0:00

It's time for Twitter this week in Tech.

0:02

We have a great show. Tim Stevens is

0:05

here. Harry McCracken, Christina

0:07

Warren, three of my favorite people, and

0:09

course, AI is the topic.

0:12

It's an amazing world we live in and some

0:14

of the new things that are happening with AI.

0:16

And some of the old things that maybe aren't

0:19

so good. We'll talk about the

0:22

Microsoft quarterly results.

0:24

That's so hot. Intel, the

0:27

worst quarter in a long time.

0:30

And about the Oscar campaign that

0:32

took Twitter by storm and worse.

0:35

It's all coming up

0:36

next. On TWiT,

0:39

podcasts you love from

0:41

people you trust. This

0:45

is TWiT. This

0:51

is TWiT. This weekend tech

0:54

episode nine hundred twelve recorded

0:56

Sunday, January twenty ninth twenty

0:58

twenty three. Let me consult

1:01

my AI lawyer. This

1:03

weekend, Tech is brought to you by worldwide

1:06

technology TWiT an innovative culture,

1:09

thousands of IT engineers, application

1:11

developers, unmatched labs and

1:13

integration centers for testing and deploying

1:16

technology at scale. WWE

1:18

helps customers bridge the gap between

1:21

strategy and execution. To learn

1:23

more about WWT, visit WWT

1:26

dot com slash TWiT.

1:29

And by ACI Learning.

1:31

Tech is one industry where opportunities

1:34

outpaced growth, especially in cybersecurity,

1:36

one of information security jobs require

1:39

a cybersecurity certification to

1:41

maintain your competitive edge across

1:43

audit, IT, and cybersecurity

1:46

readiness, visit go dot ACI

1:48

learning dot com slash quit.

1:51

And by bitWarden, get the password

1:54

manager that offers a robust and cost

1:56

effective solution that can drastically

1:58

increase your chances of staying

2:00

safe online. started with

2:03

a free trial of a Teams or enterprise plan,

2:05

or get started for free across all devices

2:07

as an individual user. At bid

2:09

wharton dot com slash

2:11

TWiT. Thanks for listening

2:14

to this show as an ad supported network.

2:16

We are always looking for new partners

2:19

with products and services that will benefit

2:21

our qualified

2:22

audience. Are you ready to grow your business?

2:25

Reach out to advertise at TWiT dot

2:27

tv and launch your campaign now.

2:35

It's time for TWiT this week at tech, the show we cover

2:37

the week's. Tech news. I'm

2:40

just gonna put a little black iron band if you don't

2:42

mind on

2:43

the San Francisco niners. Gold

2:46

throwback jacket. That's

2:48

life. Tim Stevens is

2:49

here. Hello. Oh, that was Harry.

2:52

Hello, Tim. Good to see you.

2:54

Hey, Leo. Good to see you as

2:55

well. Thank you for having me. Tim, of

2:57

course, has been on for many years. He is now a freelancer

2:59

at Jalopnik, at TechCrunch, at Motor

3:01

Trend. The virgin, he has his very

3:04

own substack. Tim that's

3:06

substack dot com. Great article. On

3:09

your visit to the Dicar,

3:12

rally in Saudi Arabia.

3:14

Wow.

3:15

Yeah. That was quite a A really interesting

3:18

social experience on a lot of levels and it

3:20

optimized a lot of things, but amazing

3:22

events in doing doing a lot of great travel

3:24

lately, so I've been very lucky.

3:25

Nice. Well, it's great to have you back. Not much

3:28

ice race. Thanks,

3:28

Leo. Sadly not.

3:30

Are

3:30

you aware of that, Tim? Can't really

3:32

be heard at least. Right? You

3:34

can't hear him, not real just barely. Alright.

3:37

Hold on. We're not ready to begin yet.

3:42

Oh, you have a little he's

3:44

not guy. He doesn't have headphones on. So

3:47

we can't do a

3:48

bleed. I understand why we had the bleed.

3:49

That's what I did.

3:50

Harry, do you mind now? Headphones.

3:52

I'm happy with whatever works for me. We could provide

3:54

you with headphones. Sure. I apologize.

3:57

No problem. Do you have some

3:59

work? You can get them out of my office if you don't.

4:05

Give them some nice ones. The

4:07

good stuff. Give them good stuff.

4:09

Give them the good

4:10

ones. I think there's an unopened box in

4:12

my covered on the left

4:14

there. Give him some

4:17

sterile headphones. I'm

4:20

sorry, Harry. I --

4:21

Okay. -- wasn't paying attention. Yeah. Usually,

4:23

we use a bleed, but I think

4:24

I suddenly realize you might not be aware of that.

4:26

I can't remember. Yeah. That would be kind of

4:28

a disadvantage to the overall

4:30

program. Be

4:33

kind of a bad thing. Good.

4:40

Nobody should hire Demiko Ryan's as

4:42

a head coach. That

4:44

would be a terrible idea. It's

4:50

Lisa's birthday, and I really wanted her to have

4:52

a nice birthday.

4:56

Happy birthday, Lisa.

4:57

Happy birthday, Lisa. It's also our anniversary

4:59

because I Foolishly

5:01

thought I if we got married on

5:03

her

5:03

birthday, I would only have to give her one

5:05

gift.

5:07

You didn't think that. That wasn't white

5:09

I just thought it'd be easier to remember. One

5:12

less one fewer date to

5:14

remember or

5:15

something, I don't know. Yeah.

5:18

That was fun. Unfortunately,

5:21

the

5:21

place we got married, Calisto

5:23

a ranch has burned to the ground and the wildfires is

5:26

gone.

5:27

Oh. So Which makes me sad. She says,

5:29

as life, I think at least for this. No.

5:32

Because we we used to go

5:33

there, you know, on our anniversary and stuff.

5:35

TWiT was really nice. So

5:37

but I've just learned that we had a collar

5:40

from the new Kona village, which is opening

5:42

this summer in Kona, Hawaii.

5:45

I've just learned it that. Reopening, and

5:47

that's somewhere I've always that was on my

5:49

bucket list somewhere to stay. So that's

5:51

where Steve Jobs was staying when the

5:54

iPhone four antennagate

5:56

happened. Mhmm. He didn't wanna

5:58

come back, but they made

6:00

him come back. That's how good it

6:02

is, I guess. You

6:04

you know, there's no there were at the time, there were

6:06

no TVs, no phones, no Internet. It

6:08

was, like, you were in a staying in a

6:11

traditional Hawaiian holiday. Sounds

6:14

alright. How's

6:18

that? You can hear it and you get

6:20

the volume there so you can control that as

6:22

don't deafen

6:23

yourself. Can you hear me,

6:24

Harry? Harry, can you

6:26

hear

6:26

me? Harry,

6:28

can you hear me? I can hear you because you're

6:30

sitting next to me. Oh, It's

6:33

not a good test.

6:38

One, two, you can you hear me? Are

6:40

you receiving me? Should I tell you what I have for

6:42

breakfast tonight? Okay. Good. Alright.

6:43

Awesome. Alright. I think we're good.

6:46

Yay.

6:50

Alright. Here we go. Yeah. Someday we'll have

6:52

four in the studio again. That

6:55

has that has happened. Christina

6:57

was here, but I don't

6:59

think we've had it. When was the last time we had an all in

7:01

person show? It's been a it's been a while.

7:04

Alright. Starting over. You can

7:06

hear. I can hear it. I think.

7:10

It's time TWiT this week to check the show week cover

7:12

the week's tech news. With

7:15

a panel of fabulous people,

7:17

I'll start over on my right with

7:19

mister Tim Stevens. We haven't

7:21

seen in a while freelance writer now.

7:23

You've you see Tim's stuff

7:25

all over the place, Jalopnik and TechCrunch,

7:28

and motor trend in the verge. He also has

7:30

his very own substack

7:33

called around the next bend.

7:35

Oh. Leave it myself.

7:38

I'll buy your loan some. That's

7:40

all I said to me, Lou, it's great to be here. No. It's great to

7:42

see you. I missed you. And

7:45

lots of stuff to talk about. You you

7:47

just came back from the the car and

7:49

the Descartes Road rally. And I loved

7:51

your pictures, but it's a was an interesting

7:53

mixed, I guess mixed bag

7:55

of

7:56

experiences. Yeah.

7:57

Yeah. Thanks. It was a great treat to Saudi Arabia.

7:59

Learned a lot of things, both good and bad. Yeah.

8:02

Also with us in studio, because

8:05

COVID is over no.

8:08

McCracken, global

8:11

tech editor at Fast Company,

8:13

Fingers

8:13

crossed. Hello, Harry. Good

8:15

to see you. Nice to actually see you in

8:17

person. Yay. And you brought your wonderful

8:19

wife Marie with you. Great

8:20

to see you

8:21

all. She has custody of Lily.

8:24

The TWiT pet. She

8:25

will be taking Lily home with us. I'm sure.

8:28

Lily is about the best dog you ever

8:30

saw in your LifeWorks

8:31

dog. But she lives here. Well, I

8:33

shouldn't say that because I think it's in our lease.

8:36

She's not allowed to spend any time here,

8:38

but I didn't

8:40

say that. I don't think the

8:42

landlord watches TWiT. Also

8:46

great to see Christina Warren

8:48

from GitHub last time you were in

8:50

Studio. Senior deck out of the kit

8:52

over there at GitHub. Good to see you.

8:54

Glad to be here. You made the

8:57

move this week when Ivory

8:59

came out. Tappots was one

9:01

of the third party apps

9:03

that mister Musk clobbered

9:05

at first without warning then

9:07

with a lie saying, you've

9:10

been violating the rules. For

9:12

fifteen years, you just noticed.

9:15

Finally, they said, oh, they retroactively changed

9:17

the rules, no third parties. But

9:20

that wasn't enough to push you to Mastodon IV

9:22

was the thing that did it. Tapots. Was

9:24

a very, very nice or TWiT rather. It

9:26

was a very nice Twitter client

9:28

from Tapbots. And Ivory is basically

9:31

TWiT for Mastodon.

9:35

Yeah. Yeah. And and honestly,

9:38

it was kind of a combination of things. It was

9:40

that that was, I think, really kind of

9:42

the final straw. Also, as a lot of

9:44

people have commented on. My

9:46

posts don't show up in people's feeds, and I

9:48

don't see replies, and I don't see other people's

9:50

posts. And so the whole experience was

9:52

becoming degraded. And then not

9:54

only did I have I

9:54

have, like, Ivory, which was great, but there's

9:57

ice cubes, which is a great opening I like

9:58

ice cubes a lot. There's a lot. Tell her, ice cubes.

10:00

It's fantastic. There's There's

10:02

elf dot zone, which is a great web

10:04

interface. There are I've actually

10:06

have a get help list I've been making of different

10:08

-- Yeah. -- cool people. This

10:09

is the beauty of open source. And an and an

10:11

open standard.

10:12

You can

10:12

anybody can develop and they can't cut you

10:15

off. Right. Right.

10:17

And and so and I've already got

10:19

it, like, about, you know, there

10:21

will be some people who follow me on Mastodon who didn't

10:23

follow me on Twitter, but I've got ten

10:26

percent of, you know, the followers that I

10:28

had on Twitter now on Mastodon,

10:30

which not

10:30

bad, you know, for for four or five days in.

10:33

So A lot of people are reporting increased

10:35

engagement even though they're fewer

10:37

followers.

10:38

Unvested. I'm

10:38

sure maybe just that. Yeah. Absolutely. Look at the

10:41

quality as

10:41

guys. I have two.

10:42

Yeah. I think No. That no.

10:44

That it will change over time. Right? Like, I

10:46

think that it's more people join, you

10:48

will see less of that high signal to

10:50

noise. But right now, I totally agree. Like, I'm

10:52

definitely seeing higher higher engagement higher

10:54

quality. It's pretty clear that Elon

10:56

has decided to heavily

10:59

algorithm it ties the feed

11:01

on Twitter. He's he's even said you

11:03

pay eight bucks and more people

11:05

will see you. And I think he's

11:08

I don't know why he thinks

11:10

eight bucks from a

11:12

a few hundred million at best.

11:14

Users is gonna make enough money to pay for

11:16

TWiT, but and the loss of ad

11:19

revenue, but he's doing whatever he can. And

11:21

that's the problem, though, is that it then

11:23

tells people, oh, you

11:26

know, nobody's Nobody's engaging with me. I

11:28

don't wanna be here. So you're driving

11:30

off your creators. Actually, Corey,

11:32

doctor Roe, wrote a

11:34

good story about this this week.

11:38

It's an impolite title. So I'm gonna

11:40

say, TikTok's in

11:43

certification. Using

11:45

the good places euphemism for

11:47

that word. It

11:50

I thought was quite insightful as

11:52

usual Corey made something that's, you know,

11:54

been around and obvious to all of us

11:56

crystal clear, put in words

11:58

that a light bulb goes off. He

12:00

says, here's how platforms die.

12:02

First, they're good to their users. Then

12:05

they abuse their users to make things better

12:07

for their business customers. Finally,

12:10

they abuse those business customers.

12:12

To claw back all the value for

12:14

themselves, then they die.

12:17

And he gives an example, Amazon, which was

12:19

customer first, customer first,

12:21

and then you know, as the as the

12:23

customer base got locked in with a variety

12:25

of techniques like Amazon Prime and

12:27

DRM and so forth, then they said,

12:29

alright. Now businesses businesses then

12:31

the marketplace fifty percent Amazon sales are

12:33

in the marketplace, third party sellers.

12:36

But they got locked in even though they

12:38

lose forty five percent of revenue

12:40

to Amazon and

12:40

fees. And now Amazon

12:44

says screw you, you're locked

12:45

in and they start monetizing

12:48

he says the company's

12:50

thirty one billion dollar, and he puts

12:52

it in quotes, advertising program is

12:54

really a payola scheme. That pits

12:56

sellers against each other forcing them to bid on the

12:58

chance to be at the top of your

13:00

search. But what ultimately happens

13:02

is you've ensured to find your

13:04

platform to the point where no one wants to use it

13:06

anymore. This is very

13:08

clearly where Elon is. Twitter TWiT first

13:10

was all about the users. They couldn't figure out how

13:12

to monetize it. Then they got brands to

13:14

go there. In fact, that's one of the reasons all of

13:16

us were there. Right? That's the best you have

13:18

to be there to promote your

13:20

brand, to to build your audience. Then

13:22

once they got them locked in, now they

13:24

can say, hey, if you wanna reach that

13:27

audience which we own, it'll

13:29

be eight dollars, please. But

13:31

You do that at the risk of driving people like Christina

13:34

away. He's talking about the in

13:36

this article, particularly about TikTok.

13:38

Doing this, but it happens

13:41

to every one of these companies. His

13:43

position, which I really agree

13:45

with, is, you know,

13:47

that's this is the way it is, and

13:49

you just move. You go to the next

13:51

thing. You leave May space for Facebook, you leave

13:54

Facebook for somewhere else. But

13:56

what we need regulation for is to make

13:58

sure it's as friction free as

14:00

possible to move, to avoid the

14:02

lock in. You need

14:04

interoperability. You need to make

14:06

it easy to move somewhere

14:07

else. And then then

14:10

you can let the

14:11

market you could

14:11

let the market rule. He

14:13

says, as I said at the start of this essay, this is

14:16

towards the end. In certification, exerts

14:18

a nearly irresistible gravity on

14:20

platform capitalism. You

14:23

know, the staff, the

14:25

executives the shareholders,

14:28

eventually, they all say,

14:30

no, you got to ensure to fight. We need the

14:32

money. But

14:35

even the most locked in user eventually reaches

14:37

a breaking point and walks away

14:40

or or gets pushed. Individual

14:45

product managers, executives,

14:47

activists, shareholders all give preference to quick

14:49

returns at the cost of sustainability

14:51

and they're in a race to see who can love

14:54

love, Corey. Eat their seed corn

14:56

first. In certification,

14:58

has only lasted for as long as it has because the Internet

15:01

has devolved into five giant websites.

15:03

Each filled with screenshots of the

15:05

other four. Gary's. Gary's.

15:07

Gary's getting a little little

15:09

cranky in his old age. I

15:11

don't know. In certification

15:13

kills, Google just laid off twelve

15:15

thousand. And the company's in a

15:17

full blown panic over the rise of AI

15:19

chatbots. What

15:23

are you what are your thoughts, Tim? I

15:26

I definitely, you know, the the

15:28

pattern is very clear when we certainly seen it before. What's

15:30

missing though, I think, is the actual death of these platforms.

15:33

I I think, you know, Twitter is certainly struggling,

15:35

and I think a lot of us are are kind of

15:37

thinking that these days are numbered, but it's

15:39

still incredibly huge and incredibly popular

15:41

and as much close to say, you

15:43

know, engagement numbers are up because everyone's

15:45

kinda watching the dumpster virus

15:47

motors. So I think it's a little too early to say that

15:49

Amazon has died, that that Google has died,

15:51

that Twitter has died, and so I think

15:53

that's the piece of the pattern that's

15:55

that's missing in this case for

15:57

better or for

15:57

worse. We certainly there are, of course,

15:59

plenty of companies. And this is, by the

16:01

way, not just tech companies, any company with

16:03

any

16:03

consumers.

16:04

There are plenty of companies in that graveyard.

16:07

We're in the process of watching these companies

16:09

move in that direction. But you're right. I mean, it's it's

16:11

hard to imagine Google going

16:13

away. Facebook. Maybe it's

16:15

not so hard to.

16:16

Although Facebook's latest data

16:19

on engagement,

16:21

since they started pushing videos

16:23

from people I don't even follow into

16:25

my feed. Apparently, that's actually

16:27

working at least right now in terms of engagement

16:30

that the AI they're using. To put

16:32

videos in front of you. It actually does seem

16:34

to determine stuff that people will

16:36

watch. And so it's the

16:38

numbers are a little encouraging

16:39

lately, especially given how

16:41

you know,

16:43

a

16:43

little good news. Facebook has had any

16:46

front in the last couple of years. Yeah.

16:48

Mhmm.

16:50

Christina, is does this

16:52

process end with

16:55

the end of

16:55

TWiT, or does Twitter just kind of

16:57

drag on? Well,

16:59

it can it can be both. Right? And we've seen

17:01

both because we've definitely seen social networks

17:04

just go under and

17:06

and just to disappear and that has

17:08

happened. And Google Plus is a great example of that

17:10

where obviously, you know, Google

17:12

put a lot of money and a lot of effort into that,

17:14

and it failed, and then they shut it down and got rid

17:16

of all the archives even, which I actually

17:18

thought that was not a great move to not

17:20

even keep the public archives

17:22

available, but that was like a

17:24

high profile failure. There have

17:26

been other ones. But then you also have

17:28

instances where they continue to

17:31

kind of stick around until

17:33

they're sold and and deleted and

17:35

what not myspace being a great example of that

17:37

where, you know, that has now

17:39

had god only knows how many owners and

17:42

people trying to use that very

17:44

worthless at this point, email list.

17:46

Of of users. But myspace,

17:48

you know, was bigger than Facebook

17:50

up until about two thousand

17:52

nine. I wanna say, And

17:54

then you started seeing a really big migration of

17:56

people from MySpace to

17:59

Facebook. To the point that that MySpace just

18:01

kind of became a dead zone except

18:03

for a very specific niche of

18:05

people. And that wasn't

18:07

really unlike Live Journal and

18:09

Geocities and and some and

18:11

Tumblr and and some other things. It wasn't

18:13

really because of any policy changes

18:15

that my space made. It was just because

18:17

the masses were all on Facebook. And

18:20

you know, Twitter is interesting because as

18:23

Tim says, it's still this giant

18:25

place. I think that

18:27

what will potentially be pushing people

18:29

off of it is less the

18:31

alternatives and more when the

18:33

overall experience becomes

18:35

degraded, whether because more

18:37

toxicity is there or, you know, just because

18:39

you you can't you're having errors in your

18:41

feed, you know, you're not able to post things

18:43

the right way. You can't refresh as quickly.

18:45

You don't see all of your replies. You

18:48

know, that's the sort of thing that makes

18:50

people go, okay, why am I investing time

18:52

in this? And and, arguably, you could say that

18:54

the demise of Twitter started

18:56

probably, you know, two thousand sixteen.

18:58

Ironically, when its engagement was higher,

19:01

when you started to see lot of the the previous, like, high

19:03

profile users of Twitter leave

19:05

the platform for Instagram and

19:07

then later, you know, TikTok,

19:10

but you you stopped seeing the celebrities on Twitter.

19:13

And I I don't know.

19:15

I mean, is this one of those well,

19:17

what's the Neelos

19:19

doesn't like it happens, you

19:19

know, slowly and then all at once. Yeah.

19:22

It's it's like the collapse of the

19:24

Give it give it give it give it's set

19:26

at first. Have if Scott Fitzgerald

19:28

said it about somebody going bankrupt, but I think it

19:30

was Gibbons who said the Roman Empire collapsed

19:33

slowly at at first and then

19:34

suddenly. And then

19:35

it's been applied to a lot of things. You're

19:38

right. Although Instagram seems to be

19:40

quite suddenly collapsing in

19:43

on

19:43

itself. But if I were wrong Totally. Well, again,

19:45

to you're not. And and I think Instagram

19:47

was one of those interesting ones where if

19:49

they had just stuck their guns like,

19:51

when they copied Snapshot,

19:54

that was brilliant because they did stories

19:56

better than Snapshot did. They had a

19:59

bigger audience and and they add us some features that made it better. So that

20:01

was like a perfect example of copy in the

20:03

right way. With TikTok, I think

20:05

they just have fundamentally misunderstood

20:07

their audience. They've misunderstood that

20:09

it's a completely different expectation.

20:11

And if they wanted to create a TikTok

20:13

compete, they should have created an

20:15

app called Instagram Reels that I bet would have

20:17

been very popular. But, you

20:20

know, by by loading it down with

20:22

stuff that people you don't follow,

20:24

you're not even necessarily interested in,

20:27

algorithm that is not as good as TikToks, and

20:29

then you don't even see, you know, your friend's photos,

20:31

the whole reason why people are there to begin TWiT. Yeah,

20:34

I I spend a lot less time on on

20:36

Instagram because I'm like, what's the point? I I

20:38

used to come come here for a specific

20:40

reason. Now, this isn't there. And

20:42

even

20:42

worse, it's a watered down version of this

20:44

other thing that already exists. But

20:46

we are creatures of habit, and

20:49

you're right, Tim, these things don't

20:51

die. But they don't exactly

20:53

thrive. There

20:54

will be something called Twitter ten years from now.

20:56

It's just not entirely clear whether anyone will go mad at all.

20:58

Am

20:58

I think

20:59

I believe there's still a friendster.

21:00

Uh-huh. By the way, I'm sorry.

21:03

Sunosi rises. It was Hemingway.

21:05

You win. I thought it

21:05

was You win

21:06

in Germany. I wish realized

21:09

wasn't? It was it was like a Fitzgerald quote.

21:11

But

21:11

maybe he still left from Gibbon. I

21:14

think Gibbon said it first, but I might be wrong on that

21:16

as well. How did you go bankrupt? Two

21:18

Gradually, then suddenly. Did

21:20

you collapse? Two ways,

21:23

gradually, then suddenly. This

21:27

actually runs into this there are tributories

21:29

off of this into a lot of the stories that

21:31

we're talking about these

21:34

days. And I don't wanna do another

21:36

Elon Musk filled TWiT so

21:38

we're not. Everybody's going, oh, thank god.

21:40

I thought he was gonna start talking about Elon. But,

21:43

really, it's about companies

21:45

in general going through this this

21:47

business cycle. Cement

21:49

your former employer, Tim Stevens,

21:51

has been accused of

21:54

some interesting shenanigans. We had

21:56

Conoco Yama on two weeks ago. Right when

21:58

this was breaking, remember the seventy five

22:01

stories and AI had written in their personal finance

22:04

section. She said, well, these are stories

22:06

no reporter wants to write.

22:08

You know, the kind of the basic boring

22:09

stories. We had the AI

22:11

write a

22:12

first draft and then an editor

22:14

look at it. TWiT correct

22:16

it, finish it, and then put it out.

22:19

But

22:19

now it's coming out that in fact there were far more

22:21

errors that were not That

22:23

a lot of the content wasn't very good.

22:26

And that perhaps cnet has been using it

22:28

more than just those seventy five

22:30

articles. She said, yeah. What we've used

22:32

for years, as many publications do, programs

22:35

to put in stock prices, that I

22:37

mean, I don't that's not using AI to write

22:39

a story. There's a very different thing

22:41

there. I don't I don't blame him for that. The verge

22:43

though has been really kinda hammering

22:46

on seeing it. I don't know. Maybe

22:48

they have a vested interest

22:50

in in in knocking

22:52

down a competitor. I I don't know.

22:54

But they're accusing CNN of

22:56

doing something a little bit more nefarious.

22:59

Remember, CNN, was sold

23:01

to equity capital

23:03

company called Red Ventures.

23:04

And, Tim, you have some probably direct

23:07

experience with

23:08

this? A bit. Yes, I do.

23:11

And what always happens with these

23:13

acquisitions is that the

23:15

equity capital companies raise

23:17

a lot of debt to

23:19

acquire these companies. So they're settled with big debt

23:21

and what you look at across this the

23:23

corporate landscape these days heavily

23:26

encumbered companies owning these companies. So a lot

23:28

of that. So there's pressure on them

23:30

from both their shareholders and their

23:33

and their lenders to kinda

23:34

monetize. So these companies very

23:37

often either

23:38

sell off pieces of

23:40

the company that they bought or

23:43

attempt to monetize it as Elon is

23:45

doing with Twitter. The verge

23:47

is accusing CNN

23:49

and Red Ventures of of and the way,

23:51

Red Ventures also owns

23:53

a a number of sites like the Points

23:56

Guy, bank rate, and credit

23:58

cards dot com. Which

24:00

are sites that make their money through affiliate

24:02

credit card affiliate fees,

24:04

and the verges accusing them in

24:07

effect of turning cnet into

24:09

that kind of site auto

24:12

generated link bait articles,

24:16

designer rank highly in searches,

24:18

that they can then monetize with ads or

24:21

affiliate

24:21

fees. Bank rate and credit cards have

24:24

also published AI written articles about credit

24:26

cards

24:26

with ads. For credit cards nestled within.

24:29

It turns out the same guy responsible for

24:31

this at bank rate and credit cards is responsible

24:33

for it at

24:35

cnet. Lance Davis,

24:37

Vice President of Content at Red

24:40

Ventures. And I think there's

24:42

an interesting accusation here that

24:44

Red Ventures is basically taking

24:46

this venerable highly

24:48

respected name in Technology

24:50

Journalism and turning it into an SEO

24:52

farm. Tim, I'll give you

24:55

the chance to mention. Either they

24:57

recuse yourself or

24:59

to

25:00

give us your thoughts.

25:02

Yeah. No pressure. Obviously, I need to be able

25:04

to be careful with what I say here, be

25:06

both because this is my former employer we're talking about and because

25:08

I have a lot of friends and a lot of people with my

25:10

respect. Well, that's really important. And I and I should say

25:12

that. Important to see. So many people, including

25:14

Connie, that I love and respect and

25:17

honor. I don't blame seeing it

25:19

for this one. I think this comes from

25:21

Red Ventures.

25:22

Well, so my my take on

25:24

this is a little bit complicated. I

25:26

I do think that you know,

25:28

clearly, the verge has

25:30

an interest in in making seeing it look bad their

25:32

competitors. That's fine. I don't think

25:34

think that anything that the verge has reported thus far from what

25:36

I've seen has been inaccurate. I I wanna

25:39

say one thing for sure. I

25:41

wasn't aware of any AI stuff that was going on

25:43

when I was there. Left seen at around August of

25:45

last year. There was kind of rumors

25:47

and talk and that kind of thing, but I wasn't aware

25:49

of anything going

25:49

on. So I have no insider knowledge about

25:51

how any of this came to pass. But I I will say that

25:54

Siena was using tools like

25:56

Wirtieth and others. And

25:58

those are tools that a lot of outlets

26:00

use of publications use those. And basically, what they do is they help

26:02

you optimize the content that you're running to make sure

26:04

that they include the right keywords, to

26:06

make sure that they are

26:09

you know, that they perform well in an algorithm

26:11

based environment. And that is really what

26:14

consumers are operating within right now. Anyone who goes

26:16

on the Internet and searches for a thing, is

26:18

asking an algorithm what things should I read.

26:20

And so it's only natural for

26:22

publicationists to wanna make sure that their content

26:24

performs as well as

26:26

possible. The thing is when you use a tool like that, it can begin

26:28

to feel like you are basically

26:30

reverse engineering

26:31

Google, you reverse engineering a

26:34

search engine. That's really what this game comes

26:36

down to. Isn't it?

26:36

Talking the AI, isn't

26:38

it? Right. Right.

26:39

And that's what we're talking about for Leo. At

26:42

some point, you know, what's the best tool to optimize

26:44

cognitive algorithm? It would be another

26:46

algorithm effectively. And so I I

26:48

think by extension, it's a natural thing that

26:50

seen it would do

26:51

this. I don't think anybody would be surprised at seeing

26:53

that -- Well, see that one

26:54

of the -- -- when there's financial pressure

26:56

to turn around a big acquisition. Right?

26:58

Maybe so, I think that the timing is

27:01

a little bit irrelevant here. I mean, CNN has definitely

27:03

been on the cutting edge of a lot of different

27:05

publication types over the years. What would

27:07

there be integrated affiliate

27:09

links, things like that. They've definitely been at

27:11

the bleeding edge. So there's no surprise that they would be

27:13

at the bleeding edge of adopting AI technology.

27:15

My concern really is that there wasn't

27:17

enough transparency involved here. I think that's what what

27:19

my problem is. If seen that had come out

27:21

and said, hey, we're experimenting with AI. This is kind

27:23

of fun and new. We don't really know what this is gonna be,

27:26

but here's what this is where we're trying it on. This

27:28

is some content that was written by AI.

27:30

What do you think? I think that

27:32

this would have been I'm sure that they would have

27:34

gotten some blowback for sure. But

27:36

from what I could see, from my perspective in

27:38

reading through coverage on the virgin and

27:41

elsewhere, it just seemed like it was they

27:43

were hoping that nobody would notice. And

27:45

I feel like that's really the wrong way to go about

27:47

doing this. If you're going to be embracing

27:49

this kind of technology or investing in it, especially

27:51

when you're talking about giving people recommendations

27:54

where they should put their money in a mortgage,

27:56

I think it's important to be incredibly transparent.

27:58

And, you know, Connie's piece was

28:00

very transparent, but that came

28:03

out long after the story had had kind of blown up, long after the

28:05

Virtus piece. And I think it's unfortunate that

28:07

the scene that wasn't more upfront with

28:11

with what was going on. To be honest, there was certainly you know, we saw the

28:13

little disclaimers on Google and things like that, but

28:15

that was in my

28:16

opinion, it was not enough in that that's where

28:18

I'm disappointed in this whole thing.

28:20

Yeah. And

28:20

don't blame Connie at all or

28:23

even Lindsay Turgeon, also who's been a regular

28:25

on this show for many years. I think I

28:27

have huge respect for both of them.

28:30

Yeah. If anything, I feel like they might have been

28:32

sandbagged by this, and they

28:34

didn't know the full extent of what was going

28:36

on and ended up being,

28:38

you know, kind of

28:40

hung out to dry, so to

28:42

speak. The verge quotes of former senior

28:44

employees saying Red Ventures was using

28:46

automated technology for content

28:49

long before the AI by line began

28:51

cropping up in November. They

28:53

mentioned this word Smith tool, which you

28:55

talked about, Tim, nicknamed Morgotron

28:58

or Morgot I don't know how you

29:00

pronounce that. Morgotron internally because if it's

29:02

used in mortgage stories, They

29:05

said it'd been used for at least a year and a half,

29:07

but the siloed natures of

29:09

the teams across CNN and Red Ventures

29:11

makes it difficult. For journalists at

29:13

the site to understand the chain of command who's

29:16

using what tools and when. So

29:18

I no blame on our

29:20

friends at cnet. I'm

29:23

I'm very happy, frankly,

29:25

to blame Red

29:27

Ventures in any equity capital

29:29

company because I feel like these guys or

29:31

to some degree the bane of the

29:33

ex of our

29:34

existence. But it it's not just it's not just

29:37

VC firms that that are pushing

29:39

companies to use this content. A lot of editorial

29:41

properties use SEO optimization tools. If

29:43

if you wanna perform, if you wanna be in the

29:45

first page on a Google search, you have to be

29:47

using these tools I know a lot of automotive properties are

29:49

using

29:49

them. This is not,

29:50

you know, proprietary software. This is stuff that you

29:52

can go out and license that anybody can

29:55

use Now tell you what keywords that you need to inject into your content.

29:57

And, again, it it it does make you feel like

29:59

your reverse engineering is your writing, but

30:01

this is not proprietary

30:03

stuff. Yeah. Well, in a way,

30:05

that's scarier if it's if it's even more

30:07

widespread use that we don't know about. I

30:09

mean, I

30:09

think they're but for the grace of god goes everybody

30:11

in the media business, not us.

30:14

We haven't figured out a way

30:16

to do that yet with podcasts. Over the

30:18

course not

30:19

immediately, but over the course of time, I I think

30:21

you will see AI play a

30:24

role a lot more particularly

30:26

as some of the

30:28

issues seen that ran into are less

30:30

of an issue. And also, I mean,

30:32

they seen that probably made a lot of mistakes so

30:34

the rest of us don't have to and can

30:36

learn from them in terms of

30:38

disclosure. But I

30:40

feel like, well, we're not doing any

30:42

of this and have no plans to do this. And in

30:44

fact, might not really work well for

30:45

us. Anyhow, I I would not

30:48

say that

30:49

At best company, we'll never use

30:51

AI in any form because I I think things are gonna

30:53

happen quite click quickly, and there might be ways to

30:55

use

30:55

that, which are actually

30:58

completely above board and reasonable and resolve

31:00

in in better content rather than just

31:03

cheaper

31:03

content? Somebody

31:07

said I'm trying to find the article

31:09

that CHAT GPT is

31:11

the absolute definition

31:13

of BS. Yes. Because

31:16

and and check and by the way, OpenAI, the

31:18

creators of QTT GPT say

31:20

this. They said we never said you had to be

31:22

accurate. That's not in the training at all.

31:24

It

31:24

has no idea what it's saying and whether it's

31:27

correct or not. Some sometimes TWiT happens

31:29

to be accurate, but that's not what the type

31:30

of incident accident if it happens to be

31:33

accurate almost. Right? I

31:34

did a piece. I I have a new newsletter,

31:37

which I should plug at the end of the call.

31:38

Plug it now. It's called plugged in,

31:41

and you go to our Fast Company homepage, there should be a

31:43

newsletter's link that will let you subscribe. Nice.

31:45

And I because I'm interested in

31:47

the history of cartoons, I chat

31:49

GPT what the first TV cartoon was. And every

31:51

time I asked, it would give a different

31:54

answer. Many of them very convincing, and

31:56

none of them correct. There

31:58

basically, there's so many things where ChatGPT

32:01

has no idea what it's saying and unless you

32:03

already know what the answer

32:04

is, you you might well be fooled because

32:06

TWiT is able to lie in such a con convincing

32:09

fashion. But it's important to understand that that's

32:11

not its mandate to tell the truth or to be

32:12

accurate. It's not a It's not

32:14

a fact generator. It's a BS generator. It's really good

32:17

at

32:17

straining words together. You

32:19

call it a glib

32:19

bot, which I think is a very

32:22

hear you come from now on. I'm calling it a grip butt.

32:25

So in

32:28

a way then, it

32:31

makes you wonder, should we

32:33

be, you know, your company

32:35

GitHub and you can you can disclaim

32:37

this. Again, I know you have nothing

32:39

to do with scrutiny. But is getting a little heat right now

32:41

from the open source community over its AI

32:44

code generator, copilot, which

32:47

is Kind of impressive. Copilot also

32:49

uses, we should mention, the same open

32:51

AI technology as chat

32:52

GPT. It is using GPT.

32:55

Yeah. I was gonna say, I I can't comment

32:57

on any of the the lawsuits or any of that stuff,

32:59

but co pilot does use the the

33:01

GPT33 dot five

33:03

you know, a language

33:06

model that chat GPT is based on. It

33:08

uses something called codecs, which is specifically

33:10

focused on source code rather than

33:12

you know, the the corpus that the chat

33:15

GPT uses, which is much more broad.

33:17

But if you use chat GPT to

33:19

say, right, you know, program that does

33:20

this, this, and this, Most of it

33:22

is data set is probably coming from. Yeah.

33:24

Because chat, GPT, can write code.

33:26

In fact, one of the stories we had on

33:28

security now is that Scripted

33:31

kitties are having chat,

33:34

GPT, right, effective

33:36

malware. My malware

33:40

that works

33:40

-- Mhmm. -- we

33:43

know somebody who used

33:45

chat, GPT, to write a

33:47

PowerShell script for Steve on SecurityNow that

33:50

looked through your last past vault and

33:52

told you some of its attributes. And

33:55

it worked and it was a lot easier to

33:57

develop it because and I

33:59

guess co co pilot even better.

34:01

Now clearly, with co

34:03

pilot, unlike chat GPT, there must be some rules

34:05

in there to

34:05

say, oh, and by the way, make sure this isn't made

34:08

up that it actually

34:10

works. Right?

34:11

For the most part, I mean, there are

34:14

suggestions that you can get that will not run,

34:16

so it is – that's why we call it

34:18

copilot. TWiT not do it for

34:20

you. It's your co pilot. It's, you know, autofundal and suggestions, you know,

34:22

plus one. Right? So And

34:26

the more that you use it, the more that it gets to know your code, it does to

34:28

know kind of your style and

34:30

your intent, and it can give you better and

34:32

better suggestions for what you're doing.

34:35

But no, you can absolutely the same way, you

34:37

know, you could get a wrong, you know,

34:39

suggestion, you could get, you know, or

34:41

or or a wrong, I guess, paragraph from

34:43

chat GPT, you could get some incorrect code suggestions. For the most part

34:45

though, I think that the the

34:47

training model there is

34:49

a little bit better because it is is, you

34:52

know, focused more on on one thing

34:54

rather than, you know, however great the

34:56

corpus is. For for

34:58

everything that Chap GPD is

35:00

doing. And as I said, it is learning based

35:02

on your own style and and the

35:04

stuff that is is in your

35:06

project folder. But no, I mean, this is why I

35:08

always tell people, look, co pilot is

35:10

amazing that it has saved me so much time,

35:12

especially with

35:14

boilerplate stuff. But if you're trying to use it to just you think you can

35:16

just automate it to write a program for you,

35:18

you might get lucky if it's something really

35:20

simple, like a a PowerShell script or something

35:22

like that. But

35:24

you you really need to have

35:26

AAA better idea of what you're doing so

35:28

that you can actually see what code it's

35:30

suggestion suggesting and then make

35:33

edits if that needs to be the case. But even if you still need to make

35:35

edits, I think there's still value there because it

35:37

can save you, you know, a

35:40

a lot of time of of having to,

35:42

you know, manually Google and

35:43

and, you know, command c, command v from from

35:46

Stack Overflow or or wherever. Yeah. People Well,

35:49

and that's it. Every programmer knows this, but maybe

35:51

a lot of civilians don't. That almost all

35:54

code is to some degree or

35:56

another copy to paste from

35:58

somebody

35:58

else. That's

36:00

kinda how it works. So copilot is a

36:02

natural way to do this. Copilot's quite impressive.

36:04

It's quite amazing. Here's the story

36:06

from earlier this month by checkpoint

36:10

research a malware research company.

36:12

They call it OPPON

36:15

AI, cyber criminal started to

36:17

use CHAT GPT. In Check Point's

36:20

research, previous blog

36:22

we described how CHAT GPT successfully

36:24

conducted a full infection flow.

36:27

From creating a convincing spearfishing

36:30

email to running a reverse

36:32

shell capable accepting commands

36:34

in

36:34

English. That's pretty scary.

36:38

This

36:38

is a case of

36:41

something called Infostealer. Which

36:44

was created the late last year by Chat GPT. A

36:47

cyber criminal showing how

36:49

he used Chatbeat GPT to

36:51

write the code Looks

36:53

like JavaScript. A hard

36:56

code. It it had a great

36:58

code to basically steal

37:00

files. From a

37:03

FTP server. It's

37:05

it's kind of amazing

37:07

what they're doing. One of the things that

37:09

really becomes obvious is this is a conversation a

37:12

year ago we might not have had.

37:13

This has happened all of a sudden

37:15

out of

37:16

nowhere. And you can imagine we're not that far away from

37:18

these areas being able to emulate I

37:20

mean, they can already do very compelling

37:24

So how far are they from being to

37:26

emulate your voice, your mom's

37:28

voice, and make up a

37:29

call, and then say, hey, you

37:31

know, it's it's your mom, I forgot my

37:33

password. Can you can you tell me Oh, yeah. that's already I'm sure

37:35

that's already it. That should be

37:37

doable

37:37

right now. Yeah. Yeah.

37:40

There is a generative

37:42

AI music already. It's not I

37:44

don't think it's quite there yet. This is a paper

37:47

from Google Research. They call it music l

37:49

m. It's based on language model like

37:51

Lambda, generating music from

37:53

a text prompt.

37:56

Yep. Here is

37:58

the main soundtrack of this

38:00

is the prop, the main soundtrack of an arcade

38:02

game. It is fast paced. And

38:05

upbeat. didn't check my audio. Do

38:08

you do you I think we'll we'll try it.

38:10

Turn my audio on. I wanna play this

38:12

song. It's fast paced and upbeat with

38:14

a catchy Electric guitar

38:16

riff. The music is repetitive and easy to

38:18

remember TWiT unexpected sounds like symbol crashes

38:20

or drumrolls. Does this sound like

38:22

an arcade game

38:23

to you? Maybe the

38:26

front screen is batch floater.

38:28

Maybe Sonic is running

38:32

down

38:32

That's completely AI

38:36

generated.

38:37

Although, internally, it's generated by

38:40

an AI

38:42

that there's a

38:42

fair amount of plagiarizing, which is why -- Oh, yeah. --

38:44

Google is

38:45

not very releasing to

38:46

this. Yeah. It's all

38:47

about that. Totally flavor. Here's

38:50

a slow tempo, bass and drums led reggae

38:53

song. Yeah,

38:58

man.

38:58

Everybody get together. We're going down to the beach. No.

39:00

No. Ant says no to that one.

39:05

Seems like it

39:06

as the potential to blow away the stock music industry

39:08

pretty quickly. But, yeah, it's a lot

39:10

better than the crap stock music we we

39:13

we have using. I'd like to introduce me as a you you'll be

39:15

able to generate

39:16

something unique to your own two hundred

39:18

and

39:18

eighty thousand hours of real music is

39:21

the training model. To

39:23

generate coherent songs for

39:26

descriptions of significant complexity

39:28

as the creators put it You wanna

39:30

you wanna feel like you're lost in space ant

39:32

and is becoming our our

39:34

taste

39:35

tester. Let's see if ant

39:37

agrees this is.

39:41

Sounds like an AI

39:43

did it, doesn't it? Sounds

39:45

like robot music.

39:50

Now here's the

39:50

question. Can we get taken down from YouTube

39:53

for playing that?

39:55

Make up sued

39:57

by a bot. See,

39:58

that's No. That's well, that's

40:00

gonna be an interesting thing. I think actually

40:02

Who wants to be

40:03

able to generate these unique things? Right.

40:06

Well, that that's an interesting question. But also,

40:08

I think becomes a very interesting question, which you

40:10

know, I think that this

40:12

YouTube relies

40:14

on someone else being able to say, I

40:16

have copyright of this and

40:18

and usually have, like, a a file registered

40:20

in place. They're, you know,

40:22

a a content ID can go and

40:24

find the same thing. But if it's

40:26

uniquely original

40:27

file, then Continental is not gonna find it. So

40:30

that's What a world we live?

40:32

That's

40:32

cool. What a world

40:35

But do you think there might be some cool

40:37

stuff that might happen if actual

40:39

human musicians work with some of

40:41

these tools to brainstorm

40:43

and Yes.

40:44

And riff on ideas. And then it seems like that could be kinda cool.

40:46

Yeah. She was

40:47

having a real

40:47

time. No. Exactly. I mean, I honestly, I think

40:50

that the way that and I know that a lot of

40:52

creators are really freaking out generative

40:54

art and and generative music and

40:56

and all this stuff. And and I understand the

40:58

fear. But for me, what excites me

41:01

about this is that the best AI

41:03

art that I've seen has been

41:06

from actual artists. Like, those are

41:08

the people who've been using the best prompts or have

41:10

been taken some of the prompts and have taken some

41:12

of the results and then made

41:14

really great things. And I think with music,

41:16

it's the exact same way. Right? Like, you might

41:18

be able to get something that sounds

41:20

slightly better than than than

41:22

stock music. TWiT it's still not going to be great. Right? It's going

41:24

to take a real artist to

41:26

then take that and edit it and

41:28

interpolate it and do what real artists have

41:30

always done and turn something

41:32

else. And and so the what

41:34

I've tried to been trying to tell people because

41:36

this isn't going away. This whatever your

41:38

feelings on on this stuff is, it's

41:41

not going away, and it's only going to become bigger.

41:43

We can have conversations about ethics and we

41:45

should. We can have conversations about safety

41:47

rails and we should this is not going

41:49

away. And so what I've been conversation I've been having with people for the last

41:52

year or so is, like, embrace

41:54

this as a tool to

41:56

your arsenal to

41:58

make new unique and better things rather than looking at

42:00

this as some sort of existential threat because

42:03

you're not going to outpace

42:06

this. This is not going to be something that you can get away from,

42:08

but it might be something that if you

42:10

are able to use, you

42:13

could actually enhance you know, this stuff

42:15

that you you for writers as

42:16

well. Last week, Brianna Wu was on

42:19

the show, her husband writes science

42:21

fiction among many other

42:23

things said that Frank

42:24

was stuck with a story that he was

42:26

I think he was ready for analog, but he was stuck

42:29

with a story. And he gave

42:31

a very extensive prompt to chat GPT, which wrote

42:33

kind of a mediocre

42:35

story, but came up with a lot of things

42:37

that became a starting point for him

42:39

and unstuck him. And that seems

42:41

like that's a very good use of something like chat,

42:44

GPT. I've heard so many

42:46

descriptions I love

42:48

I love

42:49

I love your name for it. What is it?

42:51

Glib PT? Glibot.

42:53

Yeah. Glibot and be like, remember I wrote

42:56

that? Yeah. That's good. Yeah.

42:58

That's good. I've also heard it say the ultimate man

43:00

splinter because it because

43:02

it's confidently wrong. Right?

43:04

And it says and it's

43:06

So and it's a little

43:07

patronizing. It's like, oh, no. Let me explain to you how the world works. Although if you tell

43:10

it that it's wrong,

43:10

then it gets really humble and

43:14

and apologize.

43:15

Apologize as

43:15

a great length and says it'll never do it again. Does

43:18

it

43:18

correct itself? If you correct it, does it state correct?

43:20

Yes. In

43:20

fact, some if it says something that's correct,

43:22

you tell it that it's wrong?

43:24

I will apologize for

43:25

that too. Stephen Wolfram wrote a very

43:27

good piece about how

43:30

confidently wrong chat

43:32

GPT is on things that, well, from Alpha, his his own

43:35

kind of AI. Is it an AI? I don't

43:37

know what you call, from Alpha. Search engine

43:39

for knowledge or something.

43:42

But he said if we should partner because we

43:44

we're good at getting the math right,

43:46

which chat, GPT, is terrible. And

43:50

then if we worked if we worked together, we maybe get something

43:52

out of it. He pointed out some

43:55

really historical examples. This is

43:57

his article from his blog at

43:59

steven wolfram dot com. Some hysterical

44:02

examples of just chat TPT

44:04

getting it terribly wrong. How far

44:06

is Chicago from Tokyo,

44:08

to which Chat GPT confidently

44:10

says the distance from

44:11

Chicago, Illinois to Tokyo, Japan is

44:14

approximately seventy six hundred

44:15

miles That'd be twelve thousand two hundred kilometers. It's

44:17

a very long distance, blah blah blah.

44:19

Turns out it's not even

44:20

close. It's six thousand three hundred thirteen

44:24

miles. So You correct it. So you

44:26

you

44:26

you tell it and

44:27

it says thank you for correcting

44:29

me. You're correct. Two,

44:31

of course, is the distance

44:32

is six thousand three hundred

44:35

thirteen miles. How far is

44:37

Chicago to Tokyo, and then it

44:39

gets it

44:40

right. At least in that continued

44:42

conversation. I think that's interesting. And

44:44

but kids don't don't do your

44:47

math homework with chat GPT

44:50

stick to wolfram alpha because it doesn't even

44:52

know three to the power of seventy

44:53

three, which is

44:56

pretty pathetic. By

44:56

the

44:56

way, not even close. It

44:59

said fourteen

45:00

billion. I can't say how big

45:02

the number is. It's a lot larger.

45:05

There's that story about Jack GPT

45:07

passing an MBA exam. Yeah. But but

45:09

the article, which said TWiT also pointed out that

45:11

it wasn't capable of

45:12

doing, like, high school math, which

45:14

I sound interesting because I'm I was so many

45:15

MBAs can't do. Right. We didn't realize

45:18

you could become an MBA without having high

45:20

school

45:21

math, but I

45:24

think it just passed a lot of school exam too, didn't it? This is now

45:26

the new thing. It's for professors to give

45:28

their exams to chat

45:31

EPT.

45:32

There was there there was a paper that that a couple of the

45:35

it was from University of Chicago when someone

45:37

else did with a GPT passing the

45:39

bar, and they gave

45:42

it part of the multiple choice, parts of the bar exam, and

45:44

it did better than random selection,

45:46

and it came close to humans

45:49

in couple of categories. It got a c plus. Passing,

45:52

but it's not like TWiT right. It didn't quite

45:54

pass. But it is it did but it is impressive

45:56

because the interesting thing though was that

45:58

it did significantly better than random selection. Like, it

46:01

wasn't one of those things where, you know,

46:03

you're just randomly, you know, okay. How would you've

46:05

done if you were just randomly selecting

46:07

the answer? So TWiT had some, you

46:09

know, better accuracy. And in some

46:12

categories, it was close to

46:14

humans. But Obviously, this is only for multiple

46:16

choice parts, and it did better in certain

46:18

areas than others. But, I

46:20

mean, to me, all this really says

46:22

is, okay, then you if your

46:25

big concern whether it's high school students or or,

46:27

you know, graduate students, you know,

46:29

and and professional taking tests, if your big

46:31

concern is the AI

46:34

cheating at the test, well, then you need to start changing how

46:36

you're testing. You're obviously not testing the right things.

46:38

Like, that to me is the big takeaway. And

46:40

we shouldn't be freaked out that

46:42

you know, these these AIs are able to pass the test. It's more

46:44

like, okay, well, what's the goal of this? And

46:46

are we testing the right way? And I think in most

46:49

cases, the answer would be

46:50

no, we're not testing the

46:52

right way. Howard Bauchner: Yeah, maybe that's the flaw of the

46:55

tests. Although, as you point out, Chad, GPT doesn't

46:58

do math.

46:59

Very well. It's good in constitutional law.

47:02

How long do we think

47:03

it'll be until Chad GBT is our

47:05

public defender and that you need to pay extra if

47:07

you want to a human to to defend

47:09

you in a in a

47:10

lawsuit? No. I don't think that's gonna

47:13

happen. The guy

47:13

who seems like a a black bureau. So the guy

47:15

who was doing the robot lawyer, I think,

47:18

is just decided to hit run run away with his tail

47:20

between his legs because then so

47:22

do not pay if you could A very actually,

47:26

a really cool service, which helped you get out of traffic tickets, created

47:28

an AI powered robot

47:30

lawyer that was gonna go

47:32

into court

47:34

I don't know, you know, first of all, I think any judge that

47:36

would throw it out immediately was gonna go

47:38

into court to help fight a traffic ticket

47:41

State Barr prosecutors threatened

47:44

the Joshua Broder as the CEO

47:46

of do not pay with jail

47:49

time. And so

47:51

Joshua says, we're postponing

47:53

our Caucasian. We're gonna stick

47:55

to consumer rights.

47:58

Wow. Okay. Oh, totally. Well, that's well well, this is the whole

48:00

thing. Right? It's, like, couldn't maybe, but,

48:02

like, do you do you think that there

48:05

if if any profession Can

48:07

you think of any class of profession who would

48:09

be less likely to allow

48:11

this into and and

48:13

therefore No. Like, like, the

48:16

even if you could potentially automate

48:18

things and and do things better than, like, your

48:20

typical public defender. Do you really think

48:22

that the the, you

48:25

know, the borrower associate and and and the

48:27

various lobbying groups for for lawyers. You really think that they would

48:29

allow this and their programs absolutely not. You're

48:31

gonna protect their

48:34

own interests above and beyond more than any other industry. They're gonna be the

48:36

ones who are

48:36

like, nope. Not not happening. Yeah. That's a good

48:39

point. If you're gonna pick an an

48:41

industry to disintermediate, do

48:44

podcasters. Don't do lawyers. You know, we're we're put

48:46

shelters. It'd be a lot easier

48:49

to go after us. Your

48:54

company, Microsoft, just to

48:56

acknowledge that they're putting in a they

48:58

already put a billion dollars

48:59

in. They were one of the

49:01

founders of

49:02

OpenAI. And now they have AAA even better deal

49:04

with OpenAI. The rumor was ten additional

49:06

ten billion dollars. I think that

49:08

was confirmed by Sachin Adella. Over

49:11

a period of time, obviously. And that

49:14

chat, GPT, or that kind of

49:16

technology, will be used in Microsoft Office.

49:19

But I think number of people are saying the

49:22

real the real thing to watch

49:24

is

49:25

Bing. Mhmm. Thoughts

49:27

about that. I know Yeah.

49:29

And you work at cop you

49:31

work at GitHub. So so, you know, it's

49:33

just owned by Microsoft. Opportunity on

49:35

my own. Look, I think this is exciting. I think that, you

49:37

know, there's also been reporting that the Google's

49:39

been having kind of like AAA

49:42

crisis about how successful chat GPT has been. And I I

49:44

don't blame them because Google

49:48

has amassed and this is not

49:50

in any way to try to integrate any other

49:52

company. But but they have probably

49:54

mass, like, the largest quantity of

49:56

of AI talent

49:58

from Mac cademia and and from industry of anyone.

50:00

And the fact that it was

50:02

TWiT which was interesting to me about

50:04

that, is that It wasn't really that demonstrably different from

50:06

any of the other GPT three things that

50:08

have been available. It was just the interface that

50:11

I think made it so accessible. Has

50:14

become this very mainstream

50:16

thing where, you know, I've been thinking and

50:18

I've talked been talking about, you know,

50:20

OpenAI stuff for

50:22

several years, But now this is a mainstream thing because the interface was ripened.

50:24

And, yeah, I I definitely

50:26

think that search is a great area where

50:29

it could be helpful people have

50:31

created extensions to add, you know, chat, GPT things

50:34

alongside Google results, and it's better. And

50:36

and I think that, you know,

50:38

Google results Google

50:40

is the primary research engine that I that I use. And the

50:42

results have gotten worse over time, and

50:45

and I don't think that

50:48

it's because of the SEO stuff, I think,

50:50

is because Google has

50:52

optimized for different sorts of results,

50:54

and they've you know, wanted to

50:56

highlight other things. And so I

50:58

often end up piping Reddit into my

51:00

search -- Yeah. -- because I find that I get much

51:02

better results

51:03

Mhmm. Yes. From from Reddit than I do.

51:05

Because

51:05

those are because what you're really doing is

51:07

asking for information from real experts

51:09

about a

51:11

topic. Right? Right. Know, I just wanna actually

51:13

get the conversation, like, the info where it's actually

51:14

going to be. But searching Reddit dot com is

51:17

is a mess. So searching Google for query

51:19

and then adding Reddit to TWiT. alternative.

51:22

But people have created, you know, like,

51:24

kind of side by side extensions

51:26

to add, you know, Cheggi BT stuff.

51:29

To Google things. And I think that, yeah,

51:31

this is an opportunity for Bing. I think

51:33

it's an opportunity for a lot

51:35

of consumer products. Obviously, one

51:37

of the big wins here is for

51:39

for Azure for, you know, other businesses

51:41

who want to take advantage of those models and build

51:43

it into their their products having kind

51:46

of AI as a service. Think, you

51:48

know, look, this is to be hot,

51:50

everybody's going to be this is

51:52

going to become an arms race. Right?

51:54

Even more than it already has been.

51:56

But for whatever reason, you know, OpenAI has

51:59

been the first to really commercialize

52:02

this in a way that the mainstream understands. And

52:06

it's exciting. I mean, personally as

52:10

a technologist, to me. All all the other kind of fears we

52:12

would have around it. Like, I look at this as

52:14

a moment of this is this is exciting.

52:16

Like, to me, this is much more

52:18

than expecting versus the

52:19

metaverse. Like, this is much more exciting --

52:21

Great. -- and it's much more tangible as

52:23

to what the next big place of computing

52:25

is going to be. Forget about the vice

52:27

stuff is is I think really

52:30

what's exciting. Open AI is less to

52:32

lose than

52:32

a Google or a Microsoft. Well, that's

52:34

why

52:35

Open AI was created, really. Right? Yeah. These are

52:37

enormous companies with an enormous customer base

52:39

-- So -- and secret and

52:41

reputations and and paying

52:44

customers and open AI

52:46

not

52:46

having any of that

52:47

stuff. Why why not throw it out into the public and

52:50

see what happens? Although, Jan

52:52

Lucone, who does a is the

52:54

genius AI researcher at

52:56

met at

52:58

Facebook, said that,

53:00

oh, chat GPT isn't particularly

53:03

innovative. We've been doing that for

53:04

years. No.

53:05

I I think if you're AI scientist, you

53:07

know, about transformers. Yeah. Yeah. Which were

53:09

and Google, basically,

53:09

like, transformer technology. Right. This is

53:12

this is what Lambda

53:13

did. Yes. They did. Yeah.

53:15

And that has done some cool stuff with him

53:17

too. But

53:17

it also feels a little

53:18

bit, like, south of the

53:19

apes. Right? It's like, oh, no. We did. I

53:22

was

53:23

just gonna say. Sure you have, but you didn't productize it. I

53:25

didn't tell anybody. No. Like

53:26

right. You didn't you didn't productize it. Like, I don't

53:28

think that anybody would make the argument. I

53:31

don't think Sam Altman or do you want

53:33

from open AI? AI would be like, oh, this is the most innovative thing and no

53:35

one else has done this. I think what

53:37

they would say is,

53:39

this is the first time that the public has actually been able

53:41

to interact with it in a way that had a

53:43

really good user interface. That's what

53:44

I thought. It was

53:45

a really user interface. That's what

53:47

looked like he said. He said Chad GPT is, quote, well put together.

53:49

He said that compared to other companies in

53:51

the field, OpenAI is

53:53

not particularly

53:55

advanced Google meta, and

53:57

he said half a dozen other startups have

53:59

equivalent technologies. But

54:04

That's

54:04

the difference. They were doing this in public and letting the

54:06

public use

54:07

it. Although it makes me

54:09

wonder, is there something better under

54:11

the hood somewhere else. Well, GPT

54:13

four apparently is an enormous advance over

54:15

three point five. With Sam Alvin, CEO

54:17

of opening eyes. Don't get your hopes

54:19

up. It's not it's not it's not AGI. Right?

54:22

It's not the general --

54:23

Right. -- general intel. That I actually did

54:25

put out a AI chatbot a a

54:27

few months ago. And they

54:30

immediately got flagged for

54:31

it. Did they get racist instantly? Being

54:33

racist and anti Semitic and so forth. So

54:35

and so they they

54:37

were they tried to be bold, but they weren't quite as bold

54:39

as check DPT, so they didn't get as much credit, and they

54:41

got a lot more flack about it. I

54:43

think partially because Metabolic Under

54:46

Companies is gonna get flagged no matter what it does, which

54:48

is not true of what an AI at least

54:49

yet. I've been using a search

54:52

engine that was founded about five years

54:54

ago by former Google search executives called

54:56

Neva, NEEVA. Are you

54:58

familiar with this? I read a big story on it. Oh, it's

55:00

like that's how I learned about it. Yeah. The

55:02

CEO is

55:04

former

55:04

top guy at

55:05

YouTube. Yeah. And

55:07

he got a little

55:08

bit depressed about the

55:12

monetization of search

55:13

The certification of Google. So we

55:14

we we went off to do a a search

55:16

engine with a a paid model. And, yeah, that's the

55:18

premise. We don't run ads. We don't in

55:22

fact, even when Google started, Larry Page famously wrote,

55:24

a search engine can't have

55:26

advertising or will then become

55:28

beholden to the advertisers they

55:32

only held that off for a few years

55:34

before getting involved

55:36

in advertising. So I pay five

55:38

bucks a month for Neva.

55:40

I get a lot of they actually give

55:42

you a free one password account and and other stuff, but I

55:44

think it's really good. And also,

55:48

because, you know, they're in this arms race.

55:50

They added a AI generator

55:52

at the beginning of search results.

55:54

So I searched for chat GPT

55:57

and Bing just now. And this is the result I got from the

55:59

AI. I think AI's are very good

56:01

at synopsizing and summarizing other

56:04

content. So They even do

56:06

footnotes to say where this information

56:08

come comes from cnet, the Guardian

56:10

Observer, and the Virgin. Microsoft is

56:12

reportedly integrating

56:14

ADI technology such as chat GPT into its Bing search

56:16

engine, which could potentially revolutionize

56:18

search as we know at this technology is capable of

56:20

generating a wide variety of text and

56:22

human

56:23

like ways and sponsored written prompts. Microsoft hopes to

56:25

launch this feature before the end of

56:28

March. In a bid to make

56:29

being more competitive with Google, I think

56:31

Google should be scared. Not

56:34

just by bandwidth by Nava. I think this is pretty cool. Nava,

56:37

the only I've been using Nava

56:39

full time instead of Google everywhere,

56:41

including on my

56:44

iPhone, for about a month now. The only

56:46

negative, the only hit on it is

56:48

it's amazing how quickly Google comes

56:50

back with a result. Nava,

56:52

there's a palpable second or two. But other than

56:55

that, the results are excellent. I love

56:57

this AI thing and

56:59

there's no

57:00

ads. TWiT doesn't favor

57:02

Google content over anybody else's

57:06

content.

57:06

I think there I believe there's

57:08

a dash of Bing in nivo's technology

57:10

along with some of its own technology

57:12

as well, is it interesting? I believe they've licensed some data.

57:13

Okay. Do they have

57:15

their

57:15

own crawler? Right? Yes.

57:18

I think

57:18

they they kind of mashed together some of their

57:21

own stuff and some stuff they've licensed. And

57:23

I don't know what this thing is, but

57:25

it's pretty cool. There's a little slider here

57:27

at the top. I don't currently showing

57:29

top news from all sources. Currently

57:32

showing top news from all

57:34

sources. I don't know. There's some something going

57:36

on there. That I can move around that

57:38

slider. There's it's I think it's very innovative. I I'm, you know,

57:40

I have no relationship with them. In fact,

57:43

I meant to ask you about this because I did read your

57:46

article about it. You talked to them. You think they're

57:48

pretty compelling. They're smart

57:50

folks.

57:50

They've added a lot of stuff since they

57:52

launched. It's great to say we're gonna go against

57:54

Google. Yeah. I mean,

57:55

they're a tiny company. But maybe,

57:57

you

57:57

know, now. We've

57:59

all been used to getting our search free for the

58:01

last twenty years. But I think if

58:03

there is if there is a time when we're at

58:05

an inflection point, or the idea of

58:07

going up against Google no longer sounds

58:10

quite so insane. It's now although,

58:12

of course, Microsoft is probably in

58:14

the best place to take advantage of this inflection

58:16

point. That it's already a large company with

58:19

a large search

58:19

engine. Although

58:20

I am curious how

58:23

they could shortly

58:24

rollout chat GPT as part of Bing just because of this issue

58:27

with accuracy. Yeah. I

58:29

think the way Neva does it

58:31

with the footnotes is the

58:33

only way you could do it. Right?

58:35

Because and this is the difference

58:37

between that knowledge graph and Google,

58:39

which is almost entirely from Wikipedia almost always and is

58:41

never sourced. At least, Neva

58:44

says, you know, where this stuff came from. I've

58:46

I've found it actually quite useful. I'll ask

58:48

TWiT, you know,

58:50

kind of technical questions, like coding questions, like, what, you

58:52

know, describe Dykstra's algorithm. And it

58:54

does a really good job. It's it's

58:56

quite this is exactly what chat should

59:00

be good at. And nevertheless,

59:04

to beat Google at its own game is not maybe

59:06

not. I

59:07

don't know. Maybe that was Now is

59:09

the time to do it. This is this is your

59:11

your article from last

59:14

June in

59:15

June before last? Oh, yeah. Twenty

59:17

twenty one was right when they were first Yeah. I hope they do

59:19

well. I feel very it's

59:22

very it's an interesting bet.

59:25

And I love not having the ads in there. And I

59:27

just hope they continue to be kind

59:30

of agnostic, you know,

59:32

not picking sides. I don't I don't wanna be

59:34

cut them to be a Bing new licensee or a Duck

59:36

To Go

59:36

licensee. They also, by the way, when you

59:38

install it, they install a

59:41

a

59:41

tracker. Which an anti tracker

59:43

tool plugin in your browser, which

59:45

shows you what trackers

59:48

are on Fast Company.

59:50

There you go. Not bad.

59:52

There are far worse. Let

59:54

me tell you, there's somewhere there's thirty or

59:56

forty trackers on a single page

59:58

It's kind of kind of

59:59

amazing. We were trying to make our pages leaner and

1:00:00

leaner just because that makes them

1:00:03

run faster -- That's gonna load

1:00:04

fast.

1:00:04

-- results in happier users? Yeah.

1:00:08

Alright. Wanna take a little break. There is a lot more to

1:00:10

talk about. We've got a great panel. I couldn't

1:00:12

have a better panel for this conversation. Tim

1:00:16

Stevens, is with us now freelance doing great.

1:00:18

He's driving his way home on

1:00:20

a substack at Tim Stevens dot substack,

1:00:24

dot com. He is also on Mastodon

1:00:26

on the Mastodon dot

1:00:28

social, but still says a little bit

1:00:31

a little tiny bit TWiT

1:00:34

In there too. Thank you, Tim, for being here. We appreciate

1:00:36

it. Harry at Technologize

1:00:39

your global tech editor at

1:00:41

a fast company we we started putting people's mass thumbs up

1:00:43

on the

1:00:43

screen. I think There we go. I think

1:00:46

that's

1:00:46

great. But

1:00:46

you can't add anything more because it's, like,

1:00:48

two seems to be the maximum.

1:00:51

Yeah.

1:00:51

Well, we got you Twitter and your Mastodon. It's

1:00:53

not my Post though. There you go. Are you on Post

1:00:55

as well? I have an account, but I haven't really been

1:00:57

using that. Yeah. I know. See, to

1:00:59

me, going to post is like not learning the lesson of Twitter. It's like,

1:01:01

oh, good. Let TWiT Andresen run

1:01:03

everything. Right?

1:01:04

No. I don't I think it's better to

1:01:06

be I love the idea that we

1:01:08

know somewhere that is not owned by

1:01:09

somebody. Right? I really

1:01:11

like that. And if you'd if you suddenly

1:01:13

are you're on Twitch social TWiT you hate the

1:01:15

way I'm running TWiT. You

1:01:17

go somewhere else, you know?

1:01:19

That's easy. Also on

1:01:22

the on the a new Mastodon user

1:01:24

and more more than welcome. Christina

1:01:26

Warren, film girl. And are you using

1:01:28

film girl you are at Maston dot

1:01:30

social? I

1:01:30

am. Yeah. Yeah. I'm I'm I'm at I'm

1:01:32

at the mother's girl. I might wind up

1:01:34

change another instance at some point because the mask on You're not the big

1:01:36

one. It's so big that

1:01:37

there can be yeah. There can I've

1:01:40

had the count since two thousand

1:01:42

eighteen. I don't know. IIII

1:01:45

had it just to have TWiT,

1:01:47

but It's

1:01:49

pretty easy that it's easy to move your followers. It's

1:01:51

hard to move your toots. You can

1:01:54

do

1:01:54

it, but most of the time, I think the stuff

1:01:56

that you have tutored or tweeted, the old

1:01:58

stuff that, you know,

1:02:00

that's water under the

1:02:01

bridge. Start fresh, but you at least can bring your

1:02:04

followers with you. That's very easy to do that

1:02:06

on that.

1:02:07

III like

1:02:10

our TWiT social

1:02:12

because it's you have to be a twit listener to

1:02:14

be in there. So it is a community. You're on

1:02:16

Harry's on SFBA, which is for San Francisco Bay area -- Mhmm.

1:02:19

-- people. The local timeline

1:02:21

really gets a a

1:02:23

point of view. If

1:02:25

you choose wisely, when you're on somewhere like Mastodon,

1:02:28

that social is just like a mini Twitter,

1:02:30

basically. It's everybody who it

1:02:32

didn't didn't look farther than the

1:02:34

the biggest instants. And

1:02:36

it's also pretty big now. It's well over a

1:02:38

hundred thousand people. So That could be fun.

1:02:40

Or the people who signed up in two thousand

1:02:42

teen. Yeah. And they didn't have a lot of these other things. Nice. There

1:02:44

was no twin social

1:02:45

banks. Awesome. Yeah. It was

1:02:48

also on a a

1:02:50

smaller 1X0X0

1:02:53

dot zone for my favorite my very

1:02:55

favorite

1:02:55

conference. Yeah.

1:02:56

And I I did

1:02:57

migrate like, I

1:03:00

never it. So I did go ahead and migrate the followers that I I had mask

1:03:02

there over to and there was

1:03:04

some some overlap I'm sure, but I did

1:03:06

migrate those followers over to

1:03:09

the the the main account that I'm on. But and

1:03:11

and that was actually

1:03:14

seamless. I was worried about what that process was gonna

1:03:16

be

1:03:16

like, but it wasn't cult. So that that

1:03:19

that's good news. I've done the

1:03:20

same thing I was unmasking on that

1:03:23

social way back when when it was

1:03:25

the only mess. That instance, And

1:03:27

when I started my own, I migrated over to

1:03:29

to with TWiT social. I also have something

1:03:31

on Pixel

1:03:32

Fed, which

1:03:33

is a metaverse not mask it on, but kind of Instagram

1:03:35

clone. And I really like it on pixel

1:03:38

fed social. I

1:03:40

really like

1:03:41

it because it's Instagram like it used to

1:03:43

be with just a bunch

1:03:45

of

1:03:45

photos. Golly, whoever thought of that, what

1:03:50

no reels. No dancing

1:03:52

chipmunks. What what kind of what kind

1:03:54

of places that? So

1:03:57

and and one of the nice things

1:03:59

about Ivory and these other clients is

1:04:01

you can you can

1:04:04

actually have multiple accounts

1:04:06

in your client. So you

1:04:08

can have your photos on pixel

1:04:10

fed and your toots somewhere else.

1:04:12

Our show two day brought

1:04:14

to you by our good friends

1:04:18

at worldwide technology. Worldwide technology is at

1:04:20

the forefront of innovation. We love these guys. We

1:04:22

had it was actually the last triple lease that

1:04:24

I took before COVID. We went out there in March

1:04:27

twenty twenty to visit their

1:04:29

advanced technology center. Wow. Is that

1:04:31

cool? You guys probably remember in the

1:04:33

old days if Davis had

1:04:35

that big testing lab in

1:04:37

Foster City. And that's why PC Magazine could

1:04:39

test a hundred printers because they had the

1:04:41

the capability of doing that.

1:04:43

You remember that? Well, that's what

1:04:45

the ATC is all about for enterprise technology. That's that's worldwide

1:04:48

technologies business is

1:04:50

enterprise technology.

1:04:52

They created the ANAVANCE Technology Center.

1:04:55

To try to research all of this

1:04:57

great technology out there for enterprise,

1:04:59

it now has all

1:05:02

the technologies from all the leading OEMs, including some of

1:05:05

the big new disruptors, more than

1:05:07

half a billion dollars in

1:05:10

equipment. It started one rack in one

1:05:11

building. It spread to four or five buildings now many,

1:05:13

many racks. But here's the

1:05:14

great part. And

1:05:15

I really honored w t

1:05:17

for this. They don't keep it to themselves. Sure. Their engineers use it to spin

1:05:20

up proofs of concept and learn

1:05:22

about new

1:05:22

technologies and

1:05:22

so forth to help their clients, but they

1:05:25

also make it available to you.

1:05:27

The advanced

1:05:28

technology center, you you don't have to go

1:05:30

to St. Louis. You can use it anywhere in the world. They

1:05:33

offer hundreds of on demand and

1:05:35

schedulable labs. Featuring solutions that include

1:05:38

technologies representing all the latest, the

1:05:40

newest advances in cloud

1:05:42

and security, networking

1:05:44

primary and secondary storage

1:05:46

data analytics and AI dev

1:05:48

ops and on and on and on.

1:05:52

It is

1:05:52

it is not just for those great WBT engineers and

1:05:54

partners, it's for anybody. It's free to anybody

1:05:56

who wants to use the ATC

1:05:59

platform. Which

1:06:00

means your evaluation time can go for months to weeks. Your

1:06:03

knowledge level

1:06:03

can go through the roof. You could test out products

1:06:06

and solutions before you go

1:06:08

to market, But it's more

1:06:10

than just the labs you can access,

1:06:12

technical articles, expert insights,

1:06:14

demonstration videos, white papers, all

1:06:17

the tools you need to stay up on the

1:06:19

latest enterprise technology. They also have a great

1:06:21

community. In fact, when you go to

1:06:23

the ATC platform, check out WWE's events and communities,

1:06:26

learn about technology trends here about the latest

1:06:28

research and insights

1:06:30

from experts, Not only is the

1:06:32

ATC at that physical lab space in

1:06:34

Saint Louis, and if you get a chance to see it do,

1:06:36

it was the most amazing thing, but it's

1:06:38

completely virtual so you can use it. If you're on the ATC

1:06:40

platform anytime anywhere in the

1:06:42

world. Three hundred sixty five days a

1:06:44

year. Whatever your

1:06:45

business needs, this is the point.

1:06:48

WWE is the best partner for

1:06:50

anybody using enterprise technology.

1:06:52

Worldwide technology can deliver scalable

1:06:54

tried and tested tailored

1:06:57

solutions. Because WWT understands

1:07:00

it's in business, technology

1:07:02

is not for technology's sake. It's

1:07:04

there to support your business strategy. WWF

1:07:08

brings strategy and execution together

1:07:10

to make this exciting new world happen

1:07:12

to learn more about WWF, the advanced

1:07:14

technology center to get access to

1:07:16

all these free resources. Very easy,

1:07:19

go to WWT dot

1:07:21

com slash

1:07:21

twit, WWT dot com

1:07:24

slash twit create a

1:07:26

free account, on the ATC platform and and

1:07:27

learn and and explore and

1:07:30

grow and use these technologies the way they're

1:07:32

intended to WWT dot

1:07:34

com slash Twit. These guys are

1:07:36

the good guys. These are these are the guys you need

1:07:38

as a partner. WWT

1:07:40

dot com slash tweet. Let's

1:07:45

see. Oh,

1:07:46

I do wanna do a

1:07:49

quick plug for our I think

1:07:51

it's the last chance to take the survey. Yeah. We have only two

1:07:53

days left. Twitter dot tv slash survey twenty three.

1:07:55

We survey our audience once

1:07:58

a year. We don't wanna spy on

1:08:00

you. We don't put we can't put trackers in a

1:08:02

podcast. It's RSS. But

1:08:04

we'd like to know more about you. Our advertisers

1:08:06

would like to know who they're who those are going

1:08:08

to. We can't compete with people like

1:08:10

Spotify or spy on your every move and know

1:08:12

who you are and all that stuff. It's

1:08:15

a survey. It's our only tool, but it helps us a lot. So

1:08:17

it should only take a few minutes. It's completely optional. Of course, answer any

1:08:19

questions you want. Twitter TV slash

1:08:24

survey. Twenty three. I wanna get every I

1:08:26

wanna get people from every show participating though, so we know

1:08:28

about, you know, what we're doing

1:08:30

and whether it fits your needs.

1:08:33

TWiT dot tv

1:08:35

slash survey twenty three. Last chance,

1:08:37

don't put it

1:08:38

off, and we thank you in advance.

1:08:41

Some really interesting

1:08:43

news from the

1:08:43

Department of

1:08:48

Justice There was a

1:08:50

ransomware game called Hive. Ransomware has become a plague, obviously. It's

1:08:52

really a problem.

1:08:55

Although I saw that the

1:08:57

revenues, and they they know this because they can look at Bitcoin

1:08:59

transfers. We're significantly down

1:09:02

in twenty twenty TWiT. And

1:09:05

the thinking is because people aren't paying. It's not that ransomware is not hitting

1:09:07

you. It's just people who've said,

1:09:08

screw that. We're not

1:09:11

giving you any money. Maybe

1:09:14

they've got better strategies for mitigating

1:09:16

them ransomware attack. But also,

1:09:18

the DOJ is going after

1:09:20

them. This was a press

1:09:22

conference from deputy attorney general Lisa Omonico. It

1:09:25

turns out

1:09:26

I think this is fascinating.

1:09:30

That the US had infiltrated

1:09:32

the FBI had infiltrated the

1:09:35

Hyve ransomware

1:09:35

group last

1:09:38

July. And as a result, under and

1:09:41

maybe this is why ransomware is going

1:09:43

down too. Under the,

1:09:46

you know, under the covers, that's not quite right. Officers

1:09:49

were able to warn victims

1:09:51

of impending attacks in

1:09:54

secret. Saying, hey, watch out. They're

1:09:56

going after you. They

1:09:58

also got decryption keys, and

1:10:00

they were able to hand out

1:10:02

more than three hundred decryption keys to

1:10:04

people who had been hit by

1:10:07

the hive ransomware, saving them more than a hundred thirty million dollars. The

1:10:11

US estimates Hive and its affiliates it's

1:10:14

one of those ransomware as a service

1:10:16

companies. I

1:10:19

don't wanna use a word but that's kind of what it is. Collected over

1:10:21

a hundred million dollars for more than fifteen hundred

1:10:24

victims. They went after, and

1:10:26

this was their mistake, hospitals, school

1:10:28

districts, critical infrastructure.

1:10:30

In more than eighty countries around the world, one hospital was left unable to accept new patients because

1:10:36

of

1:10:36

Hive, They worked

1:10:38

with the UK's national crime agency

1:10:40

and other law enforcement agencies around

1:10:43

the world to help victims. And

1:10:46

the UK fifty organizations were given decryption keys. And on Thursday,

1:10:49

the FBI shut

1:10:52

it down.

1:10:54

They took Hive's website and communications networks

1:10:56

down with the help of police

1:10:58

forces in Germany

1:10:59

and the Netherlands.

1:11:01

TWiT is a successful

1:11:04

attack on

1:11:05

the attackers. I don't

1:11:08

know if They

1:11:10

arrested anybody. I don't see that.

1:11:13

And that's the

1:11:16

problem because as

1:11:18

the head of intelligence at Mandy and John Hultquist said, until you arrest him, they're not

1:11:23

gonna be gone. It's like

1:11:25

soccer soccer cup over again. They It'll down.

1:11:28

If if

1:11:32

you went to the Hive Cruise

1:11:34

website, you would see this notice from the FBI. This hidden

1:11:37

site has

1:11:40

been seized. With lots of

1:11:43

badges. This is Nihai

1:11:45

was not the biggest of

1:11:47

the ransomware gangs. There are

1:11:49

bigger ones. Although, Rival, which was perhaps the biggest in twenty twenty in

1:11:51

twenty twenty

1:11:52

one, did get

1:11:55

arrested around the world. So

1:11:58

this is good. Dark side was taken down in June of twenty twenty one. This

1:12:03

is good. This is what

1:12:06

it takes. Let's see what else. Intel. You wanna talk

1:12:12

about Intel? Not a

1:12:14

good quarter

1:12:14

for Intel, the worst beating

1:12:15

in over a decade, Andrew

1:12:18

Orr writes from Apple and Cider.

1:12:22

There may be a little happy about little happier

1:12:24

than they ought to be about this.

1:12:26

Thirty two percent drop in revenue since

1:12:30

year over year since the

1:12:32

holiday quarter of last year twenty twenty

1:12:34

one, actually. Fourth quarter results coming out

1:12:38

revenue fourteen billion dollars down thirty percent

1:12:41

year over year, entire

1:12:43

year revenue down twenty percent

1:12:45

year over year. This goes along with

1:12:47

drops of thirty, forty percent in the

1:12:49

PC sales as well. So it's just

1:12:52

been a bad year for

1:12:54

PCs. Does that mean anything, Harry? Well,

1:12:56

I think Intel has known and acknowledged for a

1:12:58

while now that it's in this rebuilding process after

1:13:01

falling way behind

1:13:03

other chip companies. And

1:13:06

that it was not gonna result

1:13:08

in fantastic numbers immediately because they have to get

1:13:10

back to where their process is competitive again.

1:13:15

With with other technologies. And I believe they've

1:13:17

said that maybe by year

1:13:19

after next, they think they'll

1:13:21

be in a place with

1:13:24

Is great again, which is maybe as

1:13:26

long as they give pet galsing or their

1:13:28

CEO, time

1:13:31

to get there. Maybe that's when we can really judge

1:13:33

them. And if the numbers are still

1:13:35

this bad, then it's a really

1:13:37

bad sign. But I think

1:13:39

that at least as of when Galesinger

1:13:42

started, and I I wrote a future about him last year. The board had given him quite

1:13:44

a bit of runway on

1:13:46

understanding that those can be difficult.

1:13:49

And there would be more bad news before

1:13:51

there was any good news. Although, they may not have anticipated the degree to which business would

1:13:55

be so crummy. And I

1:13:57

think I think people and companies may just be purchases

1:14:00

because that everybody is so

1:14:02

cautious about the economy this year.

1:14:06

Yeah.

1:14:07

Everything's down. It's not just PCs. Plus,

1:14:09

we bought a lot of PCs

1:14:11

during COVID. Right? People have

1:14:13

relatively new nice computers now in

1:14:15

a way they didn't before the

1:14:16

pandemic? Right. And if you look at, like,

1:14:18

the increase in in in shifts,

1:14:20

you know, between, like, twenty

1:14:23

twenty and and now not

1:14:25

to say that, like, the some of

1:14:27

the gains haven't been impressive, but if you're not an enthusiast, you're not actually going to really notice, I think,

1:14:29

for a lot of

1:14:32

people. And you

1:14:34

know, TWiT looking more and more like what was happening, you know, all that excess buying in

1:14:36

twenty twenty and and even a

1:14:38

little bit into twenty twenty one.

1:14:43

Was a combination of both the the

1:14:45

supply chain, you know, maybe

1:14:47

even making more of

1:14:49

a frenzy because people couldn't

1:14:51

things, people having to work from home. It's an anomaly. And I think

1:14:53

to Ben, as as a lot of businesses did, try

1:14:55

to base it up as, like, well,

1:14:57

this is the new baseline was clearly

1:15:00

a mistake. Because that's, you

1:15:02

know, has has not continued. And and I I think that

1:15:05

with the,

1:15:08

I guess, being able to

1:15:10

kind of look back, we can say,

1:15:12

no, why would we have expected those trends to continue

1:15:14

year over year? Because that's just not consumer buying patterns

1:15:19

in in the last decade or

1:15:19

so, you know, we haven't seen that. So yeah.

1:15:22

We have been saying

1:15:23

for a

1:15:23

while the

1:15:26

end of

1:15:26

desktop computing But I think it's what do you think, Tim?

1:15:29

Is the end of

1:15:31

desktop computing exaggerated?

1:15:33

I I definitely think it is. I mean, I think we've got a

1:15:36

long time to go before that that and

1:15:38

certainly people's usage patterns show that they're shifting

1:15:41

away from desktop computing if you look at

1:15:43

overall utilization, you know, what devices consuming content on and creating

1:15:45

content on. But if you look

1:15:47

at overall time, I think

1:15:49

that number is going up

1:15:51

and desktop usage probably staying pretty much static

1:15:53

for the past few years. So so, yeah, I think we still have a long way to go there, but if

1:15:55

you also look at the number of layouts

1:15:58

we've seen lately, I mean, that's a

1:16:00

lot fewer corporate

1:16:02

laptops that are being needed. And certainly with nobody hiring, that means that there are fewer laptops being needed there too. And

1:16:04

if you do

1:16:07

get hired

1:16:08

now, I think it was probably

1:16:10

a pretty

1:16:11

good chance of getting a hand me down. So

1:16:13

I think that Joe's lapped up, but

1:16:14

we fired him last night. Yeah. And -- Alright, Peter. -- yeah. Sorry,

1:16:16

Joe. Got

1:16:19

it. Twelve thousand layoffs at Google. I

1:16:21

mean, it's just been tough. We

1:16:23

had on on

1:16:24

Wednesday, on TWiT. We had

1:16:27

just completely a

1:16:28

representative because I think what you

1:16:30

know, we talk about these layoffs in I

1:16:33

think tech industry, since the beginning of the

1:16:35

year, two hundred

1:16:36

thousand jobs lost. We talk about that and

1:16:38

this just kind of abstract numbers. I wanted to bring a face to

1:16:40

it. So we had

1:16:43

Richard

1:16:43

Hayon. He was a Google engineer. He's been

1:16:46

an engineer

1:16:46

for seventeen years at Google,

1:16:47

and was one of the people just summarily dismissed, kind

1:16:52

of

1:16:52

abruptly lost his job without any warning. His boss didn't

1:16:54

even know ahead of time. And I wanted just to kinda bring home the face of it because

1:16:56

these are that's

1:16:59

two hundred thousand people with

1:17:01

families, with bills, with mortgages, with rent, and

1:17:03

and they don't know what

1:17:05

tomorrow is gonna bring. That's

1:17:08

a huge hit.

1:17:11

And I don't want to just, you know, diminish it in

1:17:13

any way by just talking my raw

1:17:15

numbers, you know. Yeah. It

1:17:17

it really is a shame how that has to happen

1:17:19

these days or, like, the the corporateification of layoffs

1:17:21

is is really tragic and and

1:17:24

nauseating. Honestly, you know,

1:17:26

having recently been through that myself. How

1:17:28

depersonalized it has been mandated that

1:17:30

you you cannot have any empathy,

1:17:32

you cannot talk to

1:17:35

anybody about the situation. You you

1:17:37

are very restricted in what you you can say, when you

1:17:39

can say TWiT. you know, as someone

1:17:41

who isn't tried to be an

1:17:43

apathy leader, someone who you

1:17:46

know, treated his employees like his

1:17:48

friends. You have to go through that is

1:17:50

really really difficult. That's the situation. So Yeah.

1:17:53

I

1:17:53

I don't know where this pattern came from or

1:17:56

why it is almost

1:17:58

legislated into corporate law

1:18:00

these

1:18:00

days, but it is really discussing that

1:18:02

that is where we've gotten to a

1:18:04

right now

1:18:05

where your ability to be an empathic leader has to end at

1:18:07

the time when it's most important for

1:18:09

you to be an

1:18:12

empathic.

1:18:12

Yes. I

1:18:13

worry that Elon Musk sent sent the bar so low for

1:18:15

doing them. Epidemic or just decent to

1:18:18

the people who worked for you

1:18:20

that if

1:18:22

these large companies beat Elon, they figure

1:18:24

that that it's okay. But, I mean, there

1:18:26

were stories about Google employees who who came

1:18:29

into work. And waved their badge to get in

1:18:31

and either turned green and they were able to go in or turned red and they knew they had been laid

1:18:33

off and that's how they got

1:18:35

the

1:18:35

news. And it don't

1:18:38

understand what the excuse is

1:18:40

for that. Yeah. Yeah.

1:18:40

That's what I mean.

1:18:41

Yeah. There's there's no good

1:18:42

way to do lay offices. Is the

1:18:47

reality, but there are ways

1:18:49

that you can do it worse.

1:18:51

Right? And I

1:18:53

I agree, like, for all the excuses

1:18:55

that this has been a thing I think I've I've noticed

1:18:58

because I I first was seeing this in media where

1:19:00

people would find out sometimes

1:19:02

that they were laid off by

1:19:05

losing access in Slack, and then people

1:19:07

would disappear. And it would be, like, you know, like, the snap. And and you

1:19:09

were, like, what happened? You

1:19:11

know, brought back TTSD

1:19:14

one day when people lost access for Slack for completely

1:19:17

unrelated reason and everybody freaked out. They're

1:19:19

like, we'll just What

1:19:21

does this mean? It just

1:19:23

expired. Yeah. And and, you know, you do this

1:19:25

for for the automation reason. So we don't want people to have access to things. It's

1:19:28

like,

1:19:29

okay. Especially for

1:19:32

people who you're paying a certain amount of money and

1:19:34

who you have worked for

1:19:34

you for a certain amount of time. It's like, have some freaking humanity. You

1:19:37

know, there's a way

1:19:39

to do it there's a way to to

1:19:42

take access away. It doesn't mean that someone is entering the office

1:19:44

at seven AM,

1:19:47

hasn't checked their email their

1:19:49

personal right email, doesn't know what's going

1:19:51

on, waves a badge, and and finds out that way. Like, that

1:19:55

that just It's awful. And there are

1:19:58

better ways to do it. There's no good way to do it period, but

1:20:00

there are ways to do

1:20:02

it that are worse than others.

1:20:07

Somebody in the chatroom just told me that

1:20:09

Chris DeBona, who was one of the

1:20:11

founders of Floss Weekly, great

1:20:14

friend is also an ex Googler. He

1:20:16

was a director of open source

1:20:19

at

1:20:19

Google. I did not realize he

1:20:21

had lost his job as well.

1:20:23

So loss for them. And that's a massive that's

1:20:25

a massive concern for open source because of all

1:20:28

the work and money and resources

1:20:30

that Google has given open source

1:20:32

projects. Over

1:20:34

the years, sponsoring conferences and and other things. That's been a discussion that has come up

1:20:36

in the last couple of weeks with with

1:20:38

these bake lay offices. What does that mean?

1:20:43

the source I don't think that it's a wrong

1:20:46

one because budgets are tight everywhere. And

1:20:48

these are things

1:20:50

that, you know, we that some people in the open source movement don't like to acknowledge,

1:20:52

but a lot of the money and

1:20:54

and a lot of the funding

1:20:58

really does come from these corporations whether those

1:21:00

checks go away, like, what

1:21:01

does that mean? Because the this

1:21:04

sustainability in open source has been a

1:21:06

really big topic for the last number

1:21:10

of years. And corporate goodwill is or

1:21:12

or corporations paying their own wafer

1:21:14

services to support is one thing.

1:21:18

Laporte but the goodwill aspect, which has

1:21:20

been increasingly a thing that we've

1:21:22

seen happen, like, I I can

1:21:25

see that potentially at some places

1:21:28

And that's really discouraging and

1:21:30

I think it have really

1:21:32

negative consequences

1:21:35

because people happen always wanted to maybe

1:21:37

acknowledge how much of of a

1:21:39

role those those checks and

1:21:41

that funding can really

1:21:43

play. A lot of small projects. Microsoft also laying off

1:21:45

about ten thousand workers. And this

1:21:47

is the thing, you know, you go

1:21:50

I'm sure you do this too, Christina. You

1:21:52

go you look and you just check

1:21:54

and see, oh, gosh. And I'm sure there's corporate, you know, Slack's and stuff that

1:21:56

you can go to and see who's

1:21:58

there apparently a a Twitter there with

1:22:01

TWiT salute icons as

1:22:03

people -- Yeah. -- dropped off

1:22:05

the face of the

1:22:07

earth. Microsoft's border Go

1:22:09

ahead. I was gonna say, I think that what made it it

1:22:11

hard for it from Microsoft. And I'm sure for Google too, is everybody you know, a

1:22:13

lot of people working

1:22:16

from home And so,

1:22:18

yeah, there were a lot of, you know, kind of, you know, shadow groups of people. Yeah. Well, you know, yeah, you don't know,

1:22:20

but people check-in with

1:22:22

one another. I mean, that's

1:22:25

what I was doing. I was checking it with my

1:22:27

friends at Microsoft. I'm I'm in a a few group shots,

1:22:29

and and that's what I was doing, and then, you know,

1:22:31

checking Twitter. And and seeing, you

1:22:33

know, and some people were were were laid off and and whatnot. And and when you're talking

1:22:35

about numbers this it's not about

1:22:38

performance. It it really

1:22:40

is you know,

1:22:42

decisions made usually about entire divisions -- Just slash series. -- so mysterious. Yeah. One of the

1:22:47

stories and again, I hate to I'm

1:22:50

not gonna put you on the spot. You don't represent Microsoft by in in any stretch of imagination. But one of things

1:22:52

that we did learn

1:22:54

is that Microsoft's

1:22:55

VR, AR, Hollow

1:22:58

lens division suffered massive cuts.

1:23:01

And and that sounds

1:23:03

to me like more

1:23:05

of a strategic decision

1:23:07

to to not

1:23:08

pursue those

1:23:08

areas. And instead of

1:23:09

making hardware to to make their software available to

1:23:12

companies like HTC

1:23:15

and and meta, that are gonna make

1:23:17

the hardware an apple. One imagines that are gonna make the hardware and then, you

1:23:20

know, make the productivity software

1:23:22

for that hardware, which actually probably

1:23:26

is a better bet than

1:23:28

than putting all your chips

1:23:30

in on on legless, sexless people

1:23:32

wandering around in low poly

1:23:35

count. Microsoft also to

1:23:38

take Corey's pros. Yeah. Microsoft

1:23:40

also gave up on

1:23:42

Allspace VR, which was a a startup

1:23:44

data query. A few

1:23:46

years ago. So what

1:23:47

was a platform? I

1:23:50

mean, I I would be

1:23:52

cautious about assuming that Microsoft doesn't

1:23:54

have any ARVR slash metaverse platform

1:23:58

ambitions forever. But but if they do,

1:24:00

maybe this seems like a little bit of a reset.

1:24:02

And

1:24:02

-- Sure. -- it seems it seems perfectly sensible

1:24:05

at this point to redeploy some of

1:24:07

that metal band weapons and resources

1:24:09

into AI, which so clearly is

1:24:11

gonna have so much impact starting

1:24:13

at this very moment as opposed to the

1:24:15

metaverse, which is still, like, maybe at some point and maybe not to

1:24:17

the degree we expected

1:24:19

kind of thing. Is

1:24:22

it risky though to chase the flavor of the

1:24:23

month? Because, I mean, that's why they VR. True. Although, I mean,

1:24:26

I

1:24:26

I don't know my

1:24:29

for all the reasons to be cautious

1:24:32

about AI. I I think even if it's only five percent

1:24:34

as impactful as people expect that it's gonna be incredibly

1:24:36

important. Laporte

1:24:39

were a lot of reasons to think VRAR was not going

1:24:41

anywhere from day one. I mean,

1:24:43

that eleven percent of

1:24:45

the people who used it were nauseated

1:24:47

is a pretty good indicator that there's gonna be this may

1:24:49

not be the mass appeal

1:24:50

product. You hope it will be. There's some

1:24:53

if you want to have magical glasses

1:24:55

up up like these, have great battery life

1:24:57

and fantastic

1:24:58

Well, that's absolutely plan. Right? Spectacles. Well, they give up on that too. There may be just there's some fundamental

1:25:00

pieces of technology. We have no

1:25:02

idea how to build so far. Right.

1:25:06

We

1:25:06

know how to do AI.

1:25:07

It's a battery life. Yeah. Exactly. Chemistry

1:25:09

moves at a glacial pace compared to

1:25:11

a digital stuff. That

1:25:13

was one of the stories from the week

1:25:16

that Mark Berman saying Apple is gonna push off its, you know,

1:25:18

spectacle based AR vision for at least a couple of years.

1:25:22

To twenty twenty five, if not

1:25:24

later, because they can't get

1:25:26

it working. Even their headset,

1:25:29

which they're still rumors

1:25:31

are strong, they're gonna offer for three thousand

1:25:33

dollars this year has a battery in your pocket

1:25:35

because it's too heavy

1:25:37

to wear on your

1:25:40

head. Right. So, yeah, I think there are

1:25:42

some fundamental technical issues with we the problem is we all

1:25:44

read the same

1:25:47

science fiction

1:25:48

stories. By William Gibson and Neil

1:25:49

Stephenson. Don't want this. We all wanna jack in the metaphor. You all want this.

1:25:52

But

1:25:52

Well It's

1:25:53

not sci fi. I mean, I

1:25:55

mean, the battery

1:25:58

No. The battery thing is one of the biggest ones. I

1:26:00

mean, I I've been a proponent

1:26:02

of of going nuclear for a decade

1:26:04

at

1:26:04

least. Do you want a little nuclear

1:26:07

power plant

1:26:07

in your head? I mean,

1:26:08

honestly, I would trust it more than lithium ion. Okay. If

1:26:10

if you look at the safety record, I honestly would.

1:26:14

Do we have that technology? I know we have pocket nuclear reactors for

1:26:16

power, but pocket means I don't think The

1:26:18

size of this room

1:26:19

that I mean,

1:26:20

I don't I don't

1:26:21

I don't I don't know if we do or not. And my

1:26:23

my point is more like, wish

1:26:25

that we had been investing more over the last decades in in looking at that as a power source

1:26:28

than in some of these other

1:26:30

things because I do think that then

1:26:32

that

1:26:34

in my mind, that's the only way you can get

1:26:37

the long lasting battery life and

1:26:39

the the microization that you'll

1:26:41

need for these things. But I I just

1:26:43

don't think it's gonna be possible with with looking

1:26:45

in polymers. I I just don't. Okay.

1:26:47

Sure.

1:26:47

Physics and

1:26:50

chemistry. I

1:26:51

understand. But I'm just not

1:26:51

sure people are anxious to

1:26:53

wear the clear power

1:26:55

plant

1:26:55

hat. You're

1:26:58

not wrong, but again, I mean, maybe it needs a rebranding. just saying,

1:27:00

like, the the brand Don't call

1:27:02

it the nuclear mask.

1:27:03

Okay. That's a good thing. Right.

1:27:05

Don't call it the

1:27:07

nuclear mask. I'm I'm saying, like, it's it's

1:27:09

a branding thing, but, like, if I think that the

1:27:11

technology like, that's it's obviously the one of

1:27:13

the only solutions that I can think of

1:27:15

that we already

1:27:16

have. Because solar

1:27:18

is certainly not going to be enough enough listeners

1:27:20

in Australia are probably

1:27:22

aware of the fact that

1:27:28

the a a tiny Cesium one

1:27:30

thirty seven capsule went missing

1:27:32

on its way to

1:27:35

purse this past week. Just one, it's let me see

1:27:37

if I can find a picture of it because it's

1:27:39

it's so small. They're they're one in

1:27:41

the public not to

1:27:44

touch

1:27:44

it. But it's so small. I don't even

1:27:46

know how you would find it. It's about the size of one of those little

1:27:48

lithium ion

1:27:49

batteries that you put in

1:27:51

your in your pocket.

1:27:56

Here it is. Here's the size next to

1:27:58

a Australian something. Ten ten pens

1:28:00

piece. I don't know what

1:28:03

that

1:28:03

is. Six millimeters by eight millimeters. If you see it,

1:28:06

yeah, don't

1:28:06

touch it. Don't

1:28:07

pick it up.

1:28:10

Don't taste

1:28:10

it. Yeah. It's like remember, your guys are too young. Remember my

1:28:13

when I was a kid, there were constant ads

1:28:15

not to touch blasting caps.

1:28:18

Do you remember that? Do you remember that,

1:28:20

John? No. Blasting caps. That was,

1:28:22

like, maybe that was a major problem.

1:28:24

I know kids. If you see

1:28:26

this, don't touch it.

1:28:28

Well, kids, if you see a six

1:28:30

millimeter by eight millimeter shiny silver capsule,

1:28:32

TWiT

1:28:34

don't touch it. It could kill you. It could kill you. So do you wanna that your head?

1:28:41

I don't know. Christina's point though, we haven't

1:28:43

seen any real progress in alternate sources for power in a while. I I remember CES about

1:28:45

a decade

1:28:47

ago, there was like, a whole

1:28:50

full of portable devices

1:28:50

for hydrogen power, basically. So you could have

1:28:53

a fuel cell

1:28:55

in your lap topper on your phone and even and

1:28:58

there was so many different vendors. It seemed like it was just a couple of years away. Yeah. And then I'm guessing they started

1:29:02

you know, exploding in people's pants, and that was probably the end of that. But we

1:29:05

haven't really seen anything since

1:29:07

then. So certainly, solid state batteries are are

1:29:09

just around the corner. I think we'll see those

1:29:11

soon. And those will will provide a

1:29:13

pretty big step forward in terms of charging speed, discharge speed, and

1:29:15

will help to reduce the overall volume of

1:29:18

a given capacity of

1:29:20

battery.

1:29:21

But really, you know, like I said, we'll chemistry gap and slowly, and there

1:29:23

really isn't any kind of shot in the darkness

1:29:27

coming soon for portal devices for cars, you

1:29:29

know, super capacitors, things like that. I think we'll have some big gains in a decade or

1:29:31

so, but there's nothing like that

1:29:34

going for smaller stuff. Yeah.

1:29:37

Microsoft's recorder was not great. Revenue was up two

1:29:39

percent. Profit down twelve percent. This primarily, I

1:29:41

think, due to this

1:29:43

PC drop off. Both

1:29:47

below Wall Street expectations,

1:29:50

Amy Hood, Microsoft's

1:29:52

chief financial officer,

1:29:55

said New business slowed in December, but it expects -- and it expects

1:29:57

growth to continue to slow in the current

1:29:59

quarter, which ends March

1:30:02

thirty one. On the other hand, I think Microsoft is very

1:30:05

well positioned. This open AI investment

1:30:07

is looking very smart.

1:30:10

Right now. Clearly, if AI

1:30:12

is taking off businesses like Azure are

1:30:14