Get in touch with technology with text Stuff from hast dot com.

Heyl and welcome to text Stuff.

I'm Jonathan Strickland, and today I thought I would address some listener mail.

I haven't done that in a while.

And by address, I mean read it and answer it, as opposed to spinning out an address.

You know what I meant at any rate to We're going to be looking at several questions and comments, including a correction, and also a little bit later, I'm going to bring in a special mystery guest to talk about one of the specific pieces that were sent in to me.

So let's get started.

The first message says, hello, longtime listener, first time emailer.

I wanted to slip in a quick correction in regards to the Aircraft Carriers episode.

In the episode, you said that the US aircraft carrier, the Hornet, was sunk.

It's actually not.

It's anchored in Oakland, California, across the Bay from San Francisco, and it's open as a museum to teach visitors about how aircraft carriers work.

The flight deck, maintenance deck, and some lower decks are open to explore with or without a tour.

The island and engine room are open for tours.

There's even an Apollo capsule on board, as the Hornet was the recovery ship for two Apollo missions.

If you're ever in or around Oakland, check it out.

Regards Alan, Alan, thanks for sending in that correction.

That's really awesome.

The Hornet was one of the stories that I thought was particularly cool, and uh, and on it was an honest mistake on my part.

I had so many notes about so many different aircraft carriers, and I should have been more careful making sure that all of those were accurate.

And uh, Scott is not his fault at all.

It's all on me.

I shall admonish myself repeatedly.

Next, Dan from Twitter asked, did I buy an Apple Uch?

No?

Okay, well we're ready to move on.

Well, I guess I can address this a little further.

Uh, I did not buy an Apple watch.

I don't have an iPhone, so an Apple watch would be useless to me, and I don't want to invest in an iPhone just to get an Apple watch.

That would end up being an incredibly expensive endeavor.

I mean, you're talking around bucks for the watch, and that's you know, before you really start looking at really nice bands.

Or anything like that.

Then you've got another couple of hundred bucks for the iPhone, plus the monthly fee for it and everything, and all of my stuff is currently in the Android operating system universe.

I use a Nexus six as my smartphone.

That's not a plug for Android, it's not a plug for the Nexus line of phones.

I like it, and I'm used to it, and all of my stuff is in that ecosystem, so it would be very arduous to switch to a different one at this point.

I think the Apple Watches an interesting product.

I'm curious how well it will do in the long run.

As I record this, there are plenty of stories about the Apple Watch getting some recalls because of a faulty vibrating motor that's in some of them, and that the production may have slowed down quite a bit, which means it's going to take even longer for people to get their pre ordered watches.

Um it's it's a bit of an issue for Apple, And honestly, the wearables field is still one of those where a lot of people are expressing a lot of skepticism about, you know what, what's actually the utility of a smart watch?

Is there a need for it?

Um, I own a Pebble, which I would say is a smart ish watch.

It is good for giving me notifications that are coming through my phone, which is actually really useful for me.

I don't always notice when my phone vibrates in my pocket, even though the next of six is as large as some countries.

I don't always notice when it's vibrating.

However, the motor that's inside my watch, when it vibrates on my wrist, I absolutely noticed that.

And uh, it gives me the notifications of who is calling or if I've received a text message or some other notification.

And I love that.

That to me is perfect, And uh, you know, it's it's not super high tech as far as smart watches go.

It's not like there's a whole bunch of other usefulness with it, apart from some of the cool apps you can have, and you can do some controls with your smartphone, like advancing through tracks in an audio list, that kind of thing.

That stuff is interesting and neat, but it doesn't have a whole lot of utility beyond that.

The Apple Watch obviously is supposed to be more robust, and I hope it does well because I want to see this kind of technology.

This, this whole branch of technology continue and advance and evolve.

And my worry is that if the Apple Watch ultimately does not do well, because it's such a high profile product, it could sink the entire branch of technology or set it back by a couple of years.

It kind of reminds me of the days when virtual reality things were coming out and people were first getting a chance to actually use them, and they became disappointed, disenchanted with the technology that they thought was going to be amazing and turned out not to be as amazing as they had hoped, and it's set the entire field back by at least five years because no one wanted to fund something that people no longer had any interest or faith in.

So my hope is that the Apple Watch does well.

I do not plan to get one again, because I don't have an iPhone.

I may end up picking up a different smart watch.

There were a few that I supported on various crowdsourcing campaigns.

I'm still waiting on all three of the one that I supported, one of which is not even a smart watch.

It's it's a just a cool concept watch.

And so I don't think i'd be eager to jump on the smart watch train even if I did have an iPhone, simply because I've already spent money on watches that I don't have.

Um and that's another discussion for another day.

The entire idea of crowdsourcing, you know, crowdfunding a technology project.

Is that a good idea?

What are the dangers?

I should probably do a full episode about that.

I know I've talked about once or twice in previous episodes, but we should probably go into an in depth discussion about what is What should your expectations be if you participate in that sort of thing.

So there's your long answer to your simple question.

Uh no, And because all right, we'll move on to the next one.

Luke on Twitter asks how soon I expect we'll see stores like Kinko's get three D printers for customers to us.

I think that's actually gonna be probably at least a couple of years further down the line from the thing I think three D pryers are really going to be useful for is gonna be on demand fabrication for small businesses, meaning that instead of having a giant furniture store with an enormous warehouse, like Ikea, where you've got, you know, tons of inventory, and you have no way of knowing how how well one is going to sell versus other versions of the same sort of stuff, Like how how well is this one cabinet gonna sell compared to this other cabinet, both of which have Norwegian or Danish or Swedish or Swiss or some Scandinavian name that has far more oom louds than is remotely necessary.

Um, how well is that going to sell versus any other one?

You have to keep enough in stock so that you meet whatever demands you expect there to be, but there's no way of knowing from one day to the next how popular an item is going to be, even if you put one on sale and one's not on sale.

Three D printers, however, can let you get around that.

You don't have to have a huge warehouse to store all of your inventory.

You can have essentially a storage space for the raw material for your printer, and then you create the items on demand when you get in order.

So if someone wants a particular cabinet, you print out the parts and then sell the kit to the customer.

You don't have to keep eight different types of cabinets and storage.

You just keep the digital plans and you print them as needed.

I think that's gonna be a very popular business model once we get to a point where these printers that are are ready for commercial use like that fall into the range that a small business can can actually make that a workable business model.

It's pretty expensive if you want a really high quality three D printer, especially one that can produced on an industrial level, even a small industrial level, asked for things like Kinko's.

I think that's gonna take more time.

Uh.

And I think there's gonna be some barriers as well, because obviously, you if you are a business like Kinkos, you want to be able to have at least some control over what is getting sent to that three D printer.

Uh.

If it's something that is illegal, for example, or the legality is questionable, for example, a three D printed gun.

That's that could be a concern, especially if there were any crime associated with that weapon down the line, and it gets brought in that your establishment was the one that allowed it to be printed.

There could be some legal complications that.

I bet a lot of the big companies don't want to get bogged down in and it may be that eventually things work out where the accountability lies with other other players and not the company that act is housing the three D printer, in which case you know it may be open season.

No pun intended with the gun thing to allow everyone to send whatever they want to the printer, and as long as they're paying the rate for however much plastic is going through the extruder, they're fine.

So I think we'll see the fabrication on demand model first.

Of course, three D printers are already being used in research and development as well as in manufacturing for other reasons.

But but the as far as customers getting access to three D printers, I think it will be fabrication on demand, where the company will be the one responsible for sending the specific plan to the specific printer, and the customer only says okay, I want this item, so they don't get to say exactly what gets printed.

And then maybe a couple of years later we'll see models where people can send their their digital file to a printer directly without having to just choose from a catalog of files.

Um In the meantime, you know there will be plenty of people who will just go out and buy their own three D printer.

Depending upon how much money they spend, the stuff they print may or may not look really great, or it may look kind of clunky.

I've seen some three D printers where the the fidelity, the resolution is incredible and you you can feel the edges of each layer, but you can't really see them easily.

And then I've seen other ones where it's very apparent where each layer is kind of overhanging the next layer by you know, a millimeter or whatever, and it's very noticeable.

Uh So we'll see.

Maybe I'm wrong about that, but I suspect not.

All right, we'll go to the next question.

Nathaniel from Twitter asked how track pads register input.

How does a track pad know where your finger is?

There are two main ways that track pads are manufactured.

There are resistive track pads and capacitive track pads or touch pads.

And I know I've talked about this on the show before, but it you know, I'll go ahead and explain how these work because I think it is interesting.

So the resist of touch pads or track pads are the easiest to understand.

Um they're also not as not necessarily as resilient as capacity, but we'll get there.

So resistive touch pads use a couple of different films with kind of a uh electrodes.

Well, one film will have an electrode on it, uh, the next film like a sandwich filling that one would be uh somewhat resistive, and then you would have a bottom plate that would also have an electrode on it.

So you've got electrodes separated by a resistive film.

When you press down on the surface of the top layer, it presses that electrode closer to the other electrode on the bottom and you end up creating essentially a circuit.

There is a connection there.

And if you think about the two films as as sort of wires, right like, you have a series of horizontal wires along one film and vertical wires along the other electrode.

Then you've got a grid.

So imagine you've got this grid there.

When you put pressure on the touch pad, it acts like these It pushes these these two wires together at some point on that grid, and by locating the the area of connection along that grid, a computer knows, oh, there has been a touch registered at this point on the touch pad.

So um, you know, it's almost like a spider web, right, So you've got this this grid of connections and when you press down, it sends the signal, and by measuring that signal, the computer can tell where along the touchpad you are touching.

Capacity is a little different, so instead of instead of it being the series of wires and the physical connection is what indicates where there's a touch, Capacity creates an electric field when your finger encounters that electric field.

Really, any capacitive or conductive surface I should say, not capacity, but a conductive surface, it alters that electric field, and sensors inside the device can detect where that that point of contact is based upon the change of the electric field itself.

Now, this is why if you use gloves that don't have a conductive surface on them, so they don't have like conductive thread woven into them or anything, it's very difficult to get a capacitive touchscreen to work because you're you don't have a conductive urface interfering with that electric field.

Generally speaking, gloves don't tend to be conductive unless again, you have that conductive thread woven into them.

It's also why I remember a story from several years ago that people.

I want to say it was in Japan, we're starting to use hot dogs as a stylus for their smartphones when it would get cold, so they would have gloves and use a hot dog, which would be conductive if they would use that to make selections and commands on their phones.

Um.

So with the resistive approach, it's all dependent upon pressure you have to press, and if you're not pressing hard enough to make the contact with those wires, then it won't be registered.

With the capacity of it's not dependent upon pressure.

You don't need to push harder to make a stronger electric field, although it may by pressing harder indicate a larger surface area coming into contact with the screen, which you know, you could design a program that would react differently depending upon how much pressure was applied, simply by many sharing how much surface areas being affected.

Ah.

The problem with the resilience that was talking about earlier is that, at least with earlier resistive touch pads, the more you would move around, the more you would use it, the more it would start to wear down, and over time your touch pad would become less responsive because it was starting to just wear away from use.

Whereas capacitive ones.

You know, you don't have to have this hard contact to make sure that it's registering a touch, and they could last longer for a while.

Capacity of touch pads were much more expensive than resistive, so resistive we're very popular, particularly in industrial uses where you might have like a tablet that has a resistive screen and it's meant for heavy duty use.

There's no point in making it delicate because it was going to be in a tough environment anyway, and resistive was really popular.

These days, the difference in price is not as great, so we mostly see capacity of screens but reicularly in things like smartphones and tablets.

So I hope that answers that question.

All right, let's move on to the next email.

This one comes from Greg and says, Hey, Jonathan, I enjoy your show Tech Stuff and forward Thinking a lot.

Y'all are some real gangster's thanks.

I don't really think of myself as gangster.

That's that's a new one for me.

So you did an episode on TV Connectors recently and mentioned impedance gain and attenuation and then said that could be its own podcast.

So I'd love to hear a podcast about those topics.

Maybe also explain decibels as well.

All right, Greg, here goes.

This is gonna be an attempt on my part to explain these concepts.

First of all, I gotta come clean, because you know, have you ever had that experience where you encounter a word and you have to say the word out loud, and then you realize as you're about to say the word out loud that you have never really heard that word spoken before, or if you had, you hadn't really paid attention, and so you just have to give a wild guess as to how it's pronounced, and then you guess wrong.

Because that is exactly what happened to me the last time I talked about impedance.

I believe I meant I pronounced it impedence, which, while hilarious, is not correct.

So if any of my listeners out there tried to follow my lead and found themselves in socially awkward situations because of their pronunciation, I humbly apologize, Um, but no, that was definitely one of those cases where in my brain I'm like, oh, it has to be spelled like are pronounced like this because the way it's spelled, and I was absolutely wrong.

And in fact, if I had thought about it for a second, I would have realized how wrong I was, because it's about impeding.

It's impedence because it impedes the flow of electricity.

But I'm getting ahead of myself.

So when when we talk about impedance, we need to first talk about voltage.

And for those of you who have taken physics and you know all about circuitry and voltage and current uh and resistance, I apologize, But in order to explain impedance, I kind of have to lay some groundwork, So just stick with me here.

Voltage is the difference in electrical energy charge of electrical potential energy between two points.

Uh.

And that sounds weird, but another way to look at is that voltage is how we describe an electric field.

We measure electric fields and volts over a distance, so you have to have a difference in voltage in order for current to flow.

Another way of thinking that is you need to have an excess of electrons on one side of your circuit and a place for electrons to go on the other side of your circuit, and that is a difference in voltage.

You have a positive charge on one side and a negative charge on the other.

And that allows the electrons to move from negative to positive.

We won't go into the difference between the flow of electrons and the flow of current here.

I know I've talked about that in previous episodes of tech Stuff, and it's not really relevant in this case.

You just need to know that voltage is this difference in electric charge and electric potential energy between two points across some distance.

When you apply voltage to a circuit, that's when you induce electricity to flow through that circuit.

Now, the circuit is made of some sort of conductive material.

Now, conductive material doesn't allow electricity to flow through effortlessly.

There's an opposition to the flow of electricity, and it's called resistance.

And resistance depends upon a couple of different factors um well several.

Actually, it depends upon the material.

It depends upon how much the material is there.

In other words, if you have a very small copper wire that actually has a much higher resistance than say, very thick copper cable.

It also depends upon temperature.

You know, super cooling semiconductors can reduce the resistance to near zero or effectively zero, and you get you know, free and clear electricity.

Flowing through without any resistance, and you know, otherwise what happens is the opposition to that electrical flow will end up having you uh, you know, the component will actually start to generate heat.

And this is where we talk about how efficient and electronic devices.

If it's generating a lot of heat, you're losing a lot of energy to heat.

That electricity is being converted into heat energy and not doing whatever it was you were hoping for it to do, unless, of course, it was a heating component, in which case it's doing exactly what you wanted it to do.

So that's resistance.

Now.

In direct current circuits, resistance suffices as the term we use for this phenomenon.

But with alternating current a C applications, these are different because the current alternates.

It flows in one direction the circuit and then flows in the opposite direction.

Uh.

And it does this in cycles different frequencies that will usually it's in a single frequency, but the frequencies can vary depending upon the technology we're talking about.

So in this case we don't use the term resistance.

We actually use a different term, and that is impedance because uh, impedance relies on two factors, not just the magnitude, which is pretty much what resistance is considered concerned with, but the phase of the of the system as well.

UH So impedance impedes the flow of electricity, but it consists of reactants as well as resistance, and reactance is the extent to which a circuit stores and releases energy as the current and voltage fluctuate with each alternating current cycle.

So we describe reactance in terms terms of ohms, and it really only applies to a C circuits.

Reactance allows us to use circuit components such as capacitors and inductors.

Capacitors store energy in an electric field and inductor's store energy in a magnetic field, and both can operate as resistors and they don't generate heat the same way that resistors do, so they're very useful in a C circuitry.

So that's impedance.

Uh it's an important factor whenever you're designing any kind of electronic component that has that's working on a C.

Next, we have gain, which is the next thing that that Greg asked us about.

This is also known as the amplification factor, and it gives you some clue as to what's going on with gain.

It's all about amplifying a signal to boost its strength from input to output.

So you've got an incoming signal, then you use a means of amplification to increase the strength of that signal, so it's the same frequency but greater amplitude, and then you put it through the output.

This is used in all sorts of applications.

The easiest one to imagine happens to be sound, where a microphone will convert acoustic waves into electrical impulses that then will go to a an amplifier before going to a speaker in order to strengthen the signals so that you can get an audible acoustic wave on the other end once you get through the stereo.

But amplifiers are used for all sorts of technology, not just acoustic technology.

At any rate, we express gain in terms of decibels, and this is where things get a little confusing for people often because I think a lot of people are familiar with decibels as a means of measuring the volume of sound.

But it's be It's not just the volume of sound, it's the strength of signals.

And the decibel is a logarithmic unit, which makes it a little tricky to understand, particularly you're mostly familiar with them.

In terms of that sound magnitude gains more about um in electronics anyway, it's about electric signal strength, not about sound volume.

Uh.

And decibels are really a way of describing two different levels of signal strength, you you know, instead of just saying this one stronger than that one, it's a way of giving a unit to it and saying this is how much stronger the signal is going out than it was coming in.

You know, we were talking about amplifying, So the incoming signal is at a certain strength.

How do you describe the outgoing strength compared to the incoming strength.

Let's talk about sound first though.

So when we're considering sound, near total silence is zero decibels, and it acts as our starting point.

So this isn't completely silent.

It's not like what you would experience if you were somehow able to live out in space where there's no sound.

Uh.

This it's a ventially the strength of the smallest audible sound.

Now, a sound that's ten times louder than that than zero decibels would be at ten decibels, but a sound one hundred times more powerful than zero isn't one hundred decibels, it's twenty decibels.

And a sound one thousand times stronger than zero decibles isn't one thousand dB or decibels, it's thirty decibels.

Now, normal conversation is in the range of sixty decibles.

A car horn is around a hundred ten decibles, a jet engine is around one twenty decibles.

A gunshot right at the sources about a hundred forty decibels.

So in this case, we're talking about the difference in intensity on a logarithmic scale from being practically silent now amplic amplification factors are a little different because you can actually have negative decibels and sound.

That doesn't make any sense.

It's not like you could be quieter than silent or get somehow more quiet than no noise at all.

That's not what negative decibels means.

Instead, what it means is that you have an incoming signal and then you step down the power of that signal before you put it through output, So you can have negative decibels the incoming signal is stronger than the outgoing signal.

Is This is also really important for um sound engineers as well.

So UH, that is an amplification that is attenuation, but more on that in a second.

So with gain, we're talking about boosting signal power and not stepping it down.

Game can boost a signal's power a few decipls two around thirty deciples and strength depending upon the application, and you can also cascade amplification to boost it further.

But there is a problem in that amplification can introduce noise into the signal, and each time you amplify that signal again, it's going to also amplify the noise, So you can't continuously amplify a signal.

Eventually you're going to have so much noise that there won't be any signal anymore.

It ends up creating distortion, so you have a limitation of practical limitation on how much you can amplify a signal.

Attenuation going back to that is the reduction of a signal strength.

So this can be any signal, doesn't have to be an electric signal.

UH.

Sunglasses attenuate the light that comes from the sun that hits your eyeballs, for example.

UH.

This is also called loss.

Often in the UH, the the engineering electrical engineering BIZ and signals attenuate over distance.

It's just something that happens.

So radio waves will attenuate over a distance, UM light will attenuate over distance across things like fiber optic cables.

Electricity will attenuate over great distances.

And it's one of those problems that engineers had to get around using various means, one of the big ones being repeaters.

So let's say you've got a cable and you've got a signal going through the cable.

It's going to attenuate as it travels down this cable.

You may have to have a repeater that would take in the incoming signal and generate a new one to continue down the line to keep on going toward its destination.

And you might have multiple repeaters across an entire network in order to get signals to go from point A to point B.

So attenuation is just a fact of physics.

You know, you can't you can't engineer your way around it entirely.

You can find better and better ways of preserving signals so that they attenuate more slowly.

And in fact, if you look at uh the attenuation in cables, the less attenuation per distance you have, the more efficient that cable is there's some cable manufacturers that really rely on this as a way of saying, hey, our cables are way better than our competitor is because we have very little loss.

We have engineered our cables so that they are superior in most cases, at least for consumers.

I would argue that the difference is negligible.

The signal loss is so small and the error correction and most electronics is so good that it would be difficult to notice as a consumer, if not impossible to notice.

Um However, depending upon what you're using it for, if it's something super high end and you're having to have truly very long cables, attenuation can be a problem.

This is particularly an issue with like quantum computing.

In fact, with quantum computing right now, we expect that there is a hard limit to how far away you could have nodes be from one another in order to communicate effectively.

I believe I don't have this figure in front of me, but I believe that the theoretical upper limit is somewhere around five kilometers, and beyond that you would not be able to have a quantum computer network be able to communicate with a node to node.

So that's an issue um.

It's one of the limiting factors.

It maybe that that means we'll have quantum computing networks, but never a quantum Internet, not a true one anyway, that's a possibility.

Anyway.

I hope that ends up explaining the concepts of impedance, gain and attenuation.

Uh.

It's pretty interesting stuff and uh, I feel like I could go on a lot more about that, probably bring in a couple of engineers and really talk about it.

But I hope that serves as a good introduction into the concepts.

Let's take another quick break to thank our sponsor.

You know, mailing and shipping can really be a big pain.

If you go to the post office, you have to stand in line, and if you don't go to the post office, then your only other choice seems to be a postage meter, which is really expensive.

But in fact, there is another way, a better way, stamps dot com.

With stamps dot com, you can buy and print official United States postage for any letter or package using your own computer and your own printer.

And I've used this.

I've used the there's a scale you can get and it's fantastic.

It plugs in via USB to your laptop.

It makes it figuring out postage incredibly simple.

You just send it to the printer you're ready to go.

Incredibly useful if you have a Kickstarter campaign and you're going to be mailing out a lot of rewards you have a small business.

I can't recommend it highly enough and in fact, right now use the promo code text stuff for this special offer.

There is a no risk trial and you will get a hundred and ten dollars and bonuses that includes fifty five dollars free postage.

Go to stamps dot com, click on the microphone at the top of the homepage and enter text stuff that stamps dot com and enter text stuff all right.

Our next piece of listener mail comes from listener Keegan McDougall, who had a really cool question.

Uh.

And Kegan was asking about a philosophical question and ethical question that is called the trolley problem and wanted to know how it relates to autonomous cars.

And I thought, well, this is this is a point where I could kind of give my own answer and that would be fine, but I think it would really benefit from an actual discussion, which is why I have invited my super secret guest star, whom I alluded to earlier in the show, and it turns out to be Ben bowling of car stuff and stuff they don't want you to know.

I think both of those shows relate to the topic we're going to talk about.

Yeah, hey, thanks for having me on.

And I gotta say I love the Listener Male episodes.

So this is a really great question that ke Getting has asked.

Yeah, I think I think first obviously we should probably talk about what the trolley problem is, what the trolley question is, and this is really kind of a cognitive science slash um ethical slash moral quandary kind of question, and it's a hypothetical situation, thankfully not not a practical exam So the the hypothetical question that the most basic version of this is that there is a track and a trolley is out of control running down this track, and the there is a person on the track, and if the trolley are actually there's five people on on the track, and if the trolley continues it will collide with and presumably kill all five of these people.

You are next to a switch that will switch the track to a separate run of track, and there's one person on that separate one.

So if you pull the switch.

You will save the five people from certain death, but you will condemn the person who's on the separate track to death.

And the question is do you pull the trick the lever or not do you actually make that move?

Uh?

And it's not like there's a a right or wrong answer to this.

There there are answers that more people are that freque will frequently give um And most people would say that to save one to you know, to sacrifice one person, to say five people makes the most sense.

We're doing the vulcan response, right, Right, the needs of the many outweigh the needs of the few or the one.

Uh.

Yeah, also known as the utilitarian response.

Right.

That's a very common way of looking at it.

And then there are variations on this this question, like what if instead of pulling a lever, you are on the trolley and the trolley is out of control and it's heading towards five people on the track, and you can stop the trolley if you push a fat person who's on the trolley with you out in front.

The trolley will hit the fat person and come to a stop before hitting anyone else.

That means you actively have to kill a person.

It's not so much you pull a switch and someone else dies instead of five people, you are actively killing them.

And now we start seeing people say I don't know that I could do that.

Taking no action and allowing people to die feels differently than taking an active role and killing someone to prevent five other people from dying.

Yeah, And other people would argue, well, you're still committing a murder by pulling the switch, just a more comfortable thing, sort of similar to uh, sort of similar to buying prepared food versus going into the forest and killing your own thing, right, or or if you want to think of a in a in a kind of a military way, it could be the general calling for the order of the big push as opposed to someone who's actually on the front lines who would make that same call.

Yeah, that's a really good point, which that actually comes up in a lot of war crime trial.

Oh yeah, I mean, well, if you look at the history of World War One, that was the history of World War One, right, there were lots of criticisms leveled against generals who were who seemed far too eager to commit huge numbers of soldiers to potential death and destruction for what we're often seen as arbitrary gains.

But getting back to the trolley problem and the question Kegan had.

The reason this relates back to autonomous cars is you could imagine scenarios in which an autonomous car would have to make a split decision, and how do you program in that decision.

So here's an example i'm I would give.

Let's say that you are in your autonomous car.

The cars driving for you.

You are not in control.

The car is going down a street and there are pedestrians all along the sidewalk to your right.

All right, an oncoming car under manual control uh loses control.

So the driver loses control of that vehicle and it begins to swerve into your lane.

Your car is unable to move off to the left because of the trajectory of the incoming car.

So what does your car do.

Your car, having traffic behind you, can't back up, So your options are essentially to hit the brakes and you're going to have a collision with an oncoming car, which could potentially put you the passenger at risk, or the car could swerve to the right avoid a collision but potentially collide with pedestrians.

In either case, you're talking about people potentially being hurt or killed.

How do you, as a programmer who's designing the autonomous car are make that call?

And is it right for an engineer to make that call knowing that the person who is inside the car might have had a different reaction.

Yeah, these these are great questions and they tie into some topical things that are going to become an increasingly big issue for car owners and car manufacturers.

As abstract and philosophical as this might sound, in some degree, this is a real life problem.

So a brief explanation here.

One one of the things that we should mention when we get to real world applications is if there is a real person driving and trolley.

Unfortunately, we do have to allow for uh, their preformed opinions of the kind of people they're saving to come into play.

So five people that they consider unsavory or you know, five elderly people versus one child.

Right.

Usually the way we present this problem is that you have no information about the potential victims, because otherwise, I mean there are variations like there's one there's one where you're on the trolley, there are five people on the tracks, and the person who's responsible for the trolley being out of control is also on the trolley.

You can push that person out, and most people would say yes, because this person has committed the act that has put everyone in danger in the first place, and it is ethically all right to have that person be held responsible by way of killing them to prevent the deaths of their otherwise potential victims.

Right, Yeah, and I and I think it's correct for at least the hypothetical nature of it to keep to keep the emphasis on the problem, which is the problematic decision, that dilemma.

But when we talk about programming UM programming, we could potentially be a death machine.

You know what we what we find quickly is that we might run into um problems of profiling through code.

Could we teach uh?

Could we teach a computer that level of fidelity?

And could we teach it to be something that people agree with directly?

Furthermore, I've got to say, Man, I was thinking about this since we were talking about it briefly off air.

The one of the first answers that a lot of people go to with an AI problem is, well, the three laws of robotics, right are perfect, but they don't work in this occasion because they're asking the robot quote unquote to uh differentiate between values of human life, which the first law doesn't allow for it.

Right.

It's a it's a no win scenario for the robot, right because no matter what choice is made, there's going to be some form of injury as a result.

So there's no way that the robot can make a decision that is going to result and no one getting hurt.

That's just not even an option.

It's it's you know, you've got the traffic behind you, so you can't magically say, oh, well just PLoP it into reverse so that way you avoid the collision.

Uh, And you can't you know, you can't say, oh, we'll find a gap in the people like there's going they're going to be situations where this hypothetical question can and will arise.

Now, you could argue that when we get closer and closer to more autonomous cars being on the road, the likelihood of the situation decreases dramatically, um particularly if you have vehicles that are capable of communicating with one another, because then they're going to know each other's positions well before you're in range of each other.

Then you just have to worry about the things that are not part of the system, like pedestrians crossing the street, jay walker's dogs, you know, that sort of stuff, which could still be a factor.

Also, we keep in mind that the autonomous cars are able to react in a fraction of the time that humans are, and they're able to detect everything that's going on.

See, we humans, we're limited by what we can focus on at any given time, So even our field of view is a limitation, right, so a car doesn't have that.

A car can have a three and sixty degree area of awareness and a fraction of the amount of time it takes a human to react to be able to make these decisions.

That being said, you still have this problem of there are scenarios that are possible where ultimately someone has to figure out how do they program this in do they have do they build it in to the cars programming where if it detects pedestrians, those have the top priority of safety because they are the ones who are least likely to be able to survive and impact.

And therefore, let's let's uh, you know, you'd never want to have a crash, but if the choices between crash or plowing into pedestrians, you'd pick the crash.

That's what I would suspect, because I would imagine also that anyone making these cars would test to make sure that they withstood the same crash test parameters that other vehicles do.

So in that case, I would say that the autonomous car is more likely to be designed.

And I think that's from a manufacturer's point of view, two to withstand the crash.

That not that not that that's necessarily always going to be the quote unquote right choice from an ethics point of view, But if I'm a car manufacturer, I definitely would rather have the message be in this very unlikely scenario, uh, the car would prepare for impact essentially rather than run a muck on the sidewalk exactly, And that that's the primary concern as um.

As cold as it may sound, the truth of the matter is that if you are a car manufacturer, part of what you're selling to car buyers is a sense of safety.

So I I think that the priorities would be a little bit different.

The number one priority probably be the passengers within the vehicle.

To the point that with this preparation for impact, which I love what you're saying, than than it may be a case where the car simply stops, uh like proactively deploys its air bags.

Uh and perhaps I don't know, maybe has some brake system that braces for impact.

It might even orient itself so that the most the crumpled zones are at their the best angle for preparing for an impact.

Which, you know, it sounds kind of crazy.

You're thinking, Oh, it's a car that would actually maneuver itself so it would suffer more damage.

But it may be that that's the ideal way to deal with this kind of situation, to make sure that everyone involved has the best chance of uh of suffering the least amount of injury.

Yes, and this brings us to something that we probably don't have time for today, but it's an interesting question.

I'd love to hear um your thoughts on this, two, Keegan.

Uh.

If we have that kind of programming, which is really really the definition of live fire, just the inherent danger here, then what happens when people modify that programming, Because I mean, you know this, Jonathan, there's this big there's this increasingly big stink over the computerization of cars and the degree of control over that that programming is still held by manufacturers.

Sure, and people are buying cars and saying, why can't I work on my car?

Right?

And then there's the hacker community that have uh said that they would be able to get control of various car systems, and that's been demons demonstrated in the past.

Now until very recently, it's all been limited to people who are able to hook a laptop directly into the wiring of a vehicle.

So it meant that you had to have physical access to the car in order to have that happened.

Uh.

However, that being said, I know there's a conference coming up this summer in which a pair of presenters are going to say that they have found a way of remotely accessing a vehicle and hacking it.

Um.

I'm very curious to hear how they did that.

I imagine the only way I can imagine it working is if I mean, you would obviously have to have some sort of of transceiver on the car that could communicate out and receive communication back.

Uh.

And typically I would imagine that would be limited to a car's entertainment system.

Um.

But if that entertainment system is also hooked into the other various computer systems on a car, then that is a potential point of vulnerability.

Uh.

Now, if you've designed a car where it can wirelessly communicate with a mechanic to allow a mechanic to get more information about the vehicle and it's its current state.

Uh, then that would also be a potential point of vulnerability.

I just don't imagine that happening because it's I don't think that it would be I don't think the convenience of that particular system would warrant the investment in it because it's not that much more difficult for a mechanic to just hook their computer up to your cars computer system.

It's it's not like it's so hard to do that.

You know, we needed to find an alternative to that.

But that being said, knowing that this is a possibility that the ability to to hack into a car could become a thing, and knowing that autonomous cars are you know, they're they're growing reality.

It's something that you know, the roboticist Henrik Christensen said the kids born today will never have to drive a car because the cars that they'll have when they reach driving age will all be autonomous.

Uh.

You know, I I think that's fairly accurate.

I mean, some people say it won't be till I don't think that's true.

I think that we're much closer to seeing it become a commercial possibility within within a decade.

That all of that is to say that because of that that reality of the autonomous cars, something like the possibility of hacking is certainly a concern.

If you are a hacker who owns your own car and you wanted to remove things like the cap on speed limits so that your car could drive much faster, even if you're not the one physically controlling it, that you just enjoy the experience of writing in a really fast car.

That's a concern if you are able to do things like, for whatever reason, remove the safety features that would prevent your car from uh breaking too late or accelerating too early, or position itself correctly for a crash.

Because you don't want to lose a big investment.

Yeah, any of those things would be you know, a concern.

So uh, I mean, these are questions that we don't have real answers to yet because we're not we're not technologically there yet, but they are definitely questions that need to be thought of on the engineering and manufacturer side of things, because these are things that are gonna come up, and you know, you might argue, well, this is a now, it's a niche thing.

It's only gonna happen like a few times, but potentially that could be devastating if you don't actually address it.

Yeah, and and also you know, it's a tough spot for carmakers because as as reasonable as it is for you and I, um and probably a lot of you guys listening out there too, as reasonable as is for all of us as individuals to say, well, the utilitarian answer is correct, Spock is correct.

If GM or someone came out and said, well, guys, autonomous cars are still in the early days, so here's what we're gonna do if someone has to die, if this situation comes up, that it's going to make sure it only kills one person, set of five.

Then immediately you've got everyone say, all right, we have it on record that this this car company has said it's okay to kill people.

Right, so when it when it happened, as we can totally go and say no, you have it on record that this is what you've decided to do.

Yeah, no, it's it's definitely Again, that's why I would think that car manufacturers would try to limit the liability.

They would be there and and say like, well, uh, you know, maybe maybe it ends up being common practice that when you buy a car that has an autonomous control feature on it, that you also end up signing waivers that wave your legal rights to go after a company should one of these crazy scenarios come along.

I don't know if that would actually hold water, because I would say, imagine there'd be a lot of pushback on that, saying, well, how can you hold a a customer culpable when they have literally no control over the vehicle.

They didn't build the vehicle, they didn't design the software, they didn't decide how it navigates.

They just told it a destination.

How can you hold them responsible for that just because they own the machine.

They didn't make it do the thing it did.

Um, these are great questions.

I love, I love thinking about it.

I wish I had more answers, But honestly, it's one of those deals that I think is going to shake out as the technology matures, and uh, you know, in the best case scenario, they will largely be moot questions because the implementation of the technology will be sophisticated.

Enough where these these problems will rarely, if ever pop up.

That's that's a dream, that's very optimistic.

But but you know, shoot for the stars, right, I want myself driving car.

I want to make that clear.

I am not against the idea of autonomous cars.

I eagerly await the day when I can have an autonomous car take me to work.

Yeah, it was sure would be nice to get here in twelve minutes in seven hour, you know, And I don't drive, so for me, I would need an autonomous car because I'm not a Driver's not into it.

I know.

I've had him on the show a couple of times to talk about In fact, we had a discussion about hacking cars, and we also did a discussion about autonomous cars.

And I can see, you know, every time I started talking about autonomous cars, he started to squirm in his seat a little.

You know, not all of us are like you know, race car drivers in our heads with lead feet, lightning like reflections.

Some of us just want to get to where we need to go and and not have to deal with the navigating through traffic on our own.

I mean, I've seen what it does to the same people.

I cannot take that Ben would not work.

H Yeah.

Well, well, luckily, it looks like every everything that we have seen tells us that there is a rise in autonomous vehicles, and they are at some point wherever you are listening right now, they are at some point coming to your town.

Yeah.

Yeah, some of them may be driving around right now, depending on where you're at.

Um.

Yeah, I've seen uh, I got to see a few when I went out to San Francisco.

It's pretty cool.

All right, Well, Ben, thank you so much for joining me for that segment, and I'm sure Kegan appreciates it too.

Yeah, thanks for having me, guys, And remember you can check out Ben's shows.

He's on car stuff, He's on stuff they don't want you to know.

He's also a writer and appears for shows like brain Stuff and What the Stuff.

Uh So Yeah, if you guys are are big into video, you should definitely check those out.

You should check out the other podcast as well.

And of course it's always a pleasure having you on and I will be asking you to be a guest probably again really soon.

Hey, thank you.

Also, just to point out to everybody, uh, Jonathan, in full disclosure, you're on like half of those shows as well.

That's true.

I also appear on those shows.

So yeah, if you if you haven't seen What the Stuff like, go ahead and give a quick description.

So What the Stuff is kind of a show where we pick interesting topics and give really cool examples of whatever that topic happens to be.

So, uh, you know, I've got one coming up about honest sports in the world, and I talk about some really crazy ones.

Uh.

And then brain Stuff kind of a dive into really cool questions about how stuff works or interesting science questions that kind of thing.

Um, and those are always a lot of fun to do to.

Yeah, you've got uh, let's see you guys know, Jonathan's a little bit of an actor in a different light, right, Yeah, that's true.

Can I say that on the air?

Absolutely?

And uh did this did this great piece on brain Stuff about why people in old movies the Transatlantic accent?

Yeah, I think we've made the title why do people talk funny in Old Movies?

I think it was that ended up being a really popular one.

Check that out if you want to see some of the guy's chops, right and um, and and you know, obviously we've got tons of other shows here and how Stuff works not to mention crazy amazing articles.

I don't know how many of you have been to the house stuff works dot com website, but you should definitely check it out.

Check out.

We're gonna have some great things to announce in the near future.

Um really excited.

I mean it's been it's always been a great place to work, but it's really becoming like crazy exciting over the last year.

So check that out, guys.

I know that I haven't really talked about the website in a long time, but you definitely need to check that out.

All right, That wraps up this discussion.

If you guys have any questions, if you want to have your listener mail, be part of the next grab bag.

I want to do a lot more of these in the future.

It's always fun to kind of respond to what you guys have to say and think.

Send me a message that email address is tech stuff at how stuff works dot com, or draw me a line on Twitter, Facebook or Tumbler at all three.

I use the handle of text stuff h s W and we'll talk to you again really soon.

For more on this and thousands of other topics Works dot Com