SIMON ADLER: It's like -- yeah, anyhow.
ROBERT KRULWICH: So what have you been up to?
JAD ABUMRAD: Annie, do you need a microphone?
ANNIE MCEWEN: Uh, I think we're going to share, 'cause what else are we gonna do? It's good. It's fine.
JAD: Annie's gonna have to sort of strike like a viper.
JAD: Hey, I'm Jad Abumrad.
ROBERT: And I'm Robert Krulwich.
JAD: This is Radiolab.
ANNIE: I will go like this.
SIMON: As long as you’re comfortable with that.
ROBERT: And today ...
JAD: A story about when things from far, far away ...
ANNIE: Let's combine our heads.
ROBERT: Hit very close to home.
JAD: Just put your hand on his forehead.
ANNIE: I will.
SIMON: Or over my mouth. That's fine. [laughs]
ROBERT: Comes to us from producers Simon Adler ...
SIMON: Buh buh buh buh bum!
JAD: And Annie McEwen.
ANNIE: Buh buh buh buh bum!
ROBERT: Okay. So you're gonna tell us something about something?
ANNIE: You, you've put ...
SIMON: Should -- should I start us?
ANNIE: Yeah, yeah.
ROBERT: You start.
SIMON: Okay. So we've been exploring what happened on May 18th, 2003, in the great city of Brussels.
SIMON: Specifically in Schaerbeek.
EMMANUEL WILLEMS: Schaerbeek. One of the municipalities in Brussels.
SIMON: Story starts with this guy.
EMMANUEL WILLEMS: So it was -- it was around 11:30 pm. And I was with one of my colleagues at a restaurant to have one of the good meals of the day.
SIMON: This is Emmanuel.
EMMANUEL WILLEMS: Emmanuel Willems. I am the -- the one everybody calls when shit hits the fan. Yeah, let's put it that way. [laughs]
SIMON: And that evening, Emmanuel, along with much of the rest of the country, was waiting around for that day's national election results.
EMMANUEL WILLEMS: The results for all the different parliamentary assemblies.
SIMON: But Emmanuel's mind with on something else. Because across the country that day, millions of people had just voted on computers. Belgium was in the early days of experimenting with electronic voting, and so instead of filling in a paper ballot, in the voting booth was a touch-screen computer, and voters would walk in, insert this white credit card-sized card into the machine and then vote by selecting names on that screen. And it was Emmanuel's job to make sure that all these hundreds of computers worked properly.
EMMANUEL WILLEMS: Mm-hmm. I'm an IT guy. So something like that.
SIMON: And at that point, at 11:30 at night, after the polls had closed ...
EMMANUEL WILLEMS: Besides the usual computer glitches ...
SIMON: It seemed like everything had gone well.
EMMANUEL WILLEMS: And so we were -- there were two of us.
SIMON: He and one of his fellow IT guys.
EMMANUEL WILLEMS: Waiting for the others, I guess, to talk about the day.
SIMON: Pat each other on the back for a job well done.
EMMANUEL WILLEMS: And that's when the ...
SIMON: That's when his phone rang.
EMMANUEL WILLEMS: It was unexpected. And it was one of the officials of the ministry. They said that they had a problem, that they had detected a problem in Schaerbeek. And so I say, "Okay, we're coming over."
SIMON: And so you pay the -- the check, you hop in the car and ...
EMMANUEL WILLEMS: I hope we paid the check!
ANNIE: We dashed out of there.
EMMANUEL WILLEMS: We just didn't leave like that.
SIMON: Fifteen minutes later ...
ANNIE: They arrive. They get out of their black Escalades, as I like to imagine and they -- they enter the polling station. [laughs]
EMMANUEL WILLEMS: The townhouse of Schaerbeek.
ANNIE: This ornate building with big tall ceilings.
EMMANUEL WILLEMS: It's a very old building.
ANNIE: Filled with this crowd of confused-looking people.
EMMANUEL WILLEMS: Well, they explained to us that, going through the numbers they noticed one of the contestants ...
SIMON: This relatively unknown Communist Party candidate, Maria Vindevoghel.
EMMANUEL WILLEMS: Had a very high number of votes from this one polling station.
EMMANUEL WILLEMS: And taking a closer look, they noticed that the number of votes for this contestant was impossible.
JAD: Wait. Impossible?
EMMANUEL WILLEMS: Yeah. A mathematical impossibility.
SIMON: Essentially, she received more votes than there were voters who could possibly vote for her.
SIMON: And so as Emmanuel's hearing this ...
EMMANUEL WILLEMS: I'm curious. Is this just a fluke? Is there an explanation?
ROBERT: You're thinking fraud?
JAD: I'm smelling fraud.
EMMANUEL WILLEMS: It could have been tampering. It could have been a bug. It could have been a hardware glitch. It could have been uh ...
SIMON: It could be anything.
EMMANUEL WILLEMS: Yeah, absolutely.
SIMON: So first things first, they got all of the ballots. Those white, plastic magnetic cards that the voters had loaded their votes onto.
EMMANUEL WILLEMS: And recounted. Reinserting every single magnetic card one by one.
SIMON: This took several hours. So now it's, like, 1:30 in the morning. And they print out the new recounted report, hold it up against the old one ...
EMMANUEL WILLEMS: And they are the exact same results for every single contestant to the vote, except the contestant which had had this abnormal number of votes.
SIMON: This time around, Maria had far fewer votes.
EMMANUEL WILLEMS: And at that point, one of my colleagues did the math and said the difference between the number of votes she had in the first count and the number of votes she had in the second count, it's exactly 4,096 votes.
ANNIE: She had 4,096 fewer votes.
ANNIE: Does that mean anything to you?
ROBERT: 4,096? No, it means nothing to me.
SIMON: Well ...
EMMANUEL WILLEMS: 4,096 is not a random number.
JAD: It is a -- it's something about this number. This ...
ANNIE: Well, tell us!
JAD: I don't know -- I don't know the answer, but I have this -- a deep sense. I'm feeling tingly.
ANNIE: So if you guys were IT people and you heard the number 4,096, you'd be like, "Great Scott!"
ROBERT: Would we?
ANNIE: Yes you would. [laughs]
EMMANUEL WILLEMS: Yeah, absolutely. To every IT guy in the room the number 4,096 was -- was very meaningful.
SIMON: In fact, it was a clue. A clue to a cosmic culprit capable ...
NEWS CLIP: Instead of ...
NEWS CLIP: ... of voting systems being hacked ...
ANNIE: Of swinging our elections ...
NEWS CLIP: Have switched votes ...
NEWS CLIP: She said, "I can't slow down."
ANNIE: Crashing our cars ...
NEWS CLIP: Runaway Toyotas. Taking off on their own ...[/i]
SIMON: And reminding us just how small and at the mercy of the universe we really are.
JAD: So wait. So why -- why is 4,096 so important?
SIMON: Okay, well to explain, I'd like to back up to the morning of the election.
SIMON: So polls open and at some point, someone slid their white card through that card reader. And Maria Vindevoghel received her first vote.
ROBERT: It registers vote number one for Maria.
SIMON: Yeah. Now when that happened, inside the computer in its memory, a zero switched to a one.
SIMON: And I mean that both in terms of her vote total, and also in how a computer counts.
EMMANUEL WILLEMS: The way a computer works is that everything is stored in memory. And memory is binary. It's either a zero or a one. Either on or off.
SIMON: Zero is off and one is on. And -- and I think we all know this, right? Or we've heard that at least, that deep down computers are zeros and ones.
JAD: Sure, yeah. Yeah, yeah.
SIMON: And so in this specific case, what that means is all these individual votes that Maria would get throughout the day, they all were stored as just ones. As these individual ones.
SIMON: Now when you're limited to counting with just ones, a way to do it is to tally like you do in groups of five on a chalkboard.
JAD: Right. Diagonal line.
ROBERT: Diagonal line.
JAD: Bundling it.
SIMON: Yeah, exactly. But when that tally starts to get high, it becomes incredibly inefficient to count that way. So instead, to save time, energy and computing power, computers do a little bit of math. They -- they count using powers of two.
JAD: Powers of two. Mm-hmm.
SIMON: You remember powers of two?
ROBERT: Something squared is a power of two.
SIMON: Exactly. Two squared, or two to the second power, two times two is four. Two to the third power, two times two times two is eight. Two to the fourth, 16. Et cetera, et cetera, et cetera.
JAD: So how is it that a computer is doing that?
SIMON: Right. Well, so to help you visualize this, imagine inside a computer's memory there's this string of 13 light bulbs. Light bulbs that IT people call bits. Each one representing a different power of two. Just stick with me here for a second.
JAD: Yeah, okay.
SIMON: So that first light bulb represents ...
COMPUTER: Two to the power up zero.
SIMON: Which is one. Because anything to the zero power equals one.
SIMON: Don't ask. Math.
SIMON: Second light bulb ...
COMPUTER: Two to the first power.
SIMON: Which is two. The third bulb ...
COMPUTER: Two to the second power.
SIMON: Or two times two.
SIMON: Et cetera, et cetera, et cetera. Now, when Maria received her first vote, a little bit of electricity was sent to that first light bulb.
COMPUTER: Two to the power of zero.
SIMON: Turning it on.
JAD: Two to the zero is one.
SIMON: Exactly. Now when she got her second vote, this time some electricity was sent to that second bulb.
COMPUTER: Two to the first power.
SIMON: Turning it on, representing two votes. And because two votes are stored in that second light bulb, what -- what the computer does simultaneously is turn off that first bulb. So to be clear, when she got her second vote, the first bulb was turned off, and the second bulb, which represents two to the power of one, was turned on. So it's now only using one bulb to represent two votes.
JAD: Oh, okay!
SIMON: Now, in order to represent numbers that aren't powers of two, all the computer does is turn on multiple bulbs at once.
EMMANUEL WILLEMS: A sequence of power of twos.
SIMON: For example, when Maria got her third, it turned on the first and second light bulb.
COMPUTER: Two to the first power.
SIMON: Because ...
ROBERT: One plus two equals ...
SIMON: Three. When she got her fourth, it turned off the first and second light bulb and turned on the third.
COMPUTER: Two to the second power.
SIMON: Fifth vote turns on the first and third.
COMPUTER: Two to the power of zero, two to the second power.
SIMON: When she got, say, her 17th vote, it turned on the fifth bulb.
COMPUTER: Two to the fourth power.
SIMON: 16 plus the first bulb.
COMPUTER: Two to the power of zero.
SIMON: One. equaling 17. And so on and so on, adding up these powers of two.
EMMANUEL WILLEMS: And with that respect, 4,096 is a power of two.
JAD: I see! So that number is one that they're, like, blink!
ROBERT: Is that what that is?
EMMANUEL WILLEMS: Yeah. It's the 13th bit of memory. Which meant that ...
SIMON: At some point during that day, something somehow turned that bit on. It flipped what should've been a zero to a one.
EMMANUEL WILLEMS: Yeah. It was -- it was a binary flip.
SIMON: Your mind just went there immediately.
EMMANUEL WILLEMS: Oh yeah, absolutely. Yeah. [laughs]
SIMON: I love that you -- all the IT people are like, "Oh!" And all the non-IT people are like, "What's going on?"
ANNIE: What, what, what? [laughs]
EMMANUEL WILLEMS: Well, to each his trade.
SIMON: Well okay, so you know that it was just one bit that flipped. What does that make you ...
EMMANUEL WILLEMS: No, we knew -- so we knew that the result could be traced back to the 13th bit flipping. What we didn't know was what caused the bit to flip.
SIMON: Okay. So what would -- that's your next question: what -- what could cause this bit to flip?
EMMANUEL WILLEMS: No idea.
SIMON: [laughs] What were the options? What did you ...
EMMANUEL WILLEMS: Well, it could have been -- it could have been any number of things. And so we each -- in the following days we got access to absolutely everything.
ANNIE: First thing they checked ...
EMMANUEL WILLEMS: It could have been the software. So we carried out random tests on the source code.
ANNIE: Looking for any bugs that that could have caused this bit to flip.
EMMANUEL WILLEMS: But there was nothing major. So we were fairly confident that it wasn't a software error.
SIMON: Okay. So if it's not software, it's hardware, right?
EMMANUEL WILLEMS: Well, we got the machine that was actually used in that polling station on the day of the election.
ANNIE: The computer that had done all that tallying and counting. And they ran a bunch of memory tests on it.
EMMANUEL WILLEMS: Memory testing to see if something came up.
ANNIE: To see if some defect in the physical memory could have caused this bit to flip.
SIMON: What came back on that front?
EMMANUEL WILLEMS: Nothing. Not a single error. So, the memory on that computer was working fine.
SIMON: Well, so what the heck are you left with?
EMMANUEL WILLEMS: Not much.
EMMANUEL WILLEMS: As far as we were concerned, this one computer had had a glitch, which caused it to give a wrong answer.
ANNIE: But isn't that sort of an unsatisfying answer?
EMMANUEL WILLEMS: That's the way things go. You know, we just simply had no proof of what caused the bit to flip. We just don't know. That's the way it -- it turned out.
SIMON: And so over the next couple of days, Emmanuel and his team sat down to write up these findings. Half-expecting that once finished, this report would be filed away and never mentioned again. However ...
SIMON: Hello? Hello? Brussels, can you hear us?
NATALIE: Hi, is that Simon? It's Natalie here in Brussels. How are you?
SIMON: That was not the case.
NATALIE: I have your guest here. Armed with plenty of tea. I'm just going to let her have a seat and have a chat with you.
ANNIE: Thank you.
SIMON: Yeah, thank you Natalie.
NATALIE: No worries. Okay.
ZOE GENOT: Merci. Hello?
SIMON: Hello? Uh, is this Zoe?
ZOE GENOT: Yeah, yeah. It's me.
SIMON: Hello, Zoe.
ZOE GENOT: Hello. Hi.
SIMON: All thanks to this lady.
ZOE GENOT: Danke. I'm Zoe Genot. I am member of the parliament.
SIMON: And back in 2003, young and freshly elected to parliament ...
ZOE GENOT: I was 28 or 29.
SIMON: Like, half the age of most of the MPs.
ZOE GENOT: Yeah. It's that.
ANNIE: She was generally aware of what went down in Schaerbeek. That there was this computer problem. Someone got some extra votes.
ZOE GENOT: It was really strange. And for me, it was really a concern because our most concern is that the people can trust the electronic system. And with this -- these sort of problem, we are not sure the people can trust the system. So we waited for the report of the specialist.
SIMON: And finally, about three weeks after the election, it came out.
EMMANUEL WILLEMS: Yeah. We only -- well, we ended up writing that at some point during the day, the 13th bit had flipped.
EMMANUEL WILLEMS: That's all we wrote. Basically.
ROBERT: It means nothing. It means nothing.
SIMON: Well, then that -- that's what she's thinking.
ZOE GENOT: No, no. No, no. There -- there must be more. Nobody can understand this sort of vocabulary. And I was confused with the explanation.
SIMON: Because way down at the bottom of the report in the final paragraph, the only thing they said about a possible cause was, and I quote: "This phenomenon is extensively described in the specialized literature."
EMMANUEL WILLEMS: Yes, we did write that.
SIMON: And what -- what were you referring to there?
EMMANUEL WILLEMS: Nothing in particular.
ZOE GENOT: I said, "What's this? Why they not explain what's really happening?"
SIMON: So Zoe went to her fellow parliamentarians and was like ...
ZOE GENOT: There was a problem. We don't know what really happened.
SIMON: Like, we've got to look into this.
ZOE GENOT: What can we check? Who can check? Who will check? But the people, the older member of the parliament, they say, "Oh, yeah. Zoe Genot she's there with the strange style. And she's there speaking about problems."
SIMON: Basically, forget about it.
ZOE GENOT: It's done. We are elected. The election are gone. There was no problem. But I said, "No, no. We must find out what was really the problems and speak about it."
DAVID GLAUDE: So Zoe contacted me and she did send me the paper report.
ANNIE: She reaches -- she reaches out to this guy named David Glaude. He is a member of a group called…
DAVID GLAUDE: Our name is Poureva. And we are fighting for ethics into electronic voting.
ANNIE: He hates electronic voting.
DAVID GLAUDE: Basically, we are against voting machine. Yes.
SIMON: But interestingly he's also, like, a hot shot IT guy, works for the European Parliament.
DAVID GLAUDE: Yeah. And as an IT people, I tell you don't trust us. And don't put everything in our hands.
ANNIE: Anyhow, Zoe had known him for a long time and so she reached out and asked, "Can you help me understand what are these random bit inversions? Like, what is this?"
DAVID GLAUDE: So the report? I read it and I said, "Hey, wait. I did computer science studies and bits are not flipping. What are they talking about?" So I -- I needed to find out. So I guess I started to Google to ask what could explain this.
ANNIE: Initially, he came across some articles or blog posts arguing that these sort of things could always be traced back to some software bug.
DAVID GLAUDE: But they did not find a bug, so they -- they blame it on something else.
ANNIE: Not totally satisfied, he tried some new search terms.
DAVID GLAUDE: Blah, blah, blah. Bit flipping.
ANNIE: And came across a couple articles that made this strange claim.
DAVID GLAUDE: They said the cosmic ray could flip a bit of memory.
ROBERT: Wait, what?
DAVID GLAUDE: So they said yes. If the cosmic ray touched your computer the right way at the right spot, it could change a bit from zero to one.
ROBERT: Gamma rays aimed at Belgium in favor of a particular Walloon!
SIMON: We're not gonna be able to use any of this tape. [laughs]
JAD: No, this is great!
ROBERT: Well, so you're saying that what created this effect in Belgium on that day was a zing from deep space?
SIMON: That was the theory that David came across.
DAVID GLAUDE: And my first reaction was to say, "Ha ha ha! Cosmic ray. Hey, what a joke!" But then it become a bit bizarre, because I found out about ...
[CLIP OF IBM AD: This is IBM ...]
DAVID GLAUDE: Some IBM research.
[CLIP OF IBM AD: Freeing man's mind to shape the future.]
SIMON: An IBM paper from the '70s. And in it ...
[CLIP OF IBM AD: Electronic data processing center.]
DAVID GLAUDE: They made a lot of studies about random bit flipping and problem with computer memory.
SIMON: The report explained how in the '60s, NASA had noticed these random bit flips in their satellites.
SIMON: And then more recently, how IBM itself ...
[CLIP OF IBM AD: Problems by the computer can be displayed in various ways.]
SIMON: ... had studied their computers at higher altitudes.
DAVID GLAUDE: So closer to a space, let's say.
ROBERT: Closer to the sky.
JAD: Closer to God.
DAVID GLAUDE: And they found out that there were more bit flipping up there than at the sea level.
SIMON: And the only explanation they could come up with for -- for what could be causing this was -- was cosmic rays.
BHARAT BHUVA: Yes. When they investigated, they found out that it could have been cosmic rays causing problems.
DAVID GLAUDE: I said, "Oh, okay."
BHARAT BHUVA: Because cosmic rays essentially are charged particle, and so electronic systems can get affected.
ANNIE: This by the way, is Professor Bharat.
BHARAT BHUVA: Bharat Bhuva. I'm a professor at Vanderbilt University in electrical engineering and computer science department.
ANNIE: And while Bharat wasn't involved in that specific IBM paper, he has essentially dedicated his life to studying the effects of cosmic rays.
BHARAT BHUVA: So we started noticing on the surface of the Earth, this problem started around 2001-2002 time frame.
BHARAT BHUVA: Right. And they held big ramifications.
JAD: Is this a thing people that know about? I mean, is this like a ...
ANNIE: Well, yeah.
JAD: Because this is the first time I hear of it.
ANNIE: Well it’s ...
ANNIE: I don’t know if we should continue on the story or go into cosmic ray land.
SIMON: Yeah, what do you want? Do you want narrative or do you want cosmic ray land?
JAD: Ooh. I love that that’s a choice we get to decide. I’m torn.
ANNIE: Well I don’t know. Maybe just go with the story. I don’t know.
SIMON: I think we dive into -- we give some of this.
ROBERT: Oh, yeah.
SIMON: Okay, so according to Bharat, to understand what happened that day in Belgium, we have to go back to a long time ago. Wait for it -- in a galaxy far, far away.
[STAR WARS THEME IN]
BHARAT BHUVA: I'm going to generalize way too much.
ANNIE: No, no! That's good, that's good.
BHARAT BHUVA: Way too much.
SIMON: Bharat says in this distant galaxy there was a very, very old star that exploded.
BHARAT BHUVA: And when the star explodes, it sends out a huge amount of material.
SIMON: All sorts of little tiny charged particles.
BHARAT BHUVA: These charged particles are thrown into space. And these particles are going to travel through space for millions of light years.
ANNIE: Uh, Simon?
SIMON: Right. Sorry. These particles are passing through other galaxies.
BHARAT BHUVA: They get deflected, they get trapped. Going around suns, the gravitational pools ...
SIMON: Until finally on May 18th, 2003, one of these particles ...
BHARAT BHUVA: With lots and lots of energy ...
SIMON: Enters our solar system. And not long after that ...
BHARAT BHUVA: Comes into the atmosphere ...
SIMON: And crashes into an earthly atmospheric particle. And in that moment ...
BHARAT BHUVA: Energy's transferred to the atmospheric particle.
SIMON: Essentially, it hands it a little bit of energy, causing that newly energized atmospheric particle to start descending towards Earth.
BHARAT BHUVA: That is correct.
SIMON: But on its way down, it smacks into a couple other particles, passing some of that energy on again. Causing those particles to fall and crash into other particles. And this happens again and again as these particles and this cosmic energy catapult towards the surface of the Earth.
BHARAT BHUVA: And so you get this cascade. It's a shower of all these particles. From the top, it would look like a Christmas tree.
SIMON: Now, down at sea level ...
[NEWS CLIP: This weekend's general election ...]
SIMON: Down on the surface ...
[NEWS CLIP: A divided country, Belgium is going to the polls.]
BHARAT BHUVA: The electronic system that was counting the votes in Belgium, zooming along, humming along, you know, in its own way.
COMPUTER: Two to the sixth power.
SIMON: Tallying the votes by turning on those light bulbs. Now, the switch that turns on those light bulbs is called a transistor.
BHARAT BHUVA: The transistor turns on or off. And that creates binary ones and zeroes.
SIMON: When it’s on, it sends out a little electricity switching on a bulb. But the thing is, these transistors can be tricked into turning on. So imagine ...
BHARAT BHUVA: Suddenly, this particle from outer space ...
SIMON: Comes zooming down and enters that voting machine, hitting precisely the transistor that controls that 13th bit.
BHARAT BHUVA: And that instant ...
SIMON: The transistor is tricked and sends out a burst of electricity to that 13th position.
COMPUTER: Two to the twelfth power.
BHARAT BHUVA: And the bit flipped.
SIMON: A zero was changed to a one, and from the heavens those 4,096 votes appeared.
DAVID GLAUDE: Yeah, what -- what?
ANNIE: Again, David Glaude.
DAVID GLAUDE: I don't know. I mean it's like, if I talk about cosmic ray on one side, I'm afraid that people will take me for somebody crazy. But wait, you can read it too, if you Google those words, you will find a lot of scientific stuff talking about that.
SIMON: And so not long after, David and Zoe met up, and David told her what he'd found. That -- that it’s possible -- who knows -- but it’s possible, that the vote was changed by the cosmos.
ZOE GENOT: I -- a cosmic ray got down in the computer and it make a mistake. I say, "It's not possible. You are really sure?"
DAVID GLAUDE: I've got no clue if that's the true explanation or not. And there is no way to prove that it's a cosmic ray. But I mean, it’s the only thing which matched the explanation.
ZOE GENOT: I said, "Oh, no! I will never can explain that in the Parliament. They’re all going to laugh!"
SIMON: Aware of all of this, at the next session of parliament ...
ZOE GENOT: I decide to ask a question to the new minister who was responsible for the election.
SIMON: So the session gets underway. And eventually the agenda turns to Zoe. Now, there’s no audio recording of this session. But based on transcripts, it went something like this. She walks up to the podium at the front of the chamber and begins.
ZOE GENOT: [speaking French]
ANNIE: She summarized what had happened in Schaerbeek.
ZOE GENOT: [speaks French]
ANNIE: And then explained what David had told her.
ZOE GENOT: That this bit inversion, it's cosmic ray who caused a problem in a computer.
ZOE GENOT: The room was there joking, ha ha ha. Zoe Genot, come here with the little green men.
ANNIE: "Ha, ha! Zoe, what a dumb dumb!"
SIMON: And the minister ...
ZOE GENOT: He say, "Yeah, yeah. Don't worry. It was a small error somewhere. But we see here there's no problem with the result of the election."
SIMON: With that, the minister sort of motioned to move on, but before her time could expire ...
ZOE GENOT: [speaking French]
SIMON: Zoe jumped back in and basically said, "Wait a second. Think about what this could mean. If this is true, like, can we trust these machines?"
ZOE GENOT: So I asked, "Are you sure there was no other cosmic ray somewhere who changed some of the result of the election?" And the minister ...
SIMON: After pausing for a moment.
ZOE GENOT: Admits he was not sure. And when the minister say, "Yeah, perhaps there are other problems we don't know," there was a strange silence. And nobody can laugh anymore.
ROBERT: Democracy and stars do not get along, apparently. Ipso facto dingo dango.
JAD: I think someone here needs to hold the skepticism. I’ll be the skeptic.
ROBERT: You be the skeptic.
JAD: What is -- what are the odds? Supposing this is the truth.
ROBERT: Well the odds are apparently, like ...
SIMON: There’s no question that this does happen, that a cosmic ray can hit a computer and goof things up.
EMMANUEL WILLEMS: I mean, the only thing I can is that ...
SIMON: I mean, even Emmanuel will concede that.
EMMANUEL WILLEMS: Yeah. Cosmic radiation can have an effect on electronic computers.
ANNIE: And Bharat and his team have actually synthesized this happening.
BHARAT BHUVA: If we suspect a bit flip, we will actually put the computer in a linear accelerator. Usually it’s a cyclotron.
ANNIE: Which is basically this mini controlled supernova.
BHARAT BHUVA: Pretty close to it. It’s all magic to me. But what they do is in cyclotrons, they can get you the ...
ANNIE: They put the computer in this concrete bunker, point the particle beam at it ...
BHARAT BHUVA: The electronic system's working, we drag all the wires out of the room. Then there are a few buttons to push: three, two, one, the light comes on, alarm sounds. Beam time.
SIMON: It's called beam time?
BHARAT BHUVA: That is correct.
ANNIE: Oh wow, that's great!
SIMON: That's so great!
SIMON: They blast the computer with these tiny particles.
BHARAT BHUVA: A million particles per second.
SIMON: Just like those cosmic rays flying through space. And as this is happening, Bharat says ...
BHARAT BHUVA: You can sit there and watch the bit flips.
SIMON: And given this fact, what Zoe was really asking was ...
DAVID GLAUDE: If the extra vote, were not 4,096, but the small bit flipping, then would you notice?
SIMON: Like, what if a candidate just got 32 extra votes, and that invisibly changed an outcome?
DAVID GLAUDE: That -- yeah, that's the point. I mean it's only when the computer produced an impossible result that you can tell that there was something wrong. And -- and we call that the tip of the iceberg.
SIMON: You want us to just throw to break, or do you guys want to come in and -- and toss.
ROBERT: Let's hear what you do and then we'll come in and ruin it.
SIMON: Classic. Music swell, music swell, music swell, and we're like: so, buckle up. When we come back, we are driving full steam ahead into that iceberg to figure out just how deep it goes.
ROBERT: What an interesting metaphor!
JAD: That's a mixed metaphor.
ROBERT: You're gonna drive a car into an iceberg.
SIMON: You can drive a boat!
JAD: Love it!
ROBERT: Oh, my! You have to ask Mrs. Vanderbilt, your third-grade teacher, to go back to the metaphor section of the lesson. In the next hour.
[ANGELA BABIARS: This is Angela Babiars from San Jose, California. Radiolab is supported in part by the Alfred P. Sloan Foundation, enhancing public understanding of science and technology in the modern world. More information about Sloan at www.sloan.org.]
[PAT WALTERS: Hey, everybody. Pat Walters here. I'm a producer at Radiolab, and I'm here because I need your help. This summer, I'm hosting a series of stories on the show and I have requests for those of you who spend a lot of time with kids: parents, aunts and uncles, teachers. We're looking for stories about what we're calling tiny moments of childhood brilliance.
PAT: Basically, I want to hear about those times when a kid you know did something that just made you lean back and say, "Whoa, how did they do that?" Maybe it was the moment that a kid you'd been reading to for months started reading back to you. Or maybe the kid was at piano lessons and you suddenly notice they were doing advanced math on the margin of their musical score. Or maybe the kid was in math class and you noticed they were writing music in the margin of their geometry homework. We're interested in those small, specific moments where a kid does something super-smart, but it doesn't have anything to do with a test. If you have a story, please share it with us and go to Radiolab.org/brilliance and record a short audio message for us. Again, that's Radiolab.org/brilliance. Thank you so much.]
JAD: Okay. Ready?
ROBERT: So, all right. So we just heard that a single cosmic ray from deep space flipped a particular election result. I guess the next question is, like, you know, is this happening all the time? Some of the time?
ANNIE: Well, if you mean, like, how often do those charged particles hit Earth?
BHARAT BHUVA: Every day, every second, we are bombarded by these particles.
ANNIE: Quite often. Because it turns out these cosmic rays ...
[MOVIE CLIP: They’re coming to get you, Barbara.]
ANNIE: They’re coming.
JAD: They are?
[MOVIE CLIP: They’re coming for you. Look, there comes one of them now!]
BHARAT BHUVA: Right now, right through your body you are getting thousands of these particles going through.
SIMON: Wha...? And ...
BHARAT BHUVA: And even if you are inside a building, you are still going to get all those particles, because most of these particles just cannot be stopped. They just go through anything and everything.
ANNIE: So Bharat seems pretty sure of himself.
BHARAT BHUVA: Yes, yes, yes.
ANNIE: But Simon and I were like, "Wait. We gotta see this for ourselves."
ANNIE: Is there -- we should turn off that screen, because we gotta make it dark in here.
SIMON: And apparently, although these particles are just too small to be seen with the naked eye ...
SIMON: We are currently in the studio, blacking the place out.
SIMON: Bharat did tell us there’s something we can build. Something called a cloud chamber, which is basically just an upside down cup you fill with vapor.
BHARAT BHUVA: Yes.
SIMON: That might let us see the trails these particles leave behind.
BHARAT BHUVA: If you can make one, that would be really, really good.
[YOUTUBE CLIP: Okay, here’s what you’re gonna need to make your very own at-home cloud chamber.]
ANNIE: Now it turns out there’s a lot of YouTube videos showing you how to do this.
[YOUTUBE CLIP: Hi, I’m Joanna. And I’m Steve ...]
ANNIE: One starring a kid who looks about four years old.
[YOUTUBE CLIP: ...comes from outer space.]
ANNIE: So ...
SIMON: Okay, so ...
ANNIE: We thought we had a pretty good chance of making it work.
SIMON: Whoa! It's cold!
ANNIE: Yeah, be careful.
SIMON: It's cold!
ANNIE: We got dry ice, cast iron pan.
ANNIE: From my kitchen.
ANNIE: Flipped it over.
SIMON: Okay, we are now putting the cast iron over the chunks of dry ice. [metal whirring]
ANNIE: That’s the sound of the metal cooling down [buzzing].
ANNIE: The most beautiful ...
JAD: That was amazing! The metal pan was singing. It was such a pure tone.
ANNIE: Good job, metal pan!
ANNIE: Next, take a see-through plastic cup.
SIMON: A 24-ouncer.
ANNIE: Some felt. Our felt is purple.
JAD: It’s like if Prince had a little bit more red in his purple.
ANNIE: Stick that felt in the bottom of the cup. Hold it in place with some magnets on either side of the felt.
ANNIE: Magnets, Jad.
ANNIE: Soak that felt in rubbing alcohol.
ANNIE: What do you think? Soaked?
SIMON: That seems pretty thoroughly soaked.
ANNIE: Turn the cup upside down onto the now very cold cast iron pan, and seal it in place with modeling clay. Overhead lights off, flashlight on. Aim it at the base of the cup, and then wait.
SIMON: Let’s watch for it.
ROBERT: How is this thing gonna work? What are we gonna -- what are we trying to see here?
ANNIE: So what we’re looking for, what should happen, is that that cold metal is going to make the alcohol vapor condense and, like, turn into a cloud.
ROBERT: So you're gonna make a cloud.
JAD: Kind of foggy looking in there. That’s what you want, right?
ANNIE: And through that cloud what we're hoping to see are these tiny little particles shooting through. And we won't be able to see the particle itself, but we will be able to see the trail they leave behind.
ANNIE: And it's supposed to be very similar to the trail that an airplane leaves behind in the sky.
ROBERT: A contrail, those things are.
ANNIE: A contrail, exactly. So -- so what we're hoping to see is these, like, tiny little airplanes with their tiny little contrails inside this cloud.
ROBERT: And that will be evidence of these very particles?
SIMON: So all three of us are on our knees in the studio. Pitch black at the moment. The only light is emitting from a headlamp on Annie’s head.
ANNIE: But ...
ANNIE: Maybe it’s one of those, like, watched pot never boils things.
ANNIE: Nothing happened.
SIMON: I don’t think there’s enough fog in there yet.
ANNIE: Come on, little particles!
ANNIE: After about an hour of this ...
[YOUTUBE CLIP: Now I’m going to put the Play-Doh around the cup.]
ANNIE: Pulling up YouTube on our phones.
JAD: I'm so jealous of this four year old.
ANNIE: I know.
ANNIE: To check our scientific method.
JAD: I don’t know.
ANNIE: Jad and Simon had to get back to work.
JAD: Good luck.
ANNIE: They had things to do.
SIMON: Okay, I’ll be back in ten minutes.
ANNIE: I stayed behind just in case.
ANNIE: Why isn’t it working?
ANNIE: Alone in the dark, with a little beam of light.
ANNIE: Oh! What was that? Come on, do it again. [ELECTRONIC ZAP] Oh! I saw one! I saw one!
SIMON: You saw one?
ANNIE: Yeah. Simon!
SIMON: What did it look like?
ANNIE: Like a tiny little ...
ANNIE: These very, very tiny white thin, thin threads ...
ANNIE: There! Did you see that?
SIMON: There. There. I’m seeing a bunch of them.
JAD: Wait. Where?
ANNIE: Suspended in the cloud. Holding their shape for just a moment. [ELECTRONIC ZAP]
ANNIE: Oh, there’s one. There’s a lot ...
JAD: Oh, wow!
ANNIE: Before disappearing.
SIMON: Oh, whoa!
ANNIE: What was that?
ANNIE: Once we started seeing them ...
JAD: Ooh, they’re just like -- they’re like little daggers.
ANNIE: We couldn’t not see them.
SIMON: A couple big ones there.
ANNIE: There’s, like, three there.
JAD: That one's coming right through me.
ANNIE: Okay, I can now tell you -- not just because someone told me, but because I saw it for myself, these things are definitely real and there are definitely a lot of them. And so I guess the next obvious question is, like, if we know they're here, besides that voting booth in Belgium, like, what else are they messing with?
SIMON: Like, has my phone ever -- are the -- is the probability such that my phone has likely been goofed by a bit flip caused by a cosmic ray?
BHARAT BHUVA: Absolutely. So all of our computers, all of our systems are getting bit flips on a daily basis.
BHARAT BHUVA: For example, the routers that we have on the internet.
SIMON: Did parts of the internet get shut off by these things?
BHARAT BHUVA: Yeah. Routinely.
BHARAT BHUVA: Even today.
SIMON: Whoa, whoa, whoa, whoa!
BHARAT BHUVA: Yes. People just assume that, "Oh, my internet went down for five minutes." All the internet companies, like Cisco Systems, they’re very keenly aware of this problem.
ANNIE: Apparently in the early-2000s, Google's indexing system was down for five months.
JAD: No kidding!
SIMON: What else has been goofed by these?
BHARAT BHUVA: A lot of manufacturing plants have been goofed by this.
BHARAT BHUVA: There is a custom manufacturing plant which made custom machines, and one of the bits flipped, so they were making things wrong. So ...
SIMON: Wait, what -- what’s a real example of that?
BHARAT BHUVA: Not sure I can discuss this on the air.
ANNIE: Oh really?
BHARAT BHUVA: Yes.
ANNIE: Can you say what -- so what kind of machine?
BHARAT BHUVA: Not sure I can say this.
BHARAT BHUVA: Sorry.
SIMON: Something that Bharat was able to tell us is that the higher up you are in our atmosphere ...
BHARAT BHUVA: The number of particles with higher energy is higher.
SIMON: And that means the chances of a bit flip are also higher.
BHARAT BHUVA: The electronics in airplane is always bombarded by these particles. And one time the plane is flying, it’s on autopilot. And suddenly one of the bit flips that takes it out of the autopilot.
BHARAT BHUVA: Yes.
ANNIE: Who -- who -- what airplane? Where?
[NEWS CLIP: Australian transport safety bureau ...]
BHARAT BHUVA: This was a Qantas Air flight in about 2008.
[NEWS CLIP: They're saying passengers and one crew member were seriously injured when the Qantas Airbus nosedived, plummeting more than a thousand feet.]
ANNIE: Oh, my God!
BHARAT BHUVA: So it essentially dropped like a rock.
ANNIE: And it was all because of one bit flipping.
SIMON: Which they believed was caused by a cosmic ray.
JAD: Wait. It just -- it just nosedived for a thousand feet and people, like, hit their heads on the ceiling?
ROBERT: Something that was on just turned off, and there was nobody ...
JAD: Oh, my God!
SIMON: Because of a bit flip.
JAD: Holy shit!
ANNIE: Now we should say that, like, even though bit flips are happening frequently, the vast majority of those flips ...
BHARAT BHUVA: They are benign.
ANNIE: We don’t even notice them flipping.
BHARAT BHUVA: Those bit flips matter only if it’s right there at the right position. So for example, in that parliament the bit flipped. But that bit flip mattered only if they were taking a vote at that time and actually tallying the votes.
ANNIE: So the likelihood of a bit flip having any consequence at all?
BHARAT BHUVA: It’s rare.
SIMON: Hello. This message is for either Grover or Barbara.
ANNIE: And that’s what made this next thing, Bharat told us ...
[MOVIE CLIP: They’re coming to get you, Barbara.]
SIMON: This is a strange call to receive, I’m sure.
ANNIE: ... even crazier.
BARBARA WALTEN: I’m sorry I didn’t answer the phone initially. I’m just tired of people who want to reduce my credit card interest or give me something for my pain or ...
SIMON: This is Barbara Walten. We reached her in Boiling Springs, North Carolina.
BARBARA WALTEN: Little one stoplight town.
SIMON: Where she lives with her husband, Grover.
BARBARA WALTEN: We’ve been here since we got married 53 years ago.
SIMON: We’re hoping to be able to understand, like, what it was like to be in one of those cars.
BARBARA WALTEN: Oh my gosh. Oh!
SIMON: Do you mind just, like, telling me what you were doing the morning of the accident?
BARBARA WALTEN: No. We live in North Carolina. We have a son that lives on the coast in South Carolina, so we were going to visit.
ANNIE: It’s about a four-and-a-half hour drive. Takes you through some pretty good-sized towns.
BARBARA WALTEN: The area we were in was more or less a rural area.
ANNIE: They’re driving a 2008 Prius. Grover is behind the wheel.
BARBARA WALTEN: And I was in the backseat because I was on my computer, and I didn’t like to do that in the front seat because of the possibility of the airbag coming out and throwing the computer in my face.
SIMON: And what were you -- if I can ask, what were you doing on your computer in the backseat?
BARBARA WALTEN: What would I have been doing? Probably solitaire. Something to occupy my time. I usually snooze or -- you know? We’ve made that trip many times, so I’ve seen the scenery.
ANNIE: For the first three-and-a-half hours of the drive, everything is totally normal. Grover calmly at the wheel, Barbara on her laptop. And then Barbara looks up from her computer.
BARBARA WALTEN: And I notice my husband is speeding. And I said, "What are you doing?" And he said, "I can’t slow down." And I said, "Well, you have to slow down." And he said, "Well, I’m applying the brake, it is not slowing down."
SIMON: When -- how -- do you guys remember exactly how fast you were going?
BARBARA WALTEN: He’s a much better judge of that than I would be. Let me ask him. G.H? When our car ran away with us, how fast would you say the car was going?
GROVER WALTEN: Probably 65.
BARBARA WALTEN: I’m sure it was more than that. Probably -- he’s saying 65. I’m saying 65, 70.
ANNIE: How are both of you feeling? I can imagine feeling, like, really panicked.
BARBARA WALTEN: Mr. Walten doesn’t panic. Mrs. Walten panics.
BARBARA WALTEN: I am freaking out.
ANNIE: At this point in the drive, they’re on a two-lane highway cutting through a huge wetland area. So on both sides it’s just swamp and water.
GROVER WALTEN: And I didn’t want to go down in the swamp.
ANNIE: That’s Grover Walten, Barbara’s husband. The driver.
GROVER WALTEN: Every time I'd attempt to brake, as quick as I took my foot off the brake it would start accelerating again. I finally just basically stood on the brake.
ANNIE: They way they described it, it was almost like one of those Hollywood movies, where you’re, like, careening down the highway.
BARBARA WALTEN: We are not slowing down.
ANNIE: Zipping by cars.
BARBARA WALTEN: Seatbelts on.
ANNIE: Weaving in and out of traffic.
BARBARA WALTEN: Computer's down, praying that this was gonna turn out all right.
ANNIE: For somewhere between 10 and 15 minutes, Grover fought with the car.
BARBARA WALTEN: I can’t remember if it -- how did you stop it? I know we crossed that swampy bridge.
GROVER WALTEN: I wore the brake pads and the rotor out.
BARBARA WALTEN: Okay, so that’s what stopped it. By just applying the brakes so long and so hard, it wore the brakes and the rotor out. And that’s how the car finally came to a stop.
ANNIE: They called their son, who did some quick research online.
BARBARA WALTEN: And apparently there were already complaints.
[NEWS CLIP: The driver of a Toyota Prius says he couldn’t get his car to stop.]
[NEWS CLIP: A terrifying account of a runaway ride in a Toyota.]
SIMON: Turns out all over the country ...
[NEWS CLIP: Just kept speeding up.
[NEWS CLIP: Even with his foot on the brake.]
SIMON: People were losing control ...
[NEWS CLIP: Plunged over a 100-foot cliff.]
[NEWS CLIP: At more than 90 miles an hour.]
[NEWS CLIP: And he couldn't stop his Toyota.]
[NEWS CLIP: Lexus cars, Toyota cars, Toyota trucks.]
SIMON: ... of their Toyotas.
[NEWS CLIP: This is a serious problem.]
JAD: I remember this. This was a big thing. This was a big news story, yeah.
PHIL KOOPMAN: From the driver point of view, what was happening was that the car would take off. Sometimes at low speeds, but sometimes at high speed. And the drivers would say, "My foot was on the brake pedal but my car kept going faster."
ANNIE: This is Phil Koopman, he’s a professor at Carnegie Mellon University. And co-founder of Edge Case Research.
PHIL KOOPMAN: And my specialty for all these years has been fault-tolerant computing.
ANNIE: And Phil told us that around the same time these accidents started happening ...
PHIL KOOPMAN: In the early-2000s ...
ANNIE: Was the same time that Toyota cars -- and all cars really -- were starting to become computerized.
PHIL KOOPMAN: That’s right.
ANNIE: Before this moment, cars were mostly mechanical, and you could, like, you know, pop the hood and basically figure out what was going on. But after this moment? It’s mostly computers. And not just in the engine.
PHIL KOOPMAN: Your car has dozens, or in some cases, more than a hundred little boxes that have electronics inside it.
ANNIE: How small?
PHIL KOOPMAN: Oh, sort of like small book-sized. Like a paperback book.
ANNIE: Embedded within the wall and floor and ceiling of the car?
PHIL KOOPMAN: Wherever they can find room.
SIMON: They do all sorts of little things.
PHIL KOOPMAN: When you take out your keys and you press the button and your doors unlock, there’s a computer inside that key sending a message to a computer inside the car telling the computer at the lock to open the locks.
ANNIE: Three computers are conversing about that unlocking moment?
PHIL KOOPMAN: At least three.
SIMON: And they're responsible for some pretty important things. Like, now when you press the gas pedal, instead of it being this direct, physical link to the throttle which makes the car go, that pedal is now ...
PHIL KOOPMAN: Basically just an input to a computer.
ANNIE: Why do they want to do that?
PHIL KOOPMAN: Well, you can get better fuel economy. You know, it’s a more pleasant drive. So that part is great. It’s a better driver experience. But, you just trusted your life to software and that’s a new thing.
SIMON: Was that talked about or known? Or was this sort of a silent transition that happened that you as the driver would never really know unless you looked it up in Popular Mechanics or something?
PHIL KOOPMAN: I don’t think that you would probably have known as a regular driver.
SIMON: And so ...
PHIL KOOPMAN: When this whole thing broke ...
SIMON: Because most people didn't know that their cars were now computers, and because most of the victims ...
BARBARA WALTEN: Oh, gosh ...
SIMON: Were on the older side.
BARBARA WALTEN: We might be older, but we’re not stupid people.
SIMON: They were basically blaming the driver for all of these.
ANNIE: Barbara says when they took the car to the Toyota dealership ...
BARBARA WALTEN: I’ll never forget the smirky grin he had on his face, telling us he didn’t find anything wrong with the car.
PHIL KOOPMAN: If you’re on the car company side you say, "So you pressed the pedal and it kept going faster. Which pedal does that?"
GROVER WALTEN: [laughs] I told them I’d been driving since I was 15, and I knew the difference in the accelerator and a brake.
ANNIE: And so these random accelerations ...
PHIL KOOPMAN: Sort of flew under the radar.
ANNIE: Until ...
[CLIP OPERATOR: 911 emergency. What are you reporting?]
PHIL KOOPMAN: August 28, 2009.
[CLIP DRIVER: We're ... [inaudible]]
[CLIP OPERATOR: I’m sorry. Your cell phone's cutting out.]
ANNIE: A 911 call came from a passenger of a Lexus Toyota.
[CLIP DRIVER: We're going North 125.]
[CLIP OPERATOR: Mm-hmm.]
[CLIP DRIVER: And the accelerator is stuck.]
[CLIP OPERATOR: I'm sorry?]
[CLIP DRIVER: My accelerator is stuck. We're on 125 [inaudible]]
[CLIP OPERATOR: Okay, northbound 125. Where are you passing?]
[CLIP DRIVER: We are passing -- where are we passing? We're going 120 in Mission Gorge. We're in trouble. We can't -- there's no brakes.]
[CLIP OPERATOR: Okay.]
[CLIP DRIVER: [inaudible]]
[CLIP OPERATOR: Okay, and you don’t have the ability to, like, turn the vehicle off or anything?]
[CLIP DRIVER: We’re approaching the intersection. We’re approaching the intersection.]
[CLIP OPERATOR: Okay.]
[CLIP DRIVER: We’re approaching the intersection. Hold on, pray, pray. Oh, shoot! Oh!]
ANNIE: Not one of the four people in that car survived.
PHIL KOOPMAN: It’s a 45-year-old male California highway patrol officer, Mark Saylor, whose specialty is vehicle inspector. And this guy can’t stop his car.
ANNIE: And this just broke that driver error narrative ...
[NEWS CLIP: Toyota has a problem.]
ANNIE: To pieces.
PHIL KOOPMAN: And so that’s what launched all the inquiries and the recalls.
[NEWS CLIP: A record $33-million in fines.]
SIMON: Toyota settled a huge $1.6-billion lawsuit.
[NEWS CLIP: No electronic based cars.]
SIMON: Without admitting any blame.
PHIL KOOPMAN: But after that ...
[NEWS CLIP: For at least 2,000 reported cases ...]
[NEWS CLIP: A four-year federal criminal investigation into ...]
ANNIE: With the death count rising to 89 people, a queue of 500 death and injury cases started to form. And when the experts, Phil being one of them, were given full access to Toyota’s software and hardware, they found a number of things wrong. But the key in all of this was, you guessed it.
PHIL KOOPMAN: If you had a bit flip, the task that controls the throttle position could die.
ANNIE: Causing it to stick in some open position.
PHIL KOOPMAN: That would happen. They reproduced that on a car.
SIMON: Well, and in the first of those death and injury cases, what -- what happened?
PHIL KOOPMAN: The jury said that they -- well, I have to make sure I only say things that I’m allowed to say. This is actually a big deal.
ANNIE: Oh, okay.
PHIL KOOPMAN: Yeah, so they -- Toyota went after one of the experts for contempt. So you have to be careful.
SIMON: Wow. Okay.
PHIL KOOPMAN: Sorry. I’m gonna have to look it up.
ANNIE: No, no. Take your time.
PHIL KOOPMAN: So the jury found that Toyota was liable for the death. And the only theory presented in the case was that the computer killed people. And the jury agreed.
ANNIE: Toyota settled that case immediately, and began to settle all 500 other cases in the queue.
SIMON: Rip the Band-Aid off here, pay everybody. Let’s not drag this thing out and put it through the public.
PHIL KOOPMAN: I can’t speculate as to their motives, but I think the facts speak for themselves.
JAD: God, I mean, this is ...
ANNIE: Yeah. And, like, one quick, crazy note. This ruling actually led to the freedom of a Minnesota man named Koua Fong Lee, who had already spent two-and-a-half years in prison, charged with vehicular homicide for killing three people in the Toyota he was driving when it suddenly accelerated out of control.
ANNIE: So this ruling, like, got this guy out of prison. But he was in prison for two-and-a-half years before this, you know -- this acceleration thing.
JAD: That's nuts!
ANNIE: I know.
SIMON: Like, in your opinion and your experience, do you think this specific case and other cases were, in fact, caused by just this crazy, cosmic chain of events that led to a tiny little bit of electricity grazing a circuit board?
PHIL KOOPMAN: As amazing as it sounds, these type of events are almost certainly caused by that. Now, was that primary contribution to the Toyota UA? I don’t know, and probably no one will ever know.
ANNIE: I feel like the way you’re painting the world almost just looks like there’s just smoking wrecks everywhere. Like, crashed planes and cars into light poles and the internet’s not working, and I -- how are we -- how are we functioning? I don’t even understand.
BHARAT BHUVA: So as soon as we found out this is the problem, we went back and we started fixing the problems. Once we know that autopilot is the one that’s causing the problem, then we fixed it.
ANNIE: How did you fix it?
BHARAT BHUVA: So you get three copies of everything. The same electronic system. There are three of them, and at least two of them have to agree to do anything.
ANNIE: Oh, smart.
ANNIE: So today, like, tripling and doubling redundancy is the norm in planes and cars and, like, any other life-saving technology.
SIMON: And actually the same thing is being done in Schaerbeek after that incident in 2003. So now when they have an election, each vote is actually counted twice by the computer. And if one of those tallies doesn’t match the other, they know something's gone wrong immediately.
ANNIE: So that’s cool. But here’s the thing. This doubling and tripling is just too expensive to do with most things. And as our electronics get smaller and smaller ...
BHARAT BHUVA: And smaller and smaller and smaller.
ANNIE: These bit flips are just gonna happen more and more.
ROBERT: Why would that be?
ANNIE: Well, the smaller the transistor, the less energy you need to flip that bit.
BHARAT BHUVA: Now we are at the point where you only need about a thousand electrons to represent a one or a zero.
ANNIE: Wait, how many electrons did you need back in the '70s?
BHARAT BHUVA: Oh, probably a million.
ANNIE: Oh, and now you need a thousand. Wow, okay.
ANNIE: And transistors made today are, like, tiny, tiny.
BHARAT BHUVA: Your hair is 50 microns in diameter. So take your hair and slice it a thousand times. So that gets you down to 50 nanometers. And we are looking at seven nanometers. So you also have to pretty much slice it down seven more times, and what you are left with is the size of transistors that are coming in the market in the next two years.
ANNIE: Will they be in cellphones and laptops and things like that?
BHARAT BHUVA: They will be everywhere.
ROBERT: So as we get better and better and better at miniaturizing our technology, we are more and more and more at risk to these tiny flying projectiles.
SIMON: Right. Through our technological advancement, we have invited this problem upon ourselves.
BHARAT BHUVA: That is one way to look at it, yes. If you go back to those big cell phones, which are weighing, you know, five kg each, and about a foot long, then you wouldn’t have this problem.
JAD: Thanks Annie, Simon.
ANNIE: Actually, we have one more thing for you.
BARBARA WALTEN: Sure.
SIMON: Well, we’ve been doing some research into this and we’ve talked to a bunch of electrical engineers, and this is going to sound very, very strange, but what these researchers are saying a likely cause was, was a cosmic ray hitting one of the computers could cause the driver to lose the ability to control the gas pedal.
BARBARA WALTEN: Oh, my lord! Hmm. Well, I -- I am not a scientific person. Anything’s possible. Now you’re saying we’re bombarded by cosmic rays all the time?
BARBARA WALTEN: If that’s so, perhaps maybe we just -- maybe we just got zapped. I don’t know.
SIMON: Do you think Grover would be -- I’m just curious to hear his reaction to the possibility that your car accelerated because a cosmic ray hit it. What does he think about that?
BARBARA WALTEN: There is one theory -- you said an engineer came up with that theory?
SIMON: A bunch of electrical engineers. Very smart people.
BARBARA WALTEN: A bunch of electrical engineers have come up with a theory, you know, we’re bombarded with cosmic rays all the time.
GROVER WALTEN: With what?
BARBARA WALTEN: With cosmic rays all the time. And it’s possible that a cosmic ray, if it hits just at the right time in the right place, it can zap onboard computers and possibly zap the accelerator control. He wants to know how you feel about that? [laughs]
GROVER WALTEN: I’m somewhat open-minded, but at this point in time I, you know, I question that. But it’s not that big a deal. Okay. Well it’s -- it’s home and white wine time.
ROBERT: Well, we couldn't agree more.
ROBERT: And so what we've decided to do is provide you with a small entertainment for you to sip your white wine with.
JAD: Simon has made a video to accompany this episode that must be experienced.
JAD: It is at Radiolab.org. Watch it.
ROBERT: Very short and very, uh, weird.
JAD: Yeah. This piece was produced and reported by Simon Adler and Annie McEwen.
ROBERT: Special thanks to Lucas DeRaker and to Vadric DeBooker and to Francesca Costanzo, Kelly Gallagher, Natalie Higgins and the Hailstone Ice Company, in Crown Heights, Brooklyn.
JAD: Okay. You wanna go now?
ROBERT: Yeah, I think so. I think I'm gonna go now.
JAD: Well, it's not going forever. Just for the moment.
ROBERT: Yeah, just for the -- I feel reluctant to leave the room.
JAD: I felt that in your voice.
ROBERT: Yes, I know.
JAD: I'm Jad Abumrad.
ROBERT: And I'm just gonna stay here forever.
JAD: Okay. Well, you're -- yeah, he's Robert. And we'll see you in the next one. [laughs]. This is so much fun, so much fun!
[STAN WYNN: My name is Stan Wynn. I'm calling from Cleveland, Ohio. Radiolab was created by Jad Abumrad and is produced by Soren Wheeler. Dylan Keefe is our director of sound design. Suzie Lechtenberg is our executive producer. Our staff includes Simon Adler, Becca Bressler, Rachael Cusick, David Gebel, Bethel Habte, Tracie Hunte, Nora Keller, Matt Kielty, Robert Krulwich, Annie McEwen, Latif Nasser, Malissa O'Donnell, Sarah Qari, Arianne Wack, Pat Walters and Molly Webster. With help from Shima Oliaee, Audrey Quinn and Neil Danisha. Our fact-checker is Michelle Harris.]