit more passengers, in ways they were not designed. That is leading the way towards profit over safety.

I wish one could write the same title about software...



Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: