Latest / Elliot In The Morning / EITM: Missy Cummings 11/11/26
Transcript
- 0:00Speaker 1: Missy Cummings, please. Hello. Hi, how are you? I'm good.
- 0:06Speaker 2: How are you? Good.
- 0:07Speaker 1: Happy Veterans Day.
- 0:10Speaker 2: Thank you so much for thinking of me.
- 0:12Speaker 1: Absolutely. Absolutely. Hey, so it was weird.
- 0:15Speaker 3: I was just reading this story about the Waymo cars,
- 0:18Speaker 3: and they were talking about how these two professors were
- 0:22Speaker 3: kind of like leading the country in studying them.
- 0:26Speaker 1: And I was like, damn, she works at Mason. She's
- 0:28Speaker 1: right down the street.
- 0:30Speaker 2: That's correct.
- 0:32Speaker 1: And I apologize ahead of time for the voicemail. I
- 0:34Speaker 1: left you.
- 0:40Speaker 3: Hey, so before, before we get to Mason, can I
- 0:42Speaker 3: back up before that? I, and I apologize. I didn't
- 0:44Speaker 3: know what your history was. I just wanted to talk
- 0:46Speaker 3: to you about the autonomous cars.
- 0:48Speaker 1: Dude, that I'm sorry, ma'am.
- 0:51Speaker 3: That is, that's a pretty, that's a pretty amazing first career.
- 0:56Speaker 1: First part of your career. Yeah, yeah.
- 1:01Speaker 2: And it carries on into the latter part of my life.
- 1:06Speaker 2: I mean, you know, people will tell you I just
- 1:08Speaker 2: love to fight about everything.
- 1:11Speaker 1: Well, you came to the right place.
- 1:13Speaker 3: How did you end up being one of the first
- 1:16Speaker 3: female fighter pilots for the Navy?
- 1:20Speaker 2: Well.
- 1:22Speaker 2: When the combat exclusion law was repealed in 1993, I
- 1:28Speaker 2: had already been an adversary pilot. So I don't know
- 1:33Speaker 2: if you remember Top Gun 1, where the little plane
- 1:37Speaker 2: that Jester flew was So I had been flying A4
- 1:41Speaker 2: in a fighter pilot role, but in a training role.
- 1:45Speaker 2: So there were a group of us who became the
- 1:47Speaker 2: first group of female fighter pilots. And I was, you know,
- 1:51Speaker 2: right at the top of that list because I'd already
- 1:53Speaker 2: been doing the mission.
- 1:54Speaker 1: Oh, wow. That's awesome.
- 1:57Speaker 3: I said earlier on the air, this will not impress
- 1:59Speaker 3: you at all, but I also have been in a
- 2:01Speaker 3: fighter jet.
- 2:02Speaker 1: I was fortunate at one point to fly with the
- 2:04Speaker 1: Blue Angels.
- 2:08Speaker 2: And did you throw up?
- 2:09Speaker 1: No, actually, I didn't. I managed to keep it together.
- 2:12Speaker 1: I'll tell you, though, it was... Like, that's an amazing,
- 2:16Speaker 1: amazing experience, though.
- 2:20Speaker 2: Yeah, I think that's good. And it's... I think one
- 2:25Speaker 2: of the problems, though, with the future of technology is
- 2:28Speaker 2: it's really fun to do jobs like that, but they're
- 2:32Speaker 2: also very dangerous. So it is true that... there is
- 2:38Speaker 2: going to be an end at some point to humans
- 2:41Speaker 2: as fighter pilots because of technology.
- 2:45Speaker 1: So you think that it'll change front? By the way,
- 2:48Speaker 1: can I ask you a personal question?
- 2:49Speaker 3: Sure.
- 2:51Speaker 1: What was your call sign?
- 2:54Speaker 2: Shrew. Shrew?
- 2:55Speaker 1: That's not nice.
- 3:00Speaker 2: I loved it.
- 3:01Speaker 3: Okay, good.
- 3:02Speaker 1: As long as you like it.
- 3:05Speaker 3: So wait, so you believe that because of how dangerous
- 3:09Speaker 3: it is and the risk to human life, that fighter
- 3:16Speaker 3: pilots and fighter jets will completely be.
- 3:20Speaker 1: Gone and it'll be all drone warfare? Yes.
- 3:25Speaker 2: You know, I mean, the thing is, is that with
- 3:26Speaker 2: the rise of You know, this isn't even, we're not
- 3:29Speaker 2: even talking about things like generative AI. This has been
- 3:32Speaker 2: long in coming. And even people inside the Pentagon will
- 3:36Speaker 2: say the last fighter pilot has been born. You just can't,
- 3:41Speaker 2: humans cannot compete with what technology can do in the aircraft. Meaning,
- 3:48Speaker 2: you know, the Blue Angels are great and they do
- 3:51Speaker 2: an amazing job, but their bodies really cannot exceed 8G.
- 3:56Speaker 2: and they need sleep, and they have to go to
- 3:58Speaker 2: the bathroom. Autonomy, when you put it in a jet,
- 4:02Speaker 2: it can do 24-7. It doesn't make mistakes around the
- 4:06Speaker 2: carrier when you come aboard at night. It doesn't suffer
- 4:10Speaker 2: from fatigue, and it can pull 20 Gs, for example.
- 4:15Speaker 2: So I think that there's some real performance issues that
- 4:20Speaker 2: autonomy can really be better than us. That's not to
- 4:23Speaker 2: say that humans will be replaced altogether in war, but
- 4:28Speaker 2: they'll just be remote operators.
- 4:31Speaker 3: So do you get, do you get mad at people
- 4:33Speaker 3: like me? Because I, I obviously you're right.
- 4:35Speaker 1: Like, you know what I mean? Like, I can't argue
- 4:37Speaker 1: any of that.
- 4:38Speaker 3: I just, I just, I, I just, I, I hate
- 4:41Speaker 3: that we keep losing humans in jobs.
- 4:45Speaker 2: Well, you know, honestly, I mean, I'm glad I did it.
- 4:49Speaker 2: I'm glad I was one of those lucky few that
- 4:52Speaker 2: actually got to transition from old aircraft to new aircraft.
- 4:56Speaker 2: But it's going to be a better form of warfare
- 5:00Speaker 2: if you could call that better. I mean, less mistakes
- 5:03Speaker 2: will be made, more performance will be had. But in
- 5:09Speaker 2: all cases, humans don't actually get replaced. They just get
- 5:14Speaker 2: moved to a different place. And in fact, this whole
- 5:17Speaker 2: thing with Waymo, I think it's a perfect illustration that
- 5:21Speaker 2: What you don't know is that there are warehouses of
- 5:25Speaker 2: people that supervise these self-driving cars. There is no such
- 5:30Speaker 2: thing as a self-driving car anywhere in the world. They
- 5:33Speaker 2: all have teams of people who supervise their every move
- 5:38Speaker 2: and who jump in when the cars make mistakes.
- 5:42Speaker 3: Well, where are those people when I hit it and
- 5:44Speaker 3: then I just take off and flee the scene?
- 5:48Speaker 2: Well, it turns out one of Waymo's issues is that
- 5:52Speaker 2: they have put these people in the Philippines because they
- 5:57Speaker 2: are cheaper and they are out of the reach of regulators.
- 6:02Speaker 2: And if you hit a Waymo, it actually takes time
- 6:07Speaker 2: for the operators, because they're halfway around the world, so
- 6:12Speaker 2: they don't have the ability to immediately respond. And they
- 6:17Speaker 2: don't have the ability to respond unless you're in the car.
- 6:22Speaker 2: So they don't have any speakers on the car that
- 6:24Speaker 2: can talk to you. They don't have any way for
- 6:27Speaker 2: them to interact. So they can move the car, but
- 6:32Speaker 2: they can't tell you that they're doing anything.
- 6:37Speaker 3: So in the article that I was reading, and I
- 6:40Speaker 3: think the study was done in Phoenix and the surrounding
- 6:43Speaker 3: areas like Mesa and Gilbert and stuff like that. is
- 6:48Speaker 3: they were talking to you about how all these Waymos
- 6:52Speaker 3: are getting into accidents where people are hitting them, they
- 6:56Speaker 3: hit somebody or whatever, and it seems like it's twofold
- 6:59Speaker 3: of a human driver's reaction. Number one, they get out
- 7:02Speaker 3: of the car because they want to exchange information, they
- 7:04Speaker 3: want to do the right thing, and they walk up
- 7:06Speaker 3: to the car and they go, oh, well, there's nobody there.
- 7:09Speaker 1: There's nothing to do. I don't know what to do.
- 7:11Speaker 3: There's no representative. I'm not talking to anybody. So they
- 7:14Speaker 3: just get in their car and they leave because they
- 7:15Speaker 3: don't know what to do. And then there's the other people.
- 7:18Speaker 3: And I get it where they walk up and they go, oh,
- 7:21Speaker 3: nobody's in here.
- 7:22Speaker 1: See you later. I'm taking off. Yeah. So they end
- 7:26Speaker 1: up going.
- 7:28Speaker 2: Right. I mean, that is human nature on both counts.
- 7:31Speaker 2: You know, I think what is one supposed to do
- 7:34Speaker 2: when you're faced with a robot? And I'm It's not
- 7:40Speaker 2: a non-trivial chance that the car could take some action
- 7:45Speaker 2: after it has had an accident. We have seen that.
- 7:49Speaker 2: It wasn't a Waymo car, but there was a cruise
- 7:52Speaker 2: car that hit a pedestrian and then stopped. And then
- 8:00Speaker 2: kept moving. And then the pedestrian got drug underneath the car.
- 8:05Speaker 2: She lived, but she was seriously injured.
- 8:07Speaker 1: Right.
- 8:08Speaker 2: And these are the cases where the cars, and this
- 8:12Speaker 2: is true of all AI. And if your listeners, if
- 8:15Speaker 2: they don't remember one thing, they should remember that AI
- 8:18Speaker 2: does not think. It doesn't know. It doesn't understand. It
- 8:22Speaker 2: doesn't imagine things. So when a car is in an accident,
- 8:27Speaker 2: other than knowing that its accelerometers have indicated that there's
- 8:30Speaker 2: been an accident, it doesn't know anything else. And so
- 8:36Speaker 2: they can be highly unpredictable in these cases because they
- 8:39Speaker 2: don't know that you're behind them and that you are
- 8:43Speaker 2: trying to exchange information. You know, they know they've been
- 8:48Speaker 2: in an accident before. They may call back to the
- 8:51Speaker 2: remote operator. But if that remote operator is working on
- 8:55Speaker 2: another car in the Philippines, then they're going to have
- 8:57Speaker 2: to wait in a queue. And we know that this happens.
- 9:00Speaker 2: In fact, in San Francisco last year, there was a
- 9:04Speaker 2: power outage and all the cars freaked out. And then
- 9:08Speaker 2: they all just stopped and tried to call the remote operators.
- 9:12Speaker 2: And they didn't have enough remote operators to fix the problem.
- 9:17Speaker 1: So what would you tell somebody to do?
- 9:20Speaker 3: If you're in an accident with a Waymo, what are
- 9:25Speaker 3: you supposed to do?
- 9:26Speaker 1: Just sit there and wait until the Philippines picks up? No,
- 9:30Speaker 1: I wouldn't.
- 9:30Speaker 2: I mean, if it happened to me.
- 9:31Speaker 1: No, Shrew would flee the scene.
- 9:37Speaker 2: Well, I wouldn't approach the car at all. No way.
- 9:40Speaker 2: I would take a picture. And there's a there should
- 9:44Speaker 2: be some kind of phone number, you know, depending on
- 9:47Speaker 2: I'm not sure what they've said that they're going to
- 9:49Speaker 2: do for D.C., but you can reach them. I would
- 9:51Speaker 2: just take a picture and reach out and then try
- 9:54Speaker 2: to solve the problem.
- 9:56Speaker 3: And the reason you wouldn't approach the car goes back
- 9:58Speaker 3: to what you said.
- 9:59Speaker 1: That car doesn't know anything. It's just going to go.
- 10:03Speaker 1: It might.
- 10:03Speaker 2: And it has, you know, I mean, so it's unpredictable.
- 10:07Speaker 2: Don't approach the car. Take a picture. You don't do
- 10:10Speaker 2: the right thing. Try to contact them. But I think
- 10:13Speaker 2: one of the other things that Waymo is not being
- 10:16Speaker 2: truthful about is and I've written several papers about this.
- 10:20Speaker 2: The cars have a very serious tendency to do something
- 10:24Speaker 2: called phantom braking. And when that happens, the car sees nothing.
- 10:32Speaker 2: It's hallucinating. It's really hallucinating. And it slams on the
- 10:37Speaker 2: brakes really hard and really fast. And even if you're
- 10:40Speaker 2: behind it following at the prescribed two-second distance, you will
- 10:44Speaker 2: still hit it because robot cars can brake harder and
- 10:48Speaker 2: faster than you can. And so it is true that
- 10:53Speaker 2: people are rear-ending Waymos, but they're rear-ending Waymos at two
- 10:59Speaker 2: times the rate that normal cars would get rear-ended. And
- 11:03Speaker 2: so we know that there is a problem. This failure
- 11:07Speaker 2: mode is likely at least seriously exacerbated by the behavior
- 11:12Speaker 2: of the self-driving car. And this is actually one of
- 11:16Speaker 2: the reasons Waymo is not keen to go after people
- 11:20Speaker 2: for it, because they know that they're partly at fault
- 11:23Speaker 2: as well, and they don't want to get into a
- 11:25Speaker 2: legal tussle where they would have to prove whether they
- 11:29Speaker 2: were or were not at fault.
- 11:31Speaker 1: God, Waymo, I can't tell if Waymo loves you or
- 11:33Speaker 1: hates you. I love you.
- 11:35Speaker 2: Oh, that's me. If I were to mysteriously die, first
- 11:39Speaker 2: person who did it, Elon Musk.
- 11:41Speaker 3: That's fascinating.
- 11:42Speaker 2: Hey, one other question for you real quick.
- 11:44Speaker 1: How did you end up at Mason?
- 11:57Speaker 2: Well, I actually came to DC in 2022 to work
- 12:02Speaker 2: with the Biden administration at their invitation to help them
- 12:05Speaker 2: develop self-driving car policy. And so I was here for
- 12:09Speaker 2: a year and a temporary basis. And I just loved
- 12:12Speaker 2: it so much, I decided to stay. Oh, wow.
- 12:14Speaker 1: Good for you. Good for you. All right. Last thing.
- 12:18Speaker 1: Do I call you doctor? I don't know what to
- 12:19Speaker 1: call you. Lieutenant?
- 12:20Speaker 2: Missy's fine.
- 12:21Speaker 1: I'll just go with Missy. That's fine. Hey, can I
- 12:24Speaker 1: hold on to your number? I already remembered the last
- 12:27Speaker 1: four digits.
- 12:28Speaker 3: But can I hold on to your number in case
- 12:30Speaker 3: I need a autonomous vehicle expert?
- 12:32Speaker 2: 100%. You're going to need to call me a lot
- 12:36Speaker 2: more in the next year.
- 12:38Speaker 1: Oh, they're coming.
- 12:39Speaker 3: All right.
- 12:40Speaker 1: Very good. Very good. Missy, hold on one second. Let
- 12:44Speaker 1: me get your shirt. You don't care about that.
- 12:45Speaker 3: All right.
- 12:46Speaker 1: Hold on one second. Hold on one second, please. How
- 12:48Speaker 1: awesome is that? Yeah. Great insight.
- 12:50Speaker 2: I know.
- 12:51Speaker 1: By the way, you know what my new bid is?
- 12:54Speaker 1: What is it? No matter what accident I'm in, thought
- 12:57Speaker 1: it was a Waymo. I took a picture and I left.
- 13:00Speaker 1: There's three people in that car.
- 13:01Speaker 2: Passengers.
- 13:02Speaker 1: I don't know. Somebody behind the wheel. Short driver. Couldn't
- 13:06Speaker 1: see him. I thought somebody just got in the seat.
- 13:08Speaker 1: I hit a Waymo. Maybe that guy who lives in
- 13:10Speaker 1: the Bean. Tried to move it. Couldn't get it done.
- 13:14Speaker 1: I love her, though. I love her. She's great. It
- 13:19Speaker 1: was educational, but it was also entertaining. What was the
- 13:23Speaker 1: one thing that I was supposed to remember? AI doesn't think,
- 13:27Speaker 1: doesn't feel. Doesn't have emotions. By the way, I've said
- 13:30Speaker 1: that in the past. AI can't hug you. I don't
- 13:34Speaker 1: know if you've said it so bluntly. But you've tried
- 13:37Speaker 1: to suggest that it's a reason that humans will always
- 13:40Speaker 1: be better. Thank you.
- 13:42Speaker 2: Thank you.
- 13:42Speaker 1: Of which I am one.