0:00

Here on Gadget Lab, we dive deep

0:02

into the tech universe, tackling questions like,

0:04

is giving companies access to your genetic material

0:07

a good idea? And are

0:09

the latest phone releases really that different than the

0:11

last ones? We want to

0:13

help you make informed decisions about what

0:16

is worth your attention. And here's something

0:18

that is undeniably worth your time, a

0:20

digital subscription to Wired. Lucky

0:22

for you, we are giving Gadget

0:24

Lab listeners an exclusive discount, 20%

0:27

off an annual subscription to Wired. Just

0:30

visit wired.com and use the promo code

0:32

GL20 to get 20%

0:34

off a digital subscription. Use

0:37

GL20 to get exclusive access

0:39

to stories on the latest

0:41

innovations like AI, deepfakes,

0:43

and VR, as well

0:45

as today's most talked about people in technology.

0:50

This episode is brought to you by

0:52

Progressive. Are you driving your car or

0:54

doing laundry right now? Podcasts go best

0:56

when they're bundled with another activity. Like

0:59

Progressive Home and Auto Policies, they're best

1:01

when they're bundled too. Having these two

1:03

policies together makes insurance easier and could

1:05

help you save. Customers who save by

1:07

switching their home and car insurance to

1:09

Progressive save nearly $800 on

1:12

average. Quote a home and

1:14

car bundled today at progressive.com.

1:16

Progressive Casualty Insurance Company and

1:18

Affiliates. National average 12 month

1:20

savings of $793 by

1:22

new customers surveyed who saved with Progressive between June

1:24

2021 and May 2022.

1:28

Potential savings will vary.

1:30

AI is making waves in every

1:32

field it touches. President Biden is now

1:34

on TikTok and the election draws closer

1:37

each day. With so much going

1:39

on in the world, it is hard to keep up

1:41

with it all, let me tell you. Hi, I'm Kai

1:43

Rizdali, co-host of Make Me Smart. It's a podcast from

1:45

Marketplace. And every weekday, Kimberly Adams

1:48

and I break down the latest in business

1:50

and the economy with short daily episodes to

1:52

make it easy for you to stay in

1:54

the know. Listen to Make Me Smart wherever

1:56

you get your podcasts. Mike.

2:01

Lauren. How long have you been

2:03

covering the Internet for now as

2:05

a journalist? Oh over 20 years. How

2:08

long have you been a wire? Over 20

2:10

years. Basically since the earliest

2:12

days of the consumer Internet. Yes

2:14

I've been online since I was a

2:16

pre-teen sort of like that

2:19

character in Almost Famous following around the band while

2:21

I was a youngster. Don't

2:23

make friends with the rock stars. Yes or the nerds. How

2:26

often would you say you still get tripped

2:28

up on internet terms and protocols and acronyms?

2:30

Oh a lot. I mean it's never-ending especially

2:33

now with AI which I don't really follow

2:35

as a journalist and even if I did

2:37

I don't think I could keep track of

2:39

all the acronyms. I agree and

2:42

I really think we need to demystify all

2:45

this for people and what better way to

2:47

do it than to make ourselves look like idiots? Oh

2:49

yeah okay sure. I think I can think of better

2:51

ways. Let's do it. Hi

3:01

everyone. Welcome to Gadget Lab. I'm Lauren Goode.

3:03

I'm a senior writer at Wired. And

3:05

I'm Michael Kalori. I'm Wired's director of

3:07

consumer tech and culture. We're also joined

3:10

this week by senior writer Will Knight

3:12

who covers AI for us

3:14

at Wired. Hi Will. Hello

3:16

there. It's great to have you back on. So

3:19

we're doing something a little bit different

3:21

today. We are turning the Gadget Lab

3:23

into a quiz show but before you

3:25

all turn to another podcast seriously

3:28

stay with us because we're going

3:30

to attempt to define and explain

3:33

all of the acronyms of the earliest

3:35

and the most current consumer internet. Stuff

3:37

that you're hearing about or maybe you

3:39

even say all the time like DARPA

3:42

and PCPIP and

3:44

SMS and LLM.

3:48

How many times? Do a shot every

3:50

time you hear LLM on a Wired

3:52

podcast these days. Now this was partly

3:54

inspired by an early 2000s book that

3:57

I happened to be reading recently. It's

3:59

called dot-con by the New Yorker

4:01

writer John Cassidy. And despite the fact that

4:03

I know a little bit about the early

4:05

internet, I mean, like Mike, I've been online

4:07

since the mid-90s, I

4:10

was actually floored by how many terms I

4:12

didn't know. And it just

4:14

got me thinking, we should do a podcast on

4:16

this. Like let's break this down into three parts,

4:19

the early internet, the mobile era,

4:21

and now the era of AI.

4:24

And we've brought in Will because he is our

4:26

expert AI reporter. So first,

4:28

Will, you and I are going to quiz

4:30

Mike because he's the guy who's

4:32

already established that he's been around like forever on

4:35

the dinosaur age of the internet. Thank you. You're

4:37

welcome. And we're gonna quiz

4:39

you on all the, you have not seen these

4:41

in advance. Nope. Will and I have a shared

4:43

dock with each other. We've

4:46

crafted some ideas that we think, like

4:48

some you're going to get easily. And some

4:51

I think might stump you. Okay, I look

4:53

forward to this. I'm not allowed to look them up while you're

4:55

asking, right? No, you were not allowed to look them

4:57

up. None of us were. And

4:59

the way we're gonna break it down is after this,

5:01

you and Will are going to quiz me on the

5:03

era of kind of the mobile internet because that's when

5:05

I started covering tech. And then finally, we're gonna end

5:07

with Will on AI. Will,

5:09

would you like to go first in quizzing Mike

5:12

on the wonky acronyms of the early internet? Sure,

5:14

I would be delighted. I

5:17

wanna just

5:20

to preface it by saying, I

5:23

am terrible at remembering acronyms in the best

5:25

of times and there are a billion out

5:27

there in AI. So I'm gonna do terribly,

5:30

but the first one- Look at this copy

5:32

adding already. The

5:34

first one on our list is,

5:37

it's not DARPA, it's ARPA, but

5:39

they're related. ARPA, oh

5:42

boy, kicking it off with a

5:44

bang here. It

5:46

predates the consumer internet. See DARPA, the

5:49

D is defense because

5:51

the internet started as a

5:53

defense project, US government defense

5:55

project, right? I

6:00

can't remember what comes after Defense in DARPA,

6:02

so I can't remember what DARPA is. Research

6:06

project something, something? Yeah, no, you're getting it. Is

6:08

that right? Yeah, yeah. I

6:11

mean, if you think of the

6:13

letter A and what the internet was at the

6:15

time, it was pretty like... Autonomous?

6:17

No. I don't know. I

6:20

don't know. It has something to do with... It was

6:23

the Defense Amazing research

6:26

project. American? No.

6:28

No, it actually didn't start in America. Okay. Did

6:31

it? I don't... It was

6:33

like the earliest internet did not. I give up. What does

6:35

ARPA stand for? Will, would you like

6:37

to tell them? It's Advanced Research Projects Agency.

6:41

Advanced Research Projects Agency, okay. So

6:44

was DARPA the Defense Advanced

6:46

Research Projects Agency? Yeah. Yes.

6:48

That's right. Mm-hmm. This

6:51

one I think you're going to get. Okay. BBS.

6:53

BBS, Bulletin Board System. Yes, BBSs.

6:56

Good job. So a

6:58

bulletin board system was like a

7:00

computer in a basement

7:02

somewhere or under someone's

7:04

bed or in the closet in their mom's

7:07

bedroom where they

7:09

hosted a message board. Mm-hmm.

7:13

And you would call on your modem to

7:15

the message board and leave messages

7:17

and then hang up and go about your

7:19

day and then hours later go back and

7:22

read people's replies. It was a community forum.

7:24

Mm-hmm. Do you know where it

7:26

started? This is a bonus question. Where

7:28

BBSs started? Mm-hmm. I

7:30

would just assume they started in suburban North America. Close.

7:33

Well, it was Chicago in the 1970s

7:35

during a blizzard. Oh, nice. Yeah. Two

7:38

guys kind of patched it together. Nice. Uh-huh.

7:41

Okay. Wow. I was

7:43

an avid bulletin board boarder, I guess. Yeah. When

7:45

I was quite young. Yeah. Yeah.

7:47

I was not. Under it.

7:50

It's all there was. Are they still around? Yes.

7:53

I'm sure they are. I mean, it's basically Reddit, right?

7:55

Mm-hmm. It's all that was around. Like, you... There was

7:57

not a whole lot you could do on the internet.

7:59

rather than go to BBSs and chat rooms. Right,

8:02

and Google, I think, acquired

8:04

it. Am I remembering that correctly? What was

8:06

the- They acquired the

8:08

big users net. Oh, Usenet,

8:11

correct. They acquired Usenet. Which

8:13

is- Usenet was started, yes. BBS was started

8:15

by two guys in Chicago in the 70s. Usenet

8:18

was actually started at Duke University, I

8:20

believe. Yes. Yeah. Okay,

8:23

fun one. Similar times. Similar

8:26

times, indeed. Mike,

8:28

what about ABR? Mm-hmm.

8:31

ABR. I didn't know this one. Will put this on

8:34

there. I don't know what that means. Did I?

8:36

I think you did well. I can't, I

8:38

don't know what that means. Did I dream

8:40

it up? I

8:43

don't know what ABR is. Available bit rate. Oh,

8:46

available bit rate, okay. Okay,

8:48

so that's like when you call a server, the

8:50

maximum BOD, you

8:53

can, the maximum BBS, BOD, you

8:56

can connect that. So like 1200, 2400, 9600. You're

9:03

just staring at me. I don't know.

9:05

How about you go on. Okay,

9:08

these are good. We

9:10

don't use BOD enough. I think of- We

9:13

don't. It's a great term. I don't

9:15

know what it means, but it's a great term.

9:17

BAUD, it's BPS, basically. It's the same bits per

9:19

second. Right, yeah. Okay. Makes me think

9:21

of one of those modems going, making

9:23

that noise. Yes, that's exactly

9:25

what BOD is. Next

9:29

one, TCP, stroke IP. TCP

9:32

stroke IP, otherwise known to Americans

9:34

as TCP slash IP. Oh,

9:37

sorry, yeah. Okay. I'd

9:39

never heard that before. That's how they say slash

9:42

in Britishese. Okay,

9:46

all right then. Transfer,

9:50

transfer content protocol, internet

9:53

protocol. Mm,

9:55

close. Did I miss the C? Pretty close.

9:58

No, you had control of that. I think just had to all jump. out

10:00

there. Transmission control

10:02

protocol, internet protocol. Transmission

10:04

control protocol, internet protocol. TCPIP

10:07

is what, it's packets on

10:10

the internet. We still use TCPIP today,

10:12

right? Yes. Correct.

10:15

Internet traffic is TCPIP traffic. I just realized

10:17

I'm not counting how many you're getting, but

10:19

I think... I think I've gotten two. Two

10:21

out of four. Okay. Okay.

10:24

That's stakes and catalacts in the big leagues. I admittedly

10:26

didn't know that this was an acronym.

10:29

Basic. Basic. Mm-hmm.

10:32

Of the programming language. Oh, I don't know what it stands for. Beginners

10:36

all-purpose symbolic instruction code.

10:40

That sounds utilitarian and correct.

10:46

All right, two out of five. Okay. How

10:48

many are we doing? Okay. I

10:50

don't know how many we're doing actually. This is all very organized. GUI.

10:55

GUI? GUI. graphical

10:57

user interface. GUI. Correct.

11:01

Yeah. Correct. So,

11:03

like, that was a new thing

11:05

when we moved away from text-based

11:07

interfaces and we got a mouse

11:09

and a pointer and icons and

11:11

a desktop. That's a GUI. Yeah.

11:15

Yeah. I remember the birth of GUIs.

11:17

That's correct. Sweet. This

11:19

one comes directly from the book that I was reading. It's

11:22

a little unusual. Okay. I

11:24

would say it's not an internet protocol. Okay.

11:27

It's very related to the early 2000s internet. Okay.

11:32

PCLN. PCLN. Not

11:35

an acronym, but it stands for something. Just

11:38

breaking all the rules here.

11:46

I don't know. It's a company. It's a

11:48

company. PCLN. Okay.

11:51

What does this company make personal computers?

11:54

This company was one of

11:56

the prime examples of boom

11:58

and bust. boom and bust.

12:01

Is it Petco? No. Oh, is

12:03

it a stock ticker? Mm-hmm. Ah,

12:06

okay. It's a stock ticker for PCLN.

12:10

Um, uh, is it, it's

12:12

not, it's not Compaq. It's

12:14

not Gateway. It starts with P. Sorts

12:16

of P. I

12:20

don't know. Priceline. Priceline. Mm-hmm. Never

12:22

would have gotten that. I know. That's

12:24

kind of a tough one. Yeah, that is a tough one. Will,

12:26

would you have gotten that? No,

12:29

not at all. What is Priceline?

12:32

Priceline was a travel website.

12:34

Yeah. Oh, is that the thing with

12:36

William Shatner? With William Shatner? I should have

12:38

given you that. He built it, I believe. Yes.

12:40

Yeah. Definitely was not just

12:43

paid loads of money to promote

12:45

it. He built it. Okay, give

12:47

me one more because I'm deeply uncomfortable and I

12:49

want to end this on a pretty good value.

12:51

Oh, Will, do you want to do this next

12:53

one? It's going to make you deeply uncomfortable. Okay. Okay.

12:57

A slash S slash L.

13:00

Okay. A stroke S

13:02

stroke L. Is that what

13:04

you're saying? Yeah, no problem.

13:06

Okay. That's A, that's age,

13:08

sex, location. Is that

13:11

right? That is correct. Yeah. All right. So

13:13

like if you were chatting with somebody and

13:16

you would usually say A slash

13:18

S slash L, question mark, because you

13:20

wanted to know their age, their sex,

13:22

and their location. Sex,

13:24

of course, meaning gender. And

13:27

that was like an AOL kind

13:29

of chat thing that I

13:31

did not really participate in as much.

13:35

But I'm aware

13:37

of that because I was

13:39

editing Wired Stories where people

13:41

were referencing that acronym. That's

13:43

all I've been doing. That's

13:45

pretty great. That did make me deeply uncomfortable.

13:47

Thank you. Brought back memories, huh?

13:49

Yeah. It was just for finding friends.

13:54

Sure. Yeah. Just like the rest of the internet.

13:56

Just like the rest of the internet. Okay. So

13:58

is this where I get to start asking? asking

14:00

you questions and Will and I. Yeah,

14:02

how did you score? I really wasn't keeping track.

14:04

I think you got four out of? I

14:06

think I got all of them right. Let's go

14:09

with that. Three, four, five, six,

14:11

seven, eight, nine. We gave you eight and I

14:13

think you got about half. Okay. I

14:15

don't know what the prize is, but congratulations. The

14:17

prize is bragging, right? So as always. Okay.

14:20

All right. Okay, Will, we're gonna

14:22

quiz Lauren now. So we're moving on to

14:24

the mobile era, which I guess is like

14:27

roughly the turn of the century to about

14:29

two years ago or like 2005, six. I

14:34

would classify it as 2007 when the iPhone was launched

14:37

and the app store the following year, 2008 onward.

14:41

Okay. Yeah. Okay, we'll

14:43

do that. Aren't we still in the mobile era? No,

14:46

we're in the metaverse era. Yes, we're, it

14:48

was gonna full meta. All

14:51

right, well, okay. Okay, so here we

14:54

go for the mobile era. Okay, I want to, I

14:57

don't know if this, how big this

14:59

was in America, but WAP, W-A-P,

15:02

do you know this? Not

15:05

the song, Lauren. Not

15:08

the song. All

15:11

right, okay. No, I'm

15:14

gonna make this a wireless access

15:16

protocol. That's

15:19

like close enough. Very, very close. Okay,

15:21

what is it? What is application protocol?

15:23

Oh. This is where they were

15:25

like, we're gonna reinvent the

15:28

web for the mobile era and

15:30

it'll be really terrible and

15:32

clung in and

15:34

just little pixel-explanated websites. It

15:37

was kind of a thing on Nokia phones in Europe for

15:40

about a year. We had that

15:42

here too. Okay. Yeah, we did.

15:44

Yeah, the first mobile browsers were WAP browsers.

15:47

I don't think I realized they were called that. Yeah, Safari, I

15:49

think, was the first mobile browser that was an

15:51

actual browser. Huh. Yeah.

15:54

One note about this era, I think that

15:56

if I'm gonna get something confused, quite a bit, it's service and

15:58

system. It's a lot of these. Actors can

16:01

refer to either mobile quite.

16:03

A lot of s is that me neither

16:05

of the isn't This List member re okay

16:07

okay second one. M V

16:09

N O. O

16:11

o o on. Mobile

16:15

Video network operator. Close

16:17

Mobile. Mobile.

16:22

It's not video, Is it?

16:25

It's mobile. I'm

16:31

just gonna make up. And then mobile

16:34

sector know. Where people

16:36

use. Virtual Virtual little boy.

16:38

Virtual? Yes, No. or you remember

16:40

that. So what is it? Can

16:43

you define it? No, I don't

16:45

remember what it is. It's like

16:47

when a company leases spectrum. From.

16:50

Somebody who own the High A network.

16:52

Of yes it's all coming back in his

16:54

early mobile. yeah this is an i was

16:57

thinking up air and and okay well at

16:59

this is fun. The Okay:

17:01

next on sim. Or. As

17:03

I am. Oh. My. God. It's

17:05

the desert so yard hi I'm.

17:07

Lisa think I'm going to.

17:10

Quit. Lama hadn't just quit

17:12

Now a fan Will He

17:14

isn't as electronics them but

17:16

Sim is. In

17:23

a hockey thinking of right now know all this is

17:25

dead or bond. Life.

17:30

On this is that so Sim is is

17:33

a really hard one because but I'd never

17:35

would have guessed that at Stanford. what is

17:37

everyone. Okay so it is the

17:39

thing that gives you connectivity on

17:41

your mobile device. So I'm gonna

17:44

go with satellite. Is

17:46

that correct? Okay, I'm. System.

17:49

Know. Okay simulated

17:51

know I'm will tell

17:53

you it's very as

17:55

a direct. And and

17:58

pace. I

18:01

don't know. I don't know if

18:03

the subscriber identity module. Oh

18:05

come on, yeah you. Know

18:08

if we're done is very weird. Oh

18:10

wow oh the poor I mean it puts it probably

18:12

is a simulated I don't see more. do some where.

18:16

He loves thinking about one. Meal

18:20

is okay. Ah well. I'm I

18:22

think I'm over three or four.

18:25

Will. Get some good ones. Okay, okay,

18:27

I'm. But. It or of

18:29

a softball sms. Sort.

18:32

Messaging service. Yeah, such

18:34

the simulated miss. The.

18:37

Skies A. Bookcase

18:39

related question Mms. The.

18:42

Know Afp, Rcs, Rcs

18:44

is. This. Is one of the ones

18:46

where I'm going to get system in service confused.

18:48

It's rich and communications. Service.

18:51

Yes, okay phrase. yeah so that

18:53

is is. so there's Sms which

18:55

is sort messaging services tax base

18:57

and happens over the wireless networks.

18:59

Them and Mms is when Sms

19:01

basically got upgraded to multi media

19:04

a service and then am Now

19:06

Rcs is the wireless network that

19:08

and Google backed sort of new

19:10

era I'm at of Mms. It's

19:12

bringing richer communications yeah to what

19:14

would typically be. Texas messaging? Yes,

19:17

yes. Tax stickers? Yeah, that fun

19:19

stuff. Holloway seeking any element of

19:21

entire relationship now through tobacco. Thanks.

19:25

For that easy one guy is so great! Okay,

19:28

we'll pick another one. Okay

19:30

Bbm. Blackberry,

19:32

Messenger. which predates

19:35

I message as one

19:37

of be ah the

19:39

first at. Peer. To Peer

19:41

direct messaging services on your phone that

19:44

was actually owned and operated by the

19:46

Blackberry network. Hey now that was really

19:48

fun! Sign This Bbm. it's my

19:50

favorite bbm thing is when

19:52

they started buying product placements

19:54

in television shows so characters

19:56

would say sodium me or

19:58

b b m And

20:01

then they pull out their blackberry and the camera would

20:03

show them on their blackberry doing that. Yeah.

20:06

Terrible. Yeah. Okay.

20:09

Here's another one for you. Okay. So see, sometimes

20:12

I'm on a chip. Yes. Yeah. That

20:14

refers to when it's one piece of silicon, but it's a

20:16

put together to create a system. And so you might have

20:19

one core or basically

20:22

one chip in the system that's dedicated to

20:24

ML, and

20:26

then you might have, which is machine learning, or you

20:29

might have another one that's basically dedicated to like

20:31

the core processing and yeah, but you put

20:33

them all together as a

20:35

system. Qualcomm is a very well known maker of

20:37

system on a chip. Apple makes its

20:39

own new. Lots

20:42

of people make them. Okay. Will,

20:45

throw her another one. Okay. How about

20:47

CDMA? This one is really

20:49

tough. Yeah. Yeah.

20:55

Was it like a precursor to 3G

20:57

and it didn't? No, 3G, it was

20:59

a type of network. So

21:03

early on mobile phones, I don't, were

21:05

pretty much divided between GSM and

21:08

CDMA. GSM was more popular in

21:10

Europe. So, Will, you might've been on GSM.

21:14

Here it was, yeah, you basically,

21:16

when you bought a mobile device, you had

21:18

to specify based on which wireless carrier

21:20

you were on, what kind of device you

21:22

were using. Oh right, that's right. Because you

21:24

had different ones. Yeah. Verizon was our CDMA

21:26

network. That's right. That's

21:29

right. And 3G was more in the category of like LTE,

21:32

which stands for long term evolution, the

21:34

type of service you would get. But

21:37

okay, I'm going to- You should have waited for us to ask you that if

21:39

we could get another point on the board. Oh

21:42

shoot. Okay. CDMA, it's consolidated? Nope.

21:45

Concentrated? Communication? Uh,

21:55

can you give me the first word? Code. Code?

21:59

DMA. I would like to talk to the manager. It's

22:01

just so obvious, so logical. What are you talking about? This is

22:03

really fun though. Bring it on. What's another one? I

22:05

don't know if this is really mobile era, but it's

22:07

three emojis. An

22:13

iemojis, an iemojis, an iemojis, an iemojis. What?

22:21

An iemojis? Yep, an eyeball. And

22:23

then the mouse. And then the mouse.

22:25

And then an eyeball. This is actually

22:27

a way more recent thing, I

22:29

think. Okay. I can't believe

22:31

it. I think it's okay. I'm a little

22:38

bit too old, actually, because I think I'm going to

22:40

have to get my eye on it. This

22:47

is actually a way more recent thing, I think. Okay.

22:50

So it's I. Here. No,

22:52

that's an ear. I talk. I

22:55

have no, no one's ever sent this to me. Am

22:58

I being left out? What is this? It stands

23:00

for it is what it is. Really?

23:04

Yeah, it's kind of a TikTok thing, I think. Oh,

23:09

wow. That's amazing. I

23:14

have to say, I

23:17

was picturing a totally different era here. Well,

23:20

this was the very beginning of the

23:22

pandemic, this ramped up. Oh, I

23:25

mean, no, not what I mean

23:27

is you went really early mobile and

23:29

then you fast forward it into the

23:31

future. I was waiting for like ARPU

23:34

and ATT and GPS and stuff like

23:36

that. Oh, GPS would have been a

23:38

good one. Yeah. I think

23:40

we should have had it. App tracking transparency. I

23:44

was like clearly in kind of

23:46

like app mode. All right, though,

23:48

I didn't. You know what? I had

23:51

a lot to learn. I think you did. We

23:54

clearly need some Gen Z person

23:56

to come on and do TikTok.

24:00

All right,

24:05

so Mike, I think that you're in the lead, technically, whatever

24:07

this competition is. I

24:09

wasn't keeping score, but okay, thank you for keeping score.

24:12

Yeah, okay. This is really fun. Thank you for schooling

24:14

me, guys. We are going to take a quick break.

24:17

And when we come back, we're going to spend the

24:19

entire next segment talking to Will, quizzing

24:21

Will about AI, because it's what

24:23

everyone's talking about. So

24:26

stay tuned. Support

24:30

for today's show comes from Deloitte. If

24:33

your company is like most, your

24:35

future depends in part on technology.

24:38

Yes, that means choosing the right

24:40

technology and adopting it quickly. But

24:42

that isn't enough. To gain

24:44

advantage, your technology needs to be

24:47

as outcome focused as you. That

24:50

means helping your people be more

24:52

efficient and more inventive, reducing cost

24:54

and creating revenue streams, growing

24:57

your customer base and building trust.

25:00

At Deloitte, we combine world class

25:02

business knowledge with a full command

25:04

of technology to help clients do

25:07

more than choose cloud or adopt

25:09

AI. We help them

25:11

create advantage from it. Read

25:14

case studies at deloitte.com/US

25:16

slash engineering advantage. This

25:25

episode is brought to you by Empower. Can

25:28

you retire early? Will there be

25:30

enough money to leave an inheritance? Do

25:32

you have savings for life's important

25:34

milestones? If you have money

25:36

questions, Empower has answers so you don't have

25:38

to worry. With a real

25:41

time dashboard and real live conversations, you

25:43

can get clarity on your real life

25:45

financial goals. Join 18 million

25:47

Americans and take control of your financial

25:50

future to empower what's next. Start

25:52

today at empower.com. Hey

25:56

everybody, it's Neal. I've got some huge news.

25:58

Decoder is moving to Monday's. and

26:00

Thursdays. We're adding a second episode of the

26:02

show. On Mondays we'll have

26:04

our classic interviews with CEOs and other

26:07

troublemakers. I think we're gonna have to

26:09

start having conversations about how do we

26:11

pay those jobs that can't be done

26:14

by AI. And on Thursdays we'll be explaining

26:17

big topics in the news with Verge reporters,

26:19

experts, and other friends of the show. There's

26:21

a new generation of people on the Internet. Google

26:23

search has always sucked for them so you know

26:25

there's no reason for them to be loyal so

26:27

they can just go to TikTok. This

26:29

is gonna be really fun. I'm very excited about

26:31

all this. So go subscribe wherever you get your

26:34

podcasts now. Okay

26:37

so now we've brushed up on basically 25 years

26:40

of tech acronyms. Congratulations. You're

26:44

now everyone's favorite dinner party guest. It's

26:48

time for AI. This is the part I think

26:50

everyone is going to be most interested in right

26:52

now because you're hearing these phrases literally all the

26:54

time and some of them are probably hard to

26:58

grok. See what I did there?

27:01

That's Elon Musk. No

27:03

it's not an acronym. We can't get through an entire show

27:05

without mentioning Elon Musk. Mike, do

27:07

you want to go first and grilling Will? Sure.

27:10

Okay. Will, I'm gonna kick off with the hardest

27:12

one on the list. Okay. LLM. Oh,

27:17

okay. That's a

27:19

large language model which

27:22

is what everything's built on

27:24

these days. Chatbots

27:26

and half of the internet. Built

27:29

on language, large language model. Yeah.

27:33

What to say about LLMs? They are, yes

27:35

it is an acronym that you hear it

27:37

hearing all the time. So if someone

27:39

is at a dinner party and someone starts

27:41

talking about LLMs in the context of AI,

27:44

how would you describe like here

27:46

was this era of AI but now everyone's

27:49

talking about LLMs because they do X. Okay.

27:53

Yeah. So language

27:55

models are this example of

27:57

generative AI. Generative AI

27:59

has been... around for decades and decades in

28:01

that you generate stuff with

28:03

an algorithm. But the

28:06

large language model era is this

28:08

era where they figured out feeding

28:10

it enough information, feeding it basically the

28:12

whole of the internet and as many books, a

28:15

lot of copyrighted material as well, it turns out, into

28:18

these particular types of algorithms,

28:22

enables them to conjure up very

28:24

convincing seeming text. But the large

28:26

is the key word there because

28:29

they're absolutely enormous algorithms and also

28:32

fetch huge amounts of information or data.

28:36

Nice. Nice. Thanks, Will. And

28:38

that's our show. That's

28:41

all you needed to know. No,

28:43

there's another one that we're hearing a lot these days, GPT.

28:46

GPT, oh, shh. OK. Generative,

28:51

pre-trained transformer, right? Yes.

28:54

So it's a generative. So that's the word.

28:56

Talking about the generative models is

29:00

the type of AI algorithm that doesn't

29:03

discriminate, doesn't recognize things in

29:05

text or images, but generates

29:08

stuff when given a prompt. And

29:10

pre-trained means that they are pre-trained

29:12

actually on specific types of data

29:15

before being trained on the

29:17

entire internet. And transform is a certain

29:20

type of algorithm that

29:22

is a neural network that

29:24

can focus on lots of different stuff at once, which

29:26

turns out is very useful for language because you kind

29:29

of need to know the end of a sentence in

29:31

the beginning to make sense of what's in the middle

29:33

of it. Stephen

29:36

wrote a great story about the

29:39

paper, attention is all you

29:41

need, which came out of Google, which sort

29:43

of transformed AI

29:45

by revealing

29:48

you could do kind of amazing things

29:50

with language using these models the

29:52

first time, or yeah, pointed that way. That's

29:55

our colleague, Stephen Levy, who wrote that. We'll put

29:57

a link in the show notes too. Yeah, you

29:59

can. read it now, it's really good. And

30:02

a lot of people probably know

30:04

this, but for those who don't, when you refer

30:06

to something like ChatGPT, which was released

30:09

by the company OpenAI, the

30:11

P in ChatGPT comes from that Google

30:13

paper, that group of Google researchers.

30:16

So it's kind of a derivative of Google's

30:18

talk. Yeah. Okay.

30:21

Yeah. Pretty much, I mean, this is

30:23

in Stephen's great story, pretty much all of the

30:25

people on that paper, the Transformer

30:27

paper, ended up creating

30:30

their own, well, a lot of them went off and

30:32

created their own startups and got billions

30:34

of dollars of funding. That's

30:36

why we all do it though, right? To make billions of

30:38

dollars. Humanity

30:42

be damned. Speak to yourself. Yeah.

30:44

Okay. Here's another one for you. The third

30:46

one. You're two for two, by the way.

30:48

Congratulations. You're already whooping at us. He's going

30:50

to sweep. NLP.

30:54

Okay. That's

30:56

natural language processing. Yes.

31:00

Pardon me. Well, it's processing. Processing,

31:02

sorry. Yeah.

31:07

As opposed to NLU, natural

31:09

language understanding. So that means doing

31:12

stuff with language, not just understanding,

31:14

not comprehending the language, but doing

31:17

also synthesizing sort of stuff. Okay. So

31:19

natural language understanding is so that a

31:21

computer can understand what you're saying when

31:23

you say something to it, but natural

31:26

language processing so that it can answer

31:28

you in a way that is like

31:30

human-ish. Is that right?

31:33

Yeah. I think it refers to

31:35

the bigger field of doing stuff

31:37

with language, I think. So it

31:39

could be, yeah, I think

31:41

it can encompass voice recognition

31:43

and processing, that sort of stuff. I think. I

31:45

could be wrong. We're going to go

31:47

with your right.

31:49

Yeah. Here's the next

31:51

one, which I think you'll get. RL. Oh,

31:55

reinforcement learning. So

31:58

this is... He's crushing it. Now these are

32:00

all more recent, and I think you're being kind

32:02

to me because there are a

32:05

billion abstract acronyms out there.

32:08

RL is the type

32:11

of algorithm that AlphaGo was, so

32:13

it's quite different to things like

32:17

chat GPT in that it's the

32:19

idea of having a computer learn through

32:21

experimentation how to solve a particular task.

32:24

So they deep-mined, didn't

32:26

invent reinforcement learning, that was this

32:28

guy who was rich something or

32:31

he was a pioneer of it. And then, but

32:34

they figured out you could with more

32:36

powerful computers get machines to

32:38

do quite impressive tasks that it's impossible to

32:40

program a computer to do. So

32:42

they started off with having it play Atari games better

32:44

than a person and then they

32:46

famously demonstrated it on Go,

32:50

the board game which is very difficult to

32:52

learn and play and it's very sort of

32:55

difficult to describe what makes a good move, so

32:57

it's hard to break up. The

33:00

idea is through sort of it has reinforcement

33:02

in the form of positive or negative feedback

33:04

for a good or a bad move moving

33:07

towards a goal. Interestingly,

33:11

the big thing that people are trying to do now to

33:14

move beyond chat GPT is combine some

33:17

of these two things. A

33:19

lot of places are trying to do that because so

33:22

chat GPT sort of goes off the rails and says mad

33:24

things and it doesn't understand what numbers are,

33:26

which is kind of a problem

33:30

in an intelligence. But you can use

33:33

reinforcement learning to have it possibly figure

33:35

out how to perform specific tasks maybe

33:37

including things like math, but it's

33:40

not demonstrated that it will

33:43

work but this is sort of an idea to

33:45

take these two big things in AI and

33:47

combine them. Okay, here's another one for you.

33:52

LSTM. Okay, this is, I have

33:54

to mention the name of the

33:58

board game. of the guy

34:01

who invented this because he's famously

34:03

gets upset when people don't it's

34:05

um Jurgen Schmidhuber invented the LSTM.

34:08

Yeah he's

34:11

quite a character um long

34:13

short shoot

34:16

something memories long short term memory.

34:19

Yes correct oh boy yeah I

34:21

know we should have made this

34:23

harder. So that's a type

34:25

of neural network that can

34:29

tap into memory essentially which is something they don't

34:31

have and it's there there are a lot of

34:33

these different architectures um this

34:36

is one that's quite was quite old it sort

34:38

of predates a lot of deep learning stuff but

34:41

it was very important and influential hence why Jurgen

34:44

likes to be credited with stuff.

34:46

He's actually I spoke to him recently and he was

34:48

doing some very interesting things trying to build new

34:51

these these models that kind of argue with

34:53

each other in order to figure out a

34:55

task like a kind of

34:57

Gestalt thing that if one

35:00

isn't good at it if another one

35:02

can figure it out you have these specialized networks.

35:04

Wow is

35:07

he a big Stanislaw Lem fan? He's

35:10

a huge science fiction fan actually yeah he's

35:12

he's about why do you why is oh

35:14

because of the computers arguing with each

35:16

other and like competing to try

35:19

to figure out problems. Maybe

35:21

that's where I didn't think of that yeah

35:23

he's a huge science sci-fi nerd um it

35:26

must be probably his inspiration for us. So

35:30

this is what you're describing is the new

35:32

era of the virtual assistant remembering

35:34

what you talked about before because there

35:37

was a version of this like on Google Home Assistant

35:40

several years back at this point I would say at

35:42

least five years back where Google would

35:44

say like you would say to the Google Home

35:46

or your assistant on your phone hey Google ask

35:49

it a question like how

35:52

tall is LeBron James and

35:54

then without having to prompt it again saying

35:57

okay and what team does he play for and having

35:59

these this volley back and forth

36:02

where it had a limited amount of memory to

36:04

remember what you asked initially. But now,

36:06

right, well, it's this movement towards you

36:09

could ask Chachi BT for an itinerary

36:11

for Barcelona and then come back days

36:13

later and pick up the thread and

36:15

just something like, what else should I

36:17

add to my trip? And

36:19

it's going to remember what it was you talked

36:21

about. And people are talking about this too, not

36:24

only in the form of

36:26

like customer service agents and stuff like

36:28

that. But even in

36:30

like EQ AIs, like ones that

36:32

are meant to do more emotional tasks,

36:34

if you're using one for therapy, imagine

36:37

coming back a week later and having

36:39

it remember what you talked about before. Yeah.

36:41

Yeah. That's it. Yeah. The idea

36:43

of having Chachi BT doesn't remember anything but

36:46

beyond a long streak, these

36:48

previous prompts. Typically, I think they've added some

36:50

more memory, but that idea of having a

36:52

bigger memory. Yeah. Okay. It's a big thing.

36:54

But the other thing that he's looking

36:56

at is the idea that because you can

36:59

have a small model that's better than GPT-4

37:01

if it's very much trained

37:03

on a specific task. So this is the

37:05

idea that you have a bunch of these ones that

37:08

interact with each other and then they

37:11

can work as well as a really big one. But

37:13

also, I think they're back and forth. The idea is

37:15

that it kind of shakes out more something

37:18

cleverer. I don't

37:20

know if it works. Okay. The

37:22

next one, Palm,

37:24

which I'm pronouncing like a word, but

37:27

it's P lowercase a and then capital

37:29

L M. Oh,

37:31

God. I'm not going

37:33

to get this one.

37:35

Yeah, we stumped out.

37:37

I'm going to make

37:39

something up. I wouldn't

37:42

get it. It's a language model at the end,

37:44

probably. Oh, yeah. Well,

37:51

just so you know, before this episode, we

37:53

joked about calling this WILM, which would be

37:56

W-I-L-L-M, the Will

37:59

Learning. large language model.

38:02

One day my language model would be able

38:04

to come in my place and

38:07

get it all right because it could look it up. I

38:10

don't know. Free-trained,

38:16

amazing language model. Nope.

38:19

Good guess. It's Pathways

38:21

language model. Oh. Okay.

38:25

Yeah. Yes. We got them. So who

38:27

does this? This is the one, this

38:29

is the precursor to chatgbt.

38:32

They, right, they built

38:34

some pretty amazing chatbots on top

38:36

of it but never released them I think.

38:38

Okay. That's right. Well here's

38:40

another one with a lowercase in it.

38:42

This one is llm.

38:48

I don't

38:51

know this one either. I

38:53

think you do only because I'm

38:55

gonna give you a hint because I think

38:57

you know the company behind it and that

39:01

letter factors into this. Well,

39:06

so meta's behind llm.

39:14

Two Ls. Language,

39:18

how does meta, I don't know. I'm

39:24

stuck. No idea. It's

39:27

large language model

39:29

meta AI llama.

39:33

Makes perfect sense. Yeah, right. So

39:35

there's like extra, so there's large

39:37

language, lowercase a, model, meta AI,

39:40

but there's not two M's, it's

39:42

just llama. Should

39:44

have even got large language and made

39:46

the rest up. They really just weren't

39:48

meta in there. They did, yeah. Because

39:52

it was an alpaca I think. I think they wanted

39:54

to get something, I think.

39:56

I might be hallucinating that. Anyway,

40:01

the next one, which will be the last one, is it's

40:04

a bit of a trick because

40:06

it's not an acronym, but maybe you can describe where

40:08

it comes from. DALI. D-A-L-L

40:11

dash E. Oh, I do know this. It's

40:13

a, I might get this word wrong. Is it

40:15

a portmanteau? Is that what you say? Yes,

40:17

portmanteau. It's a portmanteau, sorry, of Salvador Dali

40:19

and Walle. Good

40:24

job. Excellent. Wow. I

40:27

just remember that because it was, every

40:30

story had to explain that

40:32

it was a portmanteau or whatever

40:34

the word is. And so I was like, shit, I

40:36

better learn that word, which I haven't done. This

40:39

is an image generator. Right, right.

40:42

This is the open AI

40:44

image generator that is

40:47

trained on lots of

40:49

nice artwork. Yes,

40:52

trained on the whole internet, which is also artwork, I

40:54

guess. All belonging to open

40:56

AI. Yes. Only

40:58

within copyright. No issues there. Yeah.

41:02

I think Will won. So I... Handily.

41:05

Did I? I think so. Yeah,

41:07

you got six out of eight. Okay. I

41:10

felt like I fell off a cliff. Well,

41:13

yeah, but you win all your games in the first

41:15

half of the season, it's okay to lose

41:17

some in the second half of the season. Is that how

41:19

it works? Don't you get closer to the playoffs in the

41:21

second half? Yeah, but your stands are

41:24

full all year round if you win all your

41:26

games in the first half of the season. Spoken

41:28

like a true capitalist. I'm an A's fan. He's

41:30

bringing A's hat right now. But

41:35

it might be slightly different if

41:37

you guys were writing about CDMA

41:40

and the like every week. That's true. Yeah,

41:43

that's true. This is true. I

41:45

mean, I referenced those things a lot.

41:49

When I first started writing about tech, I

41:52

was a video journalist at the journal and then I

41:54

started writing and I remember my first story was about

41:56

MDTV, mobile DTV. But

41:59

then after that, I just... I just sort of cruised right into like,

42:03

modern smartphones and App Store. The

42:06

early stuff is really, that was really fun.

42:08

Foundational. Super fun to learn about. You did

42:10

well. I don't know. Not

42:13

so much. I can admit defeat. What

42:15

does Will get? What's his prize? He gets to

42:17

go first on recommendations. Ooh, good

42:19

one. Let's take a break and come back with

42:21

those. This podcast is supported

42:23

by tools and weapons.

42:27

The podcast hosted by Microsoft Vice Chair

42:30

and President Brad Smith. Each

42:32

episode features insight you won't find anywhere else

42:35

from the center of the conversation surrounding emerging

42:37

technologies like AI. Right

42:39

now on the podcast, you can hear a special

42:41

episode where Brad Smith lays out Microsoft's vision for

42:44

a vibrant marketplace, driving the new AI economy. To

42:48

hear more, follow or subscribe to Tools and

42:50

Weapons with Brad Smith, wherever you get

42:52

your podcasts. Hi, I'm Jeremy Larson,

42:54

the Reviews Director of Pitchfork and

42:58

this podcast is supported by Pitchfork Music Festival. Pitchfork

43:00

Music Festival will take place July 19th through

43:03

the 21st at Union Park in Chicago, Illinois. This

43:05

year's lineup features Jamie XX, Alanis

43:08

Morissette, Black Pumas, Carly Rae Jepsen, Brittany

43:11

Howard, J. Paul, Moona, Jessie

43:13

Ware, 100 Gecs and many more. The

43:16

festival also features diverse vendors as well as

43:18

a specialty record, poster art, and a new

43:20

album. The festival features diverse vendors as well

43:22

as a specialty record, poster and craft fairs

43:24

and works to support local businesses while promoting

43:27

the Chicago arts and food communities as a

43:29

whole. For more information

43:31

on tickets and lineup, visit pitchforkmusicfestival.com.

43:45

Well, as the winner of our

43:47

acronym competition, all

43:50

we can give you is the option

43:52

to go first in recommendations. So

43:55

have at it. I wish

43:57

I'd known, because I would have deliberately failed

43:59

to buy my... have more time. I'm trying

44:01

to come up with this on the fly,

44:03

but I did think of it before. So I'm

44:05

going to recommend a book

44:09

I read recently, The

44:11

Rise, I'm holding it up for the listeners

44:14

on the Rise and the Fall of the

44:16

East by

44:18

Yashan Huang. It's a, the

44:21

East in the title stands for exams,

44:23

autocracy, stability and technology. It's a kind

44:26

of incredibly in-depth

44:29

historical analysis

44:31

of Chinese bureaucracy, which sounds like

44:33

a real page turner, but

44:35

it's actually incredibly well done

44:37

and does a lot to actually

44:39

explain how China got

44:42

to where it is and why sort

44:45

of this paradoxical place

44:47

where it seems very

44:49

innovative, but in a way, in some ways, it's

44:52

not quite, it's not as innovative or it's sort

44:54

of held back. And I find

44:56

it very convincing as to why

45:00

also China is where it is right now with the

45:02

current leadership. And

45:05

it's really fascinating if you're interested

45:07

in China and its technology. Hmm.

45:10

And how did you hear about this book?

45:12

I, so I know Yashan Huang

45:15

from MIT and

45:17

I know his work. He's quite a famous scholar on

45:19

China and it's tech industry

45:21

and they asked me

45:23

to interview him on

45:25

stage there where I lived down the road from IT.

45:27

So I had to furiously read it in two weeks,

45:29

which was kind of intense, but I sort

45:31

of knew, I knew he had one in the works because

45:36

I'd been talking to him. This is how

45:38

modest Will is. I think a lot of us would

45:40

lead with him. You know, so I spoke to this

45:42

really smart thinker and author recently and our talk was

45:44

so scintillating and also he has a book and

45:47

Will kind of led with, you know, I'm

45:49

reading this book and then how do you

45:51

know him? And oh, I happen to enter

45:53

MIT. I happen to interview him. I'm just

45:55

very incompetent at presenting things. Not

45:59

at all. Will we will not only link to

46:01

the book in the show notes, but we'll link to your talk as

46:03

well. Thank you Mike What's

46:05

your recommendation? My recommendation

46:07

is on topic for this week's show It's

46:10

called the jargon file and it's this open

46:12

source Text document that lives out on the

46:15

internet that you can print out or you

46:17

can just look at it can be copied

46:19

and Paste it in other places, but it's

46:21

basically a big Glossary

46:23

of like hacker slang and acronyms

46:25

and fun terms. There's a lot

46:27

of humor in it and

46:30

it's a fun way to like look at

46:32

computer history through like the language and the

46:35

rhetoric that people have Like

46:37

associated with computers and how hackers talk to

46:39

each other So yeah, the jargon file is

46:41

just it's basically just like an A to

46:43

Z dictionary and you look up words For

46:46

example some words that are in there. I'm

46:48

just gonna flip through bit rot

46:51

Is a good one Cyber space

46:53

and cyberpunk are both defined in here dancing

46:56

frog Kill

46:58

file The

47:01

pumpkin The

47:03

patch pumpkin the pumpkin holder Spoiler

47:07

space vaporware is defined

47:09

in here Windows

47:11

with a Z is defined in here. So

47:14

it's it's really fun like way of Looking

47:16

into into hacker culture and computer

47:19

science over the years. Cool. Yeah,

47:21

who created it? Uh, it's a it's

47:23

an open source project So a bunch of

47:25

people have edited it and contributed to it

47:27

over the years someone must have started it

47:29

I'm sure somebody did start it. Yes, I

47:31

don't know who that person is But

47:34

it's very easy to find because it's

47:36

hosted multiple places and you just need

47:38

to search for jargon file You

47:41

search for those two words together and you'll find it. It's

47:43

just a big HTML document

47:46

Super cool. Yeah, super nerdy. Yeah, that's what we're

47:48

here for. It's a lot of fun I thought

47:50

we should rename our podcast super nerdy super nerdy

47:54

Then we actually have to live up to that descriptor

47:56

and I don't know that's a couple of english majors

47:59

sometimes I Yeah, sometimes my

48:01

recommendations are very basic. Okay, I

48:03

want to hear your, is this all CAPS basic?

48:05

B-A-S-I-C basic? No, but I like how

48:07

you're bringing it back. Okay. My recommendation,

48:10

it's not at all academic or

48:12

nerdy. It's an app called

48:14

Forest. Some of you might be

48:16

familiar with the Pomodoro method of working, which

48:20

means you set a timer and you work

48:22

for 25 minutes straight,

48:24

no distractions. There are

48:26

a lot of different apps that

48:30

sort of take advantage of the Pomodoro

48:32

method and then create different user interfaces

48:34

for it and different mechanisms. There

48:36

are also web versions for people who get

48:39

really distracted on their desktop

48:41

with stuff flying in their browsers and 18 different

48:44

browser tabs open. This one

48:46

is a mobile app. I think it cost me $2.99

48:48

to download. And

48:50

I just clicked the button and showed my face and

48:52

I had it, but I'm pretty sure it was $2.99.

48:55

Is that how we buy things on the internet?

48:57

No. And

48:59

it simulates planting trees. So every time

49:02

you work 25 minutes uninterrupted, set the

49:04

timer, you've planted a tree at the

49:06

end of it. So it's the Pomodoro

49:08

method, a little bit gamified, and it's

49:11

just a good way to cut down

49:13

on distractions if you feel

49:15

like you're a little bit ADHD as

49:17

you're working. Okay. Yeah. So check out

49:19

the forest app. Then you can, if

49:22

you want to work for an extended clip, you can

49:24

just set 25 minute increments with a

49:26

five minute break in between. I'm not using it

49:29

quite like that yet. I'm just picking a 25

49:31

minute block and going for it. That's

49:33

nice. Helpful for writers. Yes. Or

49:36

anything task-based. Correct. Yeah.

49:38

All of a sudden thinking about doing administrative

49:40

work or filing expenses or something like that

49:42

becomes a lot more tolerable if you just

49:45

think, I can do

49:47

this for 25 minutes. Go make a tree.

49:50

Go plant a tree, a little virtual tree.

49:52

That's lovely. In a simulated world like

49:54

Sim. I had to bring

49:56

it back too. Before we say goodbye, I just

49:58

want to give a shout out to... So folks who

50:00

have been leading us reviews on the Apple Podcasts

50:02

app, we love the reviews. We genuinely

50:05

read them. We

50:07

appreciate the feedback and thank

50:09

you for everyone who's listened to

50:11

this point. Go leave us a review. We could

50:13

say thank you to the people who've left

50:15

reviews on the Google Podcasts app, but that's

50:17

gone. So there is a one-day. RIP Google

50:19

Podcasts. I know where else do people

50:21

leave podcast reviews? YouTube probably.

50:24

Oh, all right. We'll see.

50:26

We're not on there. Sure we are. We

50:28

are? YouTube publishes there, I think. But

50:31

you don't see our faces. You just see

50:33

the robot blowing the bubble. Yeah. Are

50:36

you guys on TikTok? Do you have

50:38

a TikTok? Sometimes. Our podcast

50:40

is on TikTok. I go on. Mike's on

50:43

TikTok a lot. He needs the Pomodoro

50:45

app to focus because TikTok

50:47

is ruining his life. He loves the

50:49

dancing teen videos. It's really weird. He's

50:52

shaking his head right now. Because I don't.

50:56

I don't. You knew what the emoji, the

50:59

TikTok emoji were. Well, yeah, because I

51:01

had the sheet that Will and I shared. Oh,

51:03

right. Okay. Speaking of, Will,

51:05

thank you so much for joining us. Thank

51:08

you for having me and for letting me win. I appreciate

51:11

it very much. You're very welcome. It's been an absolute

51:13

pleasure. And Mike, thanks for being a great

51:15

co-host. Of course, anytime. And thanks to

51:17

all of you for listening. If you have feedback, you can

51:19

find us all on the social networks. Just check the show

51:21

notes. Our producer is the

51:24

excellent BA, otherwise known as

51:26

Boon Ashworth. Goodbye for now. We'll be back

51:28

next week. Hackers

51:36

and cyber criminals have always held

51:38

this kind of special fascination. Obviously,

51:41

I can't tell you too much about what

51:43

I do. It's a game. Who's

51:45

the best hacker? And I was

51:47

like, well, this is child's play. I'm

51:49

Dina Temple-Reston. And on the

51:51

Click Here podcast, you'll meet them and the people

51:54

trying to stop them. We're not afraid of the

51:56

attack. We're afraid of the creativity

51:58

and the intelligence of the human beings. behind

52:00

it. Click here. Stories about

52:02

the people making and breaking our digital

52:04

world. AI machines,

52:06

satellites, engine ignition. Click here.

52:09

And with that, click

52:11

here every Tuesday and Friday wherever

52:13

you get your podcasts.