0:00

Welcome to the Analytics

0:02

Power Hour Hour. Analytics

0:04

topics covered conversationally and sometimes

0:07

with explicit language. Hey everybody,

0:09

welcome. It's the Analytics Power

0:11

Hour Hour. This is episode

0:13

265. And I think it

0:15

was Socrates who said the

0:18

unexamined life is not worth

0:20

living. And I believe he

0:22

said that right before putting

0:24

on his uro ring, slipping

0:26

on his loop band, and

0:28

jumping into his eight sleep

0:31

bed. One thing for sure,

0:33

though, we've got a lot more

0:35

places to collect data about ourselves

0:37

than we did back in his

0:39

day. And I think it represents

0:42

some interesting possibilities, maybe some challenges.

0:44

So we wanted to talk about

0:46

it. I mean, we're data people.

0:48

So, you know, who better to

0:50

tackle this topic. And Julie Hoyer,

0:53

manager of Analytics at Further, do

0:55

you use any of these tools

0:57

to like measure stuff about yourself?

0:59

Funny enough, I religiously wear an

1:01

Applewatch and it's collecting things, but

1:04

I couldn't tell you the last

1:06

time I looked at the, you

1:08

know, the dashboard summary data in

1:11

the app, if I'm honest. Nice.

1:13

No, that counts though. That counts.

1:15

So I'm Tim Wilson, head of

1:17

Solutions and facts and facts and

1:20

facts. measuring your heart rate.

1:22

I've got my polar H10 heart rate

1:24

monitor on right now because just want

1:26

to see how excited I get throughout

1:29

this show. So we should run that

1:31

in real time along so the podcast

1:33

to see how excited or unexcited Tim

1:36

is on a topic or how stressed

1:38

out it makes him. And I'm Michael

1:40

Helbling and yeah, I think I've got

1:43

stuff on my phone that measures how

1:45

many steps I take and things like

1:47

that. Okay, but we needed a guest.

1:50

Somebody who could help shed some light

1:52

on this topic and bring this discussion

1:54

to you or listeners. So we found

1:56

one. Michael Tiffany is the CEO and

1:58

co-founder of Fulker Dynam. was the

2:00

founding CEO, then president of Human,

2:03

a cyber security company. He also

2:05

serves on various boards as well

2:07

as advises startups. And today he

2:10

is our guest. Welcome to the

2:12

show, Michael. It's a pleasure to

2:14

be here. Me and all of

2:17

my connected devices. Nice. Are you

2:19

big? Do you do quite a

2:22

bit of that? Or just, I

2:24

assume because of your company. You

2:26

probably do a lot of testing

2:29

at least. Rock and an Apple

2:31

watch. I'm wearing an aura ring,

2:33

I've got a connected scale, I've

2:36

got an eight-sleep bed, I'm I'm

2:38

breathing into this lumen device to

2:40

instrument my metabolism by by looking

2:43

at my out-breaths. Here's how weird

2:45

I am. I'm rocking a smart,

2:47

addressable breaker box. So, so among

2:50

other things. I'm measuring, I'm like

2:52

monitoring power to the stove to

2:54

just passively monitor how often I'm

2:57

cooking. Whoa. Yep. That's a whole

2:59

nother level. What's the eight sleep

3:01

bed? What's the eight sleep bed?

3:04

Yeah, I haven't heard of that.

3:06

It's magnificent. It's a bed that

3:08

circulates through it, interwoven through the

3:11

entire bedtopper, are small channels for

3:13

water. that run to a refrigeration

3:15

slash heating unit, so the bed

3:18

can either cool you down or

3:20

warm you up, or in my

3:22

case, key for marital bliss, cool

3:25

me down while warming my wife's

3:27

side. Wow. Wow, that sounds nice.

3:29

But it also does a lot

3:32

of measuring of like your sleep

3:34

quality and stuff like that at

3:36

the same time, right? That's exactly

3:39

right. Yeah. And in owning, I

3:41

was an early adopter owning this

3:43

thing. feels like owning a Tesla

3:46

where the same hardware has been

3:48

getting better and better with OTA

3:50

updates. So while I bought it

3:53

mostly for that temperature regulation, I've

3:55

seen its sleep monitoring, its measurement

3:57

of my heart rate in the

4:00

night. like just get better and

4:02

better and more accurate, which has

4:04

been a delight. Wow. Nice. Yeah,

4:07

I used to have an if-then-that

4:09

routine running against my scale to

4:11

dump any weights I did into

4:14

a Google sheet for a long

4:16

time, but that was a long

4:18

time ago. And I think the

4:21

company that made that scale doesn't

4:23

have an API anymore, so. That

4:25

was like my gateway drug, right?

4:28

Okay, nice. If this then that

4:30

scripts to like gather this kind

4:33

of stuff. Yeah, and look at

4:35

me now. Nice. I do diligently

4:37

when I'm traveling, I miss a

4:40

little bit that I don't have

4:42

a scale because I have it

4:44

every morning that is part of

4:47

the routine. Not to the point

4:49

of having a connected scale. I

4:51

actually was given a connected scale

4:54

for Christmas, I think a year

4:56

or so ago, and I'm like,

4:58

I don't think I need that.

5:01

It's just take a measurement and

5:03

punch it into my phone while

5:05

my toothbrush is running. Who knows?

5:08

Yeah, whatever works. Okay. Hmm. All

5:10

right. So it's not a competition.

5:12

What is... Well, yeah, I hope

5:15

not, because I'm not winning. Michael

5:17

wins. Yeah. No. Yeah. All right.

5:19

So, so yeah, Michael, what, what

5:22

is the word for this? Like,

5:24

so one of the things that

5:26

gets used a lot as sort

5:29

of self quantification or self-date data.

5:31

Yeah. But like... What is sort

5:33

of the holistic term for this

5:36

or what's going on in this

5:38

space? Because obviously there's, even we've

5:40

mentioned a bunch of different companies

5:43

and things like that, but there's

5:45

more, there's many, many more. And

5:47

you can go beyond that to

5:50

like, you know, DNA, like 23

5:52

and me and those kinds of

5:54

things as well. Yes. So, so

5:57

in the early days, I would

5:59

say the pioneering hackers who were

6:01

coming together and sharing tips and

6:04

tricks tricks, we're talking about the

6:06

movement as quantified self. And that

6:08

really was in its pioneering phase.

6:11

These days, like I just showed

6:13

you my oral ring, they've surpassed

6:15

a million sales in North America.

6:18

This is now a popular device,

6:20

not just a niche device. And

6:22

while that has taken off, quantified

6:25

self as a term. of art,

6:27

I would say, has actually declined.

6:29

And this is a good thing,

6:32

not a bad thing, because what

6:34

quantified self promises you by, you

6:36

know, just the meaning of the

6:39

words is a bunch of numbers.

6:41

And that's not what people want,

6:44

right? They want insights, they want

6:46

self-knowledge, and they want increasingly connected

6:48

wellness, connected health. And I think

6:51

that captures something important, which is,

6:53

you know, the intention, the goals

6:55

here. It's not really about counting

6:58

steps. It's actually about, you know,

7:00

10 more years of a good

7:02

life. Yeah, so what is, I

7:05

guess, is there a term or

7:07

is there a singular idea or

7:09

vision that everyone says this is

7:12

what we're trying to get to

7:14

is X? Right. If I had

7:16

to pick one, it would be

7:19

connected wellness. And the reason why

7:21

it's those terms in particular is

7:23

that we're in a transition right

7:26

now based on the recognition that

7:28

health care has for many, many

7:30

years really been something more akin

7:33

to sick care. It's about fixing

7:35

you after something is broken. And

7:37

that's not awesome. There are things

7:40

you should be doing right now

7:42

to improve your wellness that mean

7:44

that less things will go wrong.

7:47

So that's the, apart from just,

7:49

you know, branding and marketing, that's

7:51

the true reason why you're seeing

7:54

the word wellness more. It's to

7:56

try to differentiate the, you know,

7:58

proactive pursuit of optimal health versus

8:01

recovery from something going wrong. We're

8:03

doing that in two ways that

8:05

are new, signaled by the word

8:08

connected. One is. that we're wearing

8:10

increasingly smart devices that in effect

8:12

make you like a type A

8:15

personality, like make you like a

8:17

really good diarist without you having

8:19

to do any work, right? I

8:22

just step on my scale, I

8:24

don't write anything down, which is

8:26

nice. And so it's connected in

8:29

that sense, the devices somehow probably

8:31

really sending bites over the wire.

8:33

And then also connected in the

8:36

sense that this data by being

8:38

digitally native is more shareable. with

8:40

a doctor with a loved one,

8:43

maybe even, you know, just shared

8:45

socially because so much about staying

8:47

fit and healthy is, it's like,

8:50

it depends on social engagement and

8:52

like doing it with others. So,

8:54

so if I had to pick

8:57

two words to capture everything that

8:59

seems to be the, the ascendant

9:02

term, it would be connected wellness.

9:04

And who's, this is funny, I

9:06

would think of their, early days

9:09

of internet of things where there

9:11

was there was talking of if

9:13

you're imagine your garage door being

9:16

able to tell you that it's

9:18

got a bearing that needs to

9:20

be greased and it's going to

9:23

go out right with sometimes those

9:25

seem kind of forced like I

9:27

don't spend a whole lot of

9:30

time feeling like I need I

9:32

need to preemptively maintain my garage

9:34

door opener like it will break

9:37

every 10 to 15 years but

9:39

when you talk about the health

9:41

the logically, early detection, early detection

9:44

slash preventative care makes sense. Is

9:46

the thinking that that is in

9:48

the hands of a, a mean,

9:51

the data collection has to, has

9:53

to, is geographically tied to the

9:55

human, but is it something that

9:58

the health care? provider will say

10:00

I need I need your historical

10:02

data if you have it or

10:05

like who's Yeah, where does it

10:07

come from? Who's driving that? Here's

10:09

how I'm approaching this in my

10:12

own life, and I found this

10:14

to be transformative. Actually, goes back

10:16

to Michael's opening observation about Socrates.

10:19

Self-knowledge is incredibly hard. It's actually

10:21

incredibly difficult to achieve extraordinary self-knowledge.

10:23

And so the way it's done,

10:26

the best way to achieve extraordinary

10:28

self-knowledge and insight for the past

10:30

several thousand years, going back to

10:33

Socrates, going back to, you know,

10:35

Vedic religions in India, or even,

10:37

I was just looking at the

10:40

rule of St. Benedict, you know,

10:42

some 1500 years ago, he's writing,

10:44

everyone does the same thing worldwide,

10:47

which is dramatic simplification. You live

10:49

like a monk, like this is

10:51

the point of of the monastic

10:54

life, it's to dramatically simplify your

10:56

life, so then you can focus

10:58

and achieve extraordinary self-knowledge and insight.

11:01

And, you know, sometimes you peer

11:03

into the very nature of reality

11:05

as well. I don't want to

11:08

do that. Like, I want to

11:10

have the self-awareness of a monk

11:13

while actually engaging with the world

11:15

like a bone of a want.

11:17

And so the challenge I set

11:20

before me, being a computer nerd,

11:22

is Like, can I use computers

11:24

to help me out in this

11:27

regard? Because computers are infinitely patient

11:29

and honestly they're really good at

11:31

counting stuff. Like, I am, I

11:34

believe that Kung Fu Masters centuries

11:36

ago really could cultivate, you know,

11:38

the ability to like, just be

11:41

constantly aware of their own heart

11:43

rate. And that was probably awesome.

11:45

I'm not going to do that.

11:48

I'm going to put on an

11:50

Apple watch. So. That's sort of

11:52

an empowering view of the world,

11:55

but I would say. that something

11:57

must be missing because the people

11:59

dawning Apple watches or aura rings

12:02

or other kinds of instrumentation are

12:04

augmenting their bodies, they're augmenting their

12:06

lives with breakthrough technology that was

12:09

sci-fi just decade ago, but I

12:11

don't think we feel like the

12:13

six million dollar man, right? You

12:16

strap this in and you don't

12:18

like just feel magically empowered. So

12:20

what is it like? What's missing?

12:23

I think that siloization is a

12:25

really big limiting factor. And I'll

12:27

give you a health care example,

12:30

and then we'll go back to

12:32

like my connected breaker box. My

12:34

bed has all these awesome instruments,

12:37

right? Like it's measuring my HRV.

12:39

It'll tell me how long I

12:41

spent in deep sleep. But it

12:44

knows nothing about what I did

12:46

the day previously that contributed to

12:48

or ruin a good night's rest.

12:51

I, for instance, learned and other

12:53

folklore users have seen the same

12:55

thing, that by getting. Passive telemetry

12:58

on my eating. So I'm not

13:00

even a big food logger that's

13:02

like a little bit too much

13:05

work for me, but I will

13:07

put on a CGM. So I've

13:09

done multiple experiments wearing a connected

13:12

glucose monitor, a continuous glucose monitor,

13:14

that's just passively recording my blood

13:16

glucose. So therefore is going to

13:19

see blood sugar spikes when I've

13:21

eaten a bunch of carbs. And

13:23

what do you know? Like a

13:26

few weeks worth of experimentation showed

13:28

that... If I want better specifically

13:31

deep sleep, I should shift my

13:33

carbs if I'm going to eat

13:35

any to the beginning of the

13:38

day. So carbs before noon, I

13:40

sleep well. Carbs afternoon, you're starting

13:42

to get into a dangerous zone.

13:45

Like dessert after dinner, forget about

13:47

it, I'm going to have an

13:49

elevated heart rate and I'm going

13:52

to have shortened deep sleep. It

13:54

is impossible to know about that

13:56

causal relationship. unless you're somehow tying

13:59

the data that's drawn from the

14:01

CGM with the data that the

14:03

bed knows about. So we've surrounded

14:06

ourselves with these ostensibly smart devices,

14:08

but they're not really smart. They're

14:10

just data-producing devices. The smartness comes

14:13

from a higher level of analysis.

14:15

And I feel like people like

14:17

me are on the leading edge.

14:20

We're geeking out on our own

14:22

data. doing data science on this

14:24

raw data in a Python notebook,

14:27

which is like too much to

14:29

ask for from maybe, you know,

14:31

the average person, but that's going

14:34

to be within the grasp of

14:36

the average person, some extent already,

14:38

and to an increasingly large extent,

14:41

because of coding copilots. So people

14:43

who've never written a lick of

14:45

code before are sometimes, you know,

14:48

getting like one shot. outputs of

14:50

functional code from, you know, colloder,

14:52

chat, chute, chute, t, that means

14:55

that what used to be really

14:57

esoteric data science skills are becoming

14:59

increasingly within the grasp of ordinary

15:02

people, but only if you've gathered

15:04

and desilode the data, hence my

15:06

focus with Volkra. I think that's

15:09

something that I've been thinking about

15:11

a lot is how, even once

15:13

you have all the data... in

15:16

one spot so that you could

15:18

use it to paint a bigger

15:20

picture, ask more helpful questions about

15:23

your health, how do we determine

15:25

what good looks like? Because what's

15:27

interesting is some of the devices

15:30

seem to be making some of

15:32

that decision and determining like what

15:34

is a good range of these

15:37

metrics. Other ones don't. They truly

15:39

do just collect the data. So

15:42

it's interesting to think when you

15:44

start to connect those things and

15:46

tie them together. Kind of back

15:49

to Tim's question, does it become

15:51

a place where that is baked

15:53

in so an individual can go

15:56

out? these questions and get those

15:58

types of answers or is it

16:00

more so that the value is

16:03

it's all together and you could

16:05

take it to a professional to

16:07

help tell you what does this

16:10

mean is this bad and based

16:12

on that like what do I

16:14

do about it yeah I'm thinking

16:17

about I'm thinking about changing the

16:19

world in this order once you

16:21

have the self-knowledge that I'm describing

16:24

then you also have new ways

16:26

of sharing how it's going in

16:28

your life with another person, which

16:31

could be a doctor, but could

16:33

just be a spouse, could be

16:35

a group of friends. So everything

16:38

starts with solving the observability problem.

16:40

Like, I think it's too hard

16:42

to get help because it takes

16:45

so much effort to just describe

16:47

to anyone else, like, this is

16:49

what's going on with me. Like,

16:52

this is how I've slept the

16:54

last week or, you know, this

16:56

is what's stressing out. is you

16:59

can think of as the human

17:01

equivalent of what we call in

17:03

Devops like observability, right? So the

17:06

instrumentation, these connected devices, they're solving

17:08

the observability problem. Then there's like

17:10

this analysis problem, which we just

17:13

sketched. And then finally, there's new

17:15

forms of sharing. And I'm like

17:17

really excited about that. Like I

17:20

want to know how. Like my

17:22

friends are sleeping in general, right?

17:24

Like how is it going with

17:27

people that I love but now

17:29

live distant from me? And also

17:31

what's normal. So what I'm hoping

17:34

is by reducing the friction and

17:36

the risk of sharing personal observability

17:38

data like this, but by making

17:41

it secure and controllable. then we'll

17:43

also be able to pull this

17:45

data to find out what's normal

17:48

across. larger groups, so you can

17:50

kind of compare yourself to averages.

17:52

Right now it's like really hard

17:55

to tell. Am I a weirdo?

17:57

And I think the internet is

18:00

sort of good at solving those

18:02

problems if you can build the

18:04

bridge between the data collection and

18:07

the kind of social sharing that

18:09

you want to do. I've got

18:11

anxiety now, as it is with.

18:14

I mean with Strava or I

18:16

mean I had Fitbit before or

18:18

Apple, I mean, I mean there

18:21

does feel like a broad parallel

18:23

that is not encouraging which is

18:25

move away from us measuring ourselves

18:28

and just kind of the world

18:30

of digital where at a corporate

18:32

level there is this obsession with

18:35

let's gather everything we can. I

18:37

mean the 360 degree view of

18:39

the customer. taken to an extreme

18:42

would be a marketer knows how

18:44

often you're cooking so they can

18:46

you know make self easy cook

18:49

meals available to you or something

18:51

yeah right I mean there's the

18:53

there's the nefarious which I feel

18:56

like insurance and government we should

18:58

get into as well right but

19:00

just the idea I mean there

19:03

there have to be people listening

19:05

and because I'm experiencing it a

19:07

little bit myself like Oh my

19:10

god, like sharing, comparing, like don't

19:12

we have, don't we have a

19:14

challenge with our youth just from

19:17

the crude form of TikTok and

19:19

Instagram comparing themselves and it's not

19:21

good for their mental health. So

19:24

it's like this, gather all this

19:26

data first, hope the analysis happens

19:28

and then we're creating community, is

19:31

there a dark side or downside

19:33

to that we need to that

19:35

we need to figure out. I

19:38

think so. I think there's there's

19:40

extraordinary benefit and an extraordinary risk

19:42

and in that That's why an

19:45

entrepreneur most known for starting cybersecurity

19:47

companies have, that's why I've waited

19:49

into this. Our design with fulcra

19:52

importantly starts with who we're working

19:54

for and how we make our

19:56

money. When you create an account

19:59

with fulcra, your data belongs to

20:01

you. You are not sharing it

20:03

with your future self. And our

20:06

revenue model is. asking for money

20:08

for that service and we need

20:11

to re-earn our customers trust every

20:13

day and if we lose that

20:15

trust then they will stop paying

20:18

us money and we will be

20:20

very sad. So I think that

20:22

being a force for personal data

20:25

sovereignty in this way is something

20:27

you have to choose to do

20:29

at the foundation of your company

20:32

and build into your DNA. I

20:34

think that if you are an

20:36

ad-funded company, even if you are

20:39

a multi-billion dollar or multi-trillion dollar

20:41

or ad-driven company, you cannot just

20:43

cite to like pivot into a

20:46

new light of business where customer

20:48

data belongs to the customers and

20:50

is, you know, encrypted in motion

20:53

and at rest and like is

20:55

just designed for whatever the customer

20:57

wants to do with it and

21:00

nobody else. The control that people

21:02

have over their own data I

21:04

think is actually going to be

21:07

of increasing importance as. AI agents

21:09

become an increasingly important part of

21:11

the future. Because as we can

21:14

see over the last especially two

21:16

years of rapid improvement in generative

21:18

AI, it's going to be very

21:21

hard to control AI models by

21:23

trying to put a cap on

21:25

their capabilities. I just I don't

21:28

even see how that's going to

21:30

work. I don't think we can

21:32

say AI can only have an

21:35

IQ of 140. No higher. Like

21:37

that's just not going to work.

21:39

So. How are ordinary people going

21:42

to have any control over an

21:44

agent that they're asking So you

21:46

want to get help from a

21:49

helpful AI assistant. How are you

21:51

going to be able to accept

21:53

that help, share enough data with

21:56

that agent that you can get

21:58

some help, but make it like

22:00

a two-way door, make it a

22:03

revocable commitment? And I think there's

22:05

only one way to do that.

22:07

And that's to control access to

22:10

your own data. So you can

22:12

grant it to an assistant. And

22:14

you say, sure, you can read

22:17

my health data. But you can't

22:19

copy it. And if I change

22:22

my mind for any reason or

22:24

no reason at all, I get

22:26

to turn that off. If instead

22:29

all of our data is going

22:31

to live with some large tech

22:33

provider that's also running the models,

22:36

if the only way you get

22:38

the help is by like uploading

22:40

all of your data in a

22:43

one-step process, you've completely lost control.

22:45

And that's like not the future

22:47

that I want to bring about.

22:50

So what we're trying to do

22:52

here. is you know empower people

22:54

as I said with self knowledge

22:57

but it's even more broadly building

22:59

an important force for personal data

23:01

sovereignty so that we can have

23:04

the benefits of AI but put

23:06

people in control. It's interesting too

23:08

with being health data. I think

23:10

it brings a very different awareness

23:12

to the world of AI and

23:14

the sharing of data and your

23:16

own data that I think people

23:18

it's very different than today right

23:20

like some people think like Oh,

23:22

you care about what I clicked

23:24

on, what ads I saw, right?

23:26

Like, it's your data, but it

23:28

feels really different when you start

23:30

to talk about, like, your personal

23:32

health metrics. And so it's, I'm

23:34

really happy to hear you talk

23:36

about it that way, and it's

23:38

really helpful to hear you talk

23:40

about that way for even just

23:42

my understanding of, like, what could

23:44

this look like, what should this

23:46

look like ethically in the future,

23:48

but I really hope that it

23:50

kind of sparks that light bulb

23:52

for other people of like. When

23:54

we're talking about your data and

23:56

privacy and the importance of it

23:58

and how it interacts with AI,

24:00

like, yeah, thinking about it the

24:02

way you think about your personal

24:04

health data for all your other

24:06

data, I don't know, it really

24:08

sparked. some clarity for me. But

24:10

it also highlights the gap we

24:12

have in the United States around

24:14

data ownership and data rights as

24:16

a person because there's not laws

24:18

in the US about if you

24:20

give that data to somebody else

24:22

what they can use it for.

24:24

And so health data. can be

24:26

predictive of many different things potentially

24:29

so just like you know the

24:31

car insurance companies want you to

24:33

take the little thing and plug

24:35

it in to track all your

24:37

movements to save you money but

24:39

in reality right it's helping them

24:41

create better predictive models for what

24:43

the likelihood you're going to get

24:45

in an accident is and what

24:47

the risk you have to them

24:49

as an insured person is and

24:51

so in the same way like

24:53

where that data goes and so

24:55

like even if you take your

24:57

data and I think this came

24:59

up with 20 and me because

25:01

I think they were. contemplating selling

25:03

the company to somewhere else and

25:05

it's like well what what happens

25:07

to all that data if someone

25:09

else comes and buys that company

25:11

what are they allowed to do

25:13

with that data if they acquire

25:15

right yeah what happens with meta

25:17

buys forkera well I mean and

25:19

so like that's a legitimate concern

25:21

because there's no underlying regulatory structure

25:23

that says does someone who comes

25:25

along and buys a company like

25:27

that can do or not do

25:29

things with that they quote unquote

25:31

own now Yeah, right. I love,

25:33

I love this kind of thinking.

25:35

And I think that when you

25:37

dig into privacy by design at

25:39

many companies, you find that there

25:41

is this end state where people

25:43

just say, well, we never do

25:45

that. And like that is an

25:47

adequate answer because you cannot guarantee

25:49

that you will always have your

25:51

hands on the wheel. So in

25:53

fact, I would encourage anyone listening

25:55

as they're thinking through what privacy

25:57

by design at a at, you

25:59

know, a Olympic level really looks

26:01

like, you have to show how

26:03

you are preserving privacy, even if,

26:05

you know, ultra super mega. Corp

26:07

acquires your company. You actually need

26:09

to limit the powers you have

26:11

as a business operator to mess

26:13

with people's data and inspect their

26:15

data, right? So that even under

26:17

the conditions where you're acquired by

26:19

a company that doesn't share your

26:21

values, they can't just, you know,

26:23

like, switch on the... the data

26:25

vacuum mode and undo all of

26:27

your work. And there are absolutely,

26:29

this is not just me thinking

26:31

about this happily, there are good

26:33

patterns of privacy by design that

26:35

are built to operate at that

26:37

high level. And I think that's

26:39

absolutely the level that literally every

26:41

company should aspire to. But there's

26:43

the, there's having, there's following all

26:45

the principles of privacy by design

26:47

and then putting something in place

26:49

and then there is also. the

26:51

I mean you sort of said

26:53

it earlier there needs to be

26:55

a trust that somebody's going to

26:57

provide their data and explaining there

26:59

still winds up being you know

27:01

truck to the masses to those

27:03

million people with an or a

27:05

ring right if you say I

27:07

mean I would guess that most

27:09

of them are saying I don't

27:11

really care I'm not getting a

27:14

whole lot of thought take my

27:16

data but if you're going to

27:18

300 million people And the truly

27:20

paranoid fringe, the, and we're in

27:22

a very weird little subset of

27:24

four people here who are happy

27:26

to spend an hour talking and

27:28

thinking about this, and we're not

27:30

remotely scratching the surface of what's

27:32

actually going on in a design

27:34

to make that happen. So actually

27:36

convincing, you know, Joe Smith, that

27:38

No, this really is okay. And

27:40

maybe this becomes just a societal

27:42

breakdown thing. They're like, says who?

27:44

My cousin Vinny said, you're going

27:46

to use this for nefarious purposes.

27:48

And no amount of rationalization will

27:50

change your mind. Right. So to

27:52

me, this is a dimension. of

27:54

business design. I'm a business nerd.

27:56

And an observation that I've had

27:58

is that whatever a company says

28:00

its mission is, if the execution

28:02

on the mission is not exactly

28:04

what earns them money, that's not

28:06

the mission. The revenue is the

28:08

mission. Over time, if these two

28:10

things are not in alignment, I'll

28:12

tell you which one wins. It's

28:14

the one that increases earnings. So

28:16

you can just know that and

28:18

then you can consider that a

28:20

constraint of business design and then

28:22

construct a revenue model that is

28:24

truly consistent and in fact even

28:26

supports your mission. That's one of

28:28

the things that I'm most proud

28:30

of with the magnificent success of

28:32

human. Cybersecurity company fights cybercrime at

28:34

scale, goes after the profit centers

28:36

of cybercrime. Importantly. doesn't have to

28:38

sell to the CSO. It's not

28:40

just another layer of protection. If

28:42

you're in the business of fraud

28:44

detection, you actually reduce losses due

28:46

to fraud. And so the reason

28:48

why you get paid is that

28:50

you charge less than the savings,

28:52

right? So then every single customer

28:54

knows exactly why they're paying you

28:56

and the incentives of that company

28:58

are such that human makes the

29:00

most money by going after the

29:02

biggest source of cyber criminal profit,

29:04

which therefore means that it is

29:06

designed to have the biggest possible

29:08

positive effect on the world, which

29:10

is super cool. Here, with fulcrab,

29:12

here's the way I see this

29:14

playing out. Lots of people. So

29:16

consider the universe of like everywhere

29:18

there's data about you, right? Everything

29:20

you use that generates some data.

29:22

You know, Facebook knows some stuff

29:24

about you and Apple knows some

29:26

stuff about you and maybe, you

29:28

know, the, where a ring has

29:30

a little bit of data. And

29:32

I don't think you need to,

29:34

I don't think you need to

29:36

go around like to. all of

29:38

that. But if you and only

29:40

you have the superset, if you

29:42

have all your data from every

29:44

single one of those sources, then

29:46

you're the only one who has

29:48

the complete picture. And you could

29:50

decide to then invoke some right

29:52

to be forgotten, or you ask

29:54

for all your data to be

29:57

deleted, and then you'll truly only

29:59

have the only copy. But I

30:01

think it's good enough. that you

30:03

are the master of the complete

30:05

set? Because that'll alter incentives going

30:07

forward where some people who just

30:09

already have some sliver of data

30:11

about you, they don't have to

30:13

ask permission. They've already got it.

30:15

But if they want to have

30:17

access to the full picture to

30:19

provide a better service or whatever,

30:21

they have to ask you. And

30:23

to a great extent, I think

30:25

that's winning. Right? Like if individuals

30:27

are just in charge and get

30:29

to say yes or no, if

30:31

they're asked at all, that would

30:33

be pretty great. Right now, in

30:35

real time bidding for like most

30:37

of the ads that are getting

30:39

served to you, even though you've

30:41

had to answer a bunch of

30:43

nonsense cookie consent pop-ups, like no

30:45

one's really asking your permission for

30:47

doing, you know, some kind of

30:49

cookie or pixel sink that is

30:51

connected to some email newsletter that

30:53

you signed up for, that they're

30:55

using to figure out how many

30:57

people are in your household and

30:59

what your income is, right? You

31:01

had, you were just not involved

31:03

in any of that. And like,

31:05

that's the little turn that I

31:07

want to just make on society.

31:09

And we could do that through

31:11

lawmaking, we try to force people

31:13

to ask for your consent. But

31:15

I think what's even better is

31:17

to reward them, to rationally motivate

31:19

them to deal with you, like

31:21

they get better data. And we'll

31:23

deliver you a better experience. So

31:25

like they'll do it if it's

31:27

in their best interest. And I

31:29

think that happens when people are

31:31

in control of like the super

31:33

corpus. You bring up a point

31:35

that I actually would love. to

31:37

kind of circle back on because

31:39

it goes in two areas we've

31:41

talked about. One, I do feel

31:43

like if there was clarity, so

31:45

say you were the owner of

31:47

all your data, I feel like

31:49

the only way to get people

31:51

to share their data openly, like

31:53

on a large scale with companies,

31:55

is if those companies could tell

31:57

us as individuals like, We would

31:59

love this type of data from

32:01

you because then we could answer

32:03

these types of questions. Here's the

32:05

benefit. Like that value tradeoff they

32:07

talk about, like if you allowing

32:09

cookies, right? Like what does it

32:11

get you in return? Why should

32:13

you share this with this company?

32:15

But what's really interesting is we

32:17

know that one doesn't happen. I

32:19

think it would be amazing if

32:21

it could, but Because we know

32:23

that people aren't starting with a

32:25

question in mind always. There is

32:27

still the obsession that we talk

32:29

about a lot on the show

32:31

that companies have of like just

32:33

collect all the data and I

32:35

do feel like it goes into

32:37

the, you know, connected health conversation

32:39

we're having of people think if

32:42

I have all the data on

32:44

myself, then I'll be. able to

32:46

answer all these amazing questions. I

32:48

don't know what questions I'm exactly

32:50

going to ask, but if I

32:52

have all the data, I'll be

32:54

able to. And then you get

32:56

into the reality of a lot

32:58

of these questions. You can't answer

33:00

or you're answering them with data

33:02

that you inherently realize has biased

33:04

or errors in it. So then

33:06

it kind of takes you down

33:08

the path too of like, there's

33:10

a whole area of the industry

33:12

that spun up then to collect

33:14

more and better data, but we're

33:16

still probably going to miss the

33:18

piece of like. What's the motivation

33:20

of collecting all this data? Like

33:22

what do companies want to ask

33:24

and use it for? What do

33:26

you yourself want to ask? What's

33:28

a helpful question to ask? What

33:30

should you be collecting data to

33:32

then get out of it? So

33:34

I know there's kind of like

33:36

a lot of branches we could

33:38

take off that, but it's just

33:40

been interesting hearing the last couple

33:42

points you've made. I'll throw this

33:44

out there is like a concrete

33:46

prediction of the future. I think

33:48

the way this plays out is

33:50

that there's like... too much data

33:52

for a human to sort through.

33:54

There are too many potential use

33:56

cases for it all. But it

33:58

really does seem to me like

34:00

we're headed to a place where

34:02

helpful AI. assistance are within everyone's

34:04

grasp. So what I think will

34:06

happen is you will have a

34:08

kind of concierge agent, concierge, that

34:10

only works for you, that has

34:12

trusted access to your data. And

34:14

it intermediates with other companies agents,

34:16

and essentially negotiates on your behalf.

34:18

So instead of you having to

34:20

deal with a whole bunch of.

34:22

questions about consent and like individual

34:24

offers, there's just going to be

34:26

too much sort through. But you'll

34:28

be able to delegate it to

34:30

your agent and just be like,

34:32

show me the two marketing offers

34:34

that you think are really going

34:36

to land with me. That's how

34:38

much human detention I actually have.

34:40

And so your agent might be

34:42

dealing with like countless kinds of

34:44

unsolicited offers or ideas and is

34:46

providing the curation layer based on

34:48

knowing you. And then in a

34:50

rule space way, can share like

34:52

the little subsets of data that

34:54

are going to be able to

34:56

activate those offers or, you know,

34:58

make them work. I see. If

35:00

I'm right, that means that agent

35:02

to agent communication is going to

35:04

be the majority of internet traffic

35:06

within like 10 years. That's kind

35:08

of scary to think they could

35:10

be talking on your behalf in

35:12

the background and then that becomes

35:14

its whole own black box. Like

35:16

it's cool, but it kind of

35:18

scares me too. Just as long

35:20

as they open the pod bay

35:22

doors. Well, but I mean back

35:25

to that, I think there is

35:27

this, and I know I've run

35:29

through it when I've made a

35:31

fitbit, which I don't know, I've

35:33

probably gone through six fit bits

35:35

over, I don't know how many

35:37

years. And when I switched to

35:39

an Apple Watch, there was, I

35:41

genuinely felt a, oh my God,

35:43

I'm like losing all this historical

35:45

data. And I draw that parallel

35:47

to the business world. The reality

35:49

is, is I bet when somebody

35:51

has a heart issue, they get

35:53

sent home with a a heart

35:55

monitor and they say, let me

35:57

collect a couple of, let me

35:59

collect a couple of weeks, wear

36:01

this for a month. So as

36:03

you're talking AI agents, my brain

36:05

went off on a, I want

36:07

to lose weight, or I want

36:09

to sleep better, or I want

36:11

to do X or Y. Here's

36:13

all the data that I'm already

36:15

collecting in an aggregated way. Here's

36:17

what's already there. What can you

36:19

do with that? have the agent

36:21

tell me, you know what, you

36:23

should put a CGM on for

36:25

a while, but not turn into

36:27

this. There's nothing in my entire

36:29

history of working with Analytics that

36:31

makes me think that anyone is

36:33

going to be good at saying,

36:35

collect this data for a while

36:37

for a specific purpose because there's,

36:39

well, just in case, imagine the

36:41

next time you ask if you've

36:43

already been collecting that, then you

36:45

don't need to collect it for

36:47

another. two weeks. So I, the

36:49

exchange you just, you two just

36:51

had, had me thinking like, is

36:53

there a data, because that's one

36:55

of the privacy by design principles

36:57

is around like, collect a minimal

36:59

amount of data. So where does,

37:01

where does that fit into it

37:03

that don't collect it just in

37:05

case you need it, collect it

37:07

once you know what you need,

37:09

but this neb, get everything, and

37:11

then we'll. have the most to

37:13

work with. Some of it's not

37:15

going to ever, you know, matter,

37:17

right, or not matter enough to

37:19

make it worth it. And in

37:21

the name of prevention, it's kind

37:23

of hard to make that case.

37:25

That's right. So in the longevity

37:27

context, I think if you ever

37:29

want to train an AI on

37:31

yourself, you kind of want to

37:33

have as much data as you

37:35

can possibly afford to have. So

37:37

the things get different when you

37:39

think about data retention, when you're

37:41

thinking about it for your own

37:43

purpose. versus regulating businesses for their

37:45

commercial services. Like one of the

37:47

reasons why we felt I honestly

37:49

compelled to create full growth was

37:51

because because of the data lossage

37:53

that you just talked about. Like

37:55

the fact is that like Geocities

37:57

died, right? Like it turns out

37:59

the internet isn't forever, like data

38:01

will just completely go away. And

38:03

you've got a host of options

38:05

for saving files, you know, Dropbox,

38:07

drop, Google Drive, apples, There's no

38:10

streaming data store for consumers. There's

38:12

no like Kafka for people. So

38:14

for data like your location history,

38:16

your calendars, any biometric, right, like

38:18

my heart rate just keeps happening,

38:20

thank goodness. So it's not a

38:22

file, right? It'll never be a

38:24

file. It is a stream. So

38:26

I need a streaming data store

38:28

for it. And there literally were

38:30

no options. So we had to

38:32

write one ourselves. And the way

38:34

I see this. being brought to

38:36

bear over time is that all

38:38

of these data streams that I

38:40

have pouring into my full-grade data

38:42

store are capturing how I live

38:44

and how I live in what's

38:46

going on with me situational awareness

38:48

is one of the things you

38:50

need to give to a potential

38:52

assistant so they can actually be

38:54

helpful. Right now we're all experimenting

38:56

with chat bots where you have

38:58

to initiate every conversation and that's

39:00

really limiting. I want to live

39:02

in a world that's more like

39:04

what you just described him, where

39:06

some external source of intelligence points

39:08

out what I'm missing, like tells

39:10

me about a thing I wouldn't

39:12

have thought of, and is like,

39:14

dude, you need to put on

39:16

a CGM for a couple of

39:18

weeks. It's not going to be

39:20

forever. We just need to sort

39:22

of sample this diet of yours

39:24

and see what is up. I

39:26

want to leave this data corpus

39:28

behind for my heirs. And to

39:30

me... make all of this data

39:32

unambiguously mine and unambiguously inheritable, I

39:34

need to collect it before I

39:36

die. Like my kids are not

39:38

going to be writing to Amazon

39:40

or whatever and being like, please

39:42

let us export the data. Like

39:44

it's over by then. You need

39:46

to have it. It needs to

39:48

unambiguously be yours before the event.

39:50

And what is all this data

39:52

add up to? It adds up

39:54

to how I lived. It adds

39:56

up to who I did the

39:58

living with, you know, like you're

40:00

going to be able to in

40:02

some cases probably recreate my tone

40:04

by transcribing this podcast and, you

40:06

know, feeding it into 11 labs

40:08

and capturing my voice and you'll

40:10

capture some of my like vocal

40:12

intonations, but none of this tells

40:14

you about all that tacit stuff,

40:16

the person. all the procedural knowledge,

40:18

right? So an AI model that's

40:20

trained on me that lives on

40:22

after me is a model that

40:24

I hope will, you know, bake

40:26

cookies with my great-grandchildren. I'm extremely

40:28

proud of my almond flour chocolate

40:30

chip cookie recipe and it's not

40:32

just about the ingredient list, it's

40:34

about how I do it, right?

40:36

So you should be able to

40:38

like walk into the kitchen in

40:40

the future and, you know, boot

40:42

up GrandPappy Michael and... And we're

40:44

going to bake cookies together. This

40:46

is going to be great. But

40:48

only in the first part of

40:50

the day, not later. Yeah, that's

40:53

right. That's right. Yes, no compensation.

40:55

I was thinking there would also

40:57

be an agent saying, you have

40:59

not asked, hey, I'm grandpapi, Michael,

41:01

and you haven't asked me to

41:03

make cookies with you in a

41:05

while. Like, oh my God. Oh,

41:07

that's a little too on the

41:09

nose. Don't you want to connect

41:11

with your ancestry? guilt tripping beyond

41:13

the grave yeah you never call

41:15

yeah I mean there's there's something

41:17

that says like you you could

41:19

always be making better day-to-day like

41:21

you there is a bit of

41:23

a bleak you know hey do

41:25

you really want that next I

41:27

know you made the cookies that's

41:29

good but you really do you

41:31

need the third one we've been

41:33

monitoring you and I don't know

41:35

I mean it's yeah it is

41:37

interesting because obviously this vision of

41:39

the world creates and it kind

41:41

of brings to life some very

41:43

interesting possibilities kind of like you've

41:45

been talking about Michael and then

41:47

some concerns as well and so

41:49

it'll be very interesting to sort

41:51

of see how this progresses and

41:53

the one thing unfortunately we can't

41:55

progress with further we do have

41:57

to start to wrap up because

41:59

we're really at a time but

42:01

but it's this is pretty fascinating

42:03

and at the same time sort

42:05

of like I think on the

42:07

downside risk part of it we

42:09

all in sort of envision that

42:11

guy Brian Johnson is his name

42:13

that sort of like measures every

42:15

possible thing and wants to live

42:17

forever and and we're sort of

42:19

like yeah I don't think that's

42:21

me but I think there's a

42:23

somewhere there's a happy media to

42:25

catch recently he found out he

42:27

was doing there was one of

42:29

the things he was doing that

42:31

was actually working in the opposite

42:33

direction I can remember what it

42:35

was but well that's comforting actually

42:37

a little bit so that's fine

42:39

But it also is kind of

42:41

exciting to sort of think of

42:43

yourself like Neo and the Matrix

42:45

and you turn around and be

42:47

like, I know Kung Fu, right?

42:49

Because like, I didn't have to

42:51

study to become a Kung Fu

42:53

master, but now I have these

42:55

AI assistance and data that helps

42:57

me do the things they could

42:59

do. like understand my heart rate

43:01

and those kinds of things. Making

43:03

a data-driven decision that one of

43:05

your health interventions wasn't working is

43:07

kind of where we all need

43:09

to be, right? Instead of absorbing

43:11

the recommendations that supposedly worked for

43:13

the 22 people in the double-blind

43:15

clinical trial that might not work

43:17

for you, the question is what

43:19

works for you, specifically you, and

43:21

then you want to double down

43:23

on those and stop the ones

43:25

that don't. I'm optimistic about that

43:27

kind of tuning over time. I

43:29

think lots of people are going

43:31

to live for a very very

43:33

long time from here. Yeah, until

43:35

we upload ourselves into the machine

43:38

guide. Oh, wait. Yes, that's right.

43:40

Yes. Bring on the still up

43:42

on brains. I mean, we didn't

43:44

even touch on neural link. So,

43:46

you know, that's a second episode,

43:48

maybe. Okay, we have to wrap

43:50

up. But one thing we like

43:52

to do is go around the

43:54

horn, share something that we think

43:56

might be of interest to our

43:58

listeners. It's been a really awesome

44:00

conversation, though. Michael, thank you so

44:02

much for joining us to do

44:04

it. But yeah, you're our guest.

44:06

Do you have a last call

44:08

you like to share? It is

44:10

outrageously cold here in coastal New

44:12

Hampshire. It's going to get down

44:14

to three degrees Fahrenheit tonight. So

44:16

the first thing that pops in

44:18

my mind is actually just like

44:20

my favorite new product. I got

44:22

innu heat gloves. So get this.

44:24

There are gloves that take a

44:26

battery pack. The battery pack. Doesn't

44:28

use some weirdo proprietary connector, right?

44:30

It's USBC, thank goodness, right? So

44:32

I like charge the the battery

44:34

packs with a USBC outlet. I

44:36

snap them onto my gloves and

44:38

oh my god, they really do

44:40

work Just that's awesome. It's so

44:42

cool. Yeah, I just got a

44:44

fleece for Christmas that does the

44:46

same thing and it's you literally

44:48

just hit a button and it

44:50

turns on and it worms you

44:52

up all over. Michael Helpling this

44:54

is, Julie got that my son

44:56

got my wife the same vest

44:58

because we were gonna go skiing

45:00

and also in my helping had

45:02

shown me his and I was

45:04

like that's weird and then realized

45:06

that actually my wife had also

45:08

gotten one for Christmas and it's

45:10

similarly like hooked and she's had

45:12

electric gloves for a while. I

45:14

know right here I showed up

45:16

and I'm like talking I'm like

45:18

oh yeah I have an addressable

45:20

breaker box right like I have

45:22

an addressable breaker box right like

45:24

I'm doing all this like crazy

45:26

mad science clothing makes me feel

45:28

like I'm living in the future.

45:30

You know, like, yeah, it's a

45:32

holiday. That's awesome. That's awesome. All

45:34

right, Julie, what about you? What's

45:36

your last call? My last call,

45:38

I'm sure everyone's heard about the

45:40

congestion tax in New York. I

45:42

know it's a big thing, and

45:44

I had found the link to

45:46

the congestion pricing tracker, and it's

45:48

got some good data visualization. I'm

45:50

really interested to see as time

45:52

goes on, like. what do they

45:54

find? They even had done I

45:56

think a good job like stating

45:58

what they're hoping will happen from

46:00

it. So I love that they

46:02

actually paired it with, hey this

46:04

is how we're visualizing things, want

46:06

to know how they're going to

46:08

analyze it, what are their conclusions

46:10

going to be, but my favorite

46:12

part, Tim, is that when I

46:14

got to the bottom of this

46:16

tracker, it actually says that it

46:18

is run by, I'm guessing two

46:21

students at Brown University and supervised

46:23

by Emily Oster. So I was

46:25

like, you know, I wonder how

46:27

I love this. Oh my God,

46:29

it's great. So I've just been

46:31

peeking at it. It hasn't been

46:33

running obviously too long just for

46:35

this year so far. But I

46:37

think it's really cool and I'm

46:39

excited to see what comes out

46:41

of it, especially knowing that Emily

46:43

is involved. So I think that

46:45

might be a last call that

46:47

needs to become a future episode.

46:49

Awesome. That's so cool. I saw

46:51

that same thing. I was like,

46:53

oh my gosh. So that's so

46:55

cool. All right, Tim, what about

46:57

you? What's your last call? So

46:59

as I am, I tend to

47:01

do, I'm going to do a,

47:03

I'm going to do a threefer,

47:05

I think. So one, I want

47:07

to call out back three. Yeah,

47:09

they'll be quick. And Cassie Kasrachav

47:11

will be included in one of

47:13

them. So, uh, one, back when

47:15

we first started talking when we

47:17

first started talking to talking to

47:19

talking to talking to talking to

47:21

the fulker talking to the fulker

47:23

team talking to the fulker team

47:25

talking to the fulker team about

47:27

talking to the fulker team about

47:29

talking to the fulker team about

47:31

talking about talking about having Michael

47:33

on for this, tried out the

47:35

kind of hooked up what I

47:37

could and also kind of interesting

47:39

even I can find things that

47:41

that I'm not that connected but

47:43

you know there's no swarm connection

47:45

like it's it's crazy how many

47:47

things we have that are tracking

47:49

and the the challenge of tracking

47:51

everything but I think there is

47:53

a there's like a seven trial

47:55

if anybody wants to just download

47:57

the app, then you kind of

47:59

hook up whatever services and you

48:01

kind of get to see what

48:03

the aggregated data looks like. Is

48:05

that right? Yeah, yeah. Everyone should

48:07

give it a trial. I think

48:09

most people are surprised by the

48:11

data that they have and just

48:13

didn't know about. You quite likely

48:15

have years worth of step count

48:17

data that you didn't even know

48:19

because it was sort of silently

48:21

turned on by your iPhone. Yeah.

48:23

So it's always, even if you

48:25

just want to look and you

48:27

just want to delete the app

48:29

after seven. And it has a

48:31

very well documented API from, you

48:33

know, playing around with it. So,

48:35

you know, we're not, this is

48:37

not a, you know, paid endorsement,

48:39

but this whole discussion, if it's

48:41

got people thinking, oh, that's kind

48:43

of worth checking out. And I

48:45

think actually hearing you talk about

48:47

sort of division kind of makes

48:49

it a little more exciting to

48:51

people thinking. So, thank you. So

48:53

that's one. Number two, just a

48:55

PSA for anybody who, if you're

48:57

not already falling, castee- caught that

48:59

she's moved over to sub stack

49:01

and has, you know, went through

49:03

her three weeks of acting training

49:06

and whatnot. So that's just in

49:08

case decision dot sub stack.com. And

49:10

then my my core last call

49:12

is completely off. Not really analytics,

49:14

but over the holidays there was

49:16

reasons that I needed to explore

49:18

new podcasts and I did not

49:20

realize that Mike Brabiglia had a

49:22

podcast called Working It Out and

49:24

he has his David Sideris comes

49:26

on as kind of for his

49:28

second appearance and oh my god

49:30

like I I don't know of

49:32

like two more delightful people than

49:34

Mike Brabiglia and David Sideris on

49:36

the Working Out podcast so if

49:38

you're looking something for that is

49:40

just if you know either of

49:42

those guys and their sensibilities and

49:44

you're into them, that was like,

49:46

oh my God, just like heaven

49:48

of listening for 40 minutes or

49:50

however long it was. It has

49:52

nothing to do with analytics, but

49:54

had to put in a plug

49:56

for that as well. What about

49:58

you, Michael? You have six? Last

50:00

call? This is just a... What's

50:02

a for-fer? Is that a thing?

50:04

No. A quad? A quadfer. I've

50:07

got an octavefer for... No. It's

50:09

just really pushing the limits. I

50:11

have an AI agent that will

50:13

talk to your AI agent about

50:15

my last calls, so... a little

50:18

negotiate. No. So no, actually... So

50:20

one of the things I ran

50:22

into recently that I thought could

50:24

be a little bit helpful to

50:26

our audience is a lot of...

50:29

folks use Google Analytics for website

50:31

tracking and one of the things

50:33

that is kind of required to

50:35

make that tool useful at all

50:37

is to export it into big query.

50:40

But what's in big query is

50:42

often quite different than what's in.

50:44

GA4 and that's kind of hard

50:46

for business users. So Hamancho Sharma

50:48

actually did a mapping of those

50:50

metrics and the calculations to get

50:52

to those metrics pretty comprehensively from

50:54

GA4 to Big Query. So if

50:56

you're in that place where you're

50:58

trying to navigate that, that could

51:01

actually be quite a good resource

51:03

to use. And then alternatively,

51:05

you could also switch to a better

51:07

tool, but you know, in the meantime.

51:09

That is one that I would say

51:11

is one you could bookmark and use

51:14

as a good reference. So all right.

51:16

So I think this has been so

51:18

fun to kind of dive into this

51:20

conversation. Michael, so thank you so much

51:22

for joining us. It's been really cool.

51:24

I love kind of hearing your vision

51:26

for the future, the way you're thinking

51:29

about this, the way that the world

51:31

is progressing on these fronts is kind

51:33

of a very cool frontier that we're

51:35

on again, both with AI and the

51:37

connected health. and the Connect itself. So

51:39

it's, I really appreciate you, you know,

51:42

kind of sharing some of your thoughts

51:44

and things about that. And I'm sure

51:46

as our listeners have been listening, they

51:48

might have some thoughts and feelings about

51:50

this as well. We'd love to hear

51:53

from you. So go ahead, reach out

51:55

to us. The best way to do

51:57

that, probably on LinkedIn, or you can

51:59

connect. on the Measure Slack chat

52:01

group or also by email contact

52:04

at analytics hour. I. Oh, Michael,

52:06

do you, are you active on

52:08

social media at all? Could people

52:10

find you out there? Yeah, principally

52:13

find me on LinkedIn. The company

52:15

Fulgro Dynamics also on LinkedIn and

52:17

I'm on X with my old

52:19

school hacker handle of Kubla, KUBLA.

52:22

Get me up there and find

52:24

Fulgro on X as well. Awesome,

52:26

thank you. So you can reach

52:28

out to him as well and

52:31

follow him on those channels as

52:33

well. So you can hear what

52:35

the latest and greatest is in

52:37

this crazy changing world. All right,

52:40

well, hey, listen, one of the

52:42

things we're trying to do this

52:44

year is make sure that people

52:46

get access to this show. One

52:48

of the ways that you can

52:51

help with that is putting a

52:53

rating or review where you listen

52:55

to the show, whether it's Spotify

52:57

or Apple or wherever, we would

53:00

love to have you. Rate the

53:02

show review the show that really

53:04

helps apparently algorithmically Until AI can

53:06

take over and recommend us to

53:09

the right people I heart radio

53:11

is that really where we're yeah,

53:13

we're really targeting that one That's

53:15

our that's where our that's where

53:18

our listener lives is Let's package

53:20

that up. We'll get a series

53:22

A. Nope no time So but

53:24

yeah if you if you're listening

53:27

and you haven't done that before

53:29

We'd love it if you could

53:31

and we always just love hearing

53:33

feedback about the show as well

53:36

And it helps us think about

53:38

the future of the show so

53:40

really appreciated if you can and

53:42

no show would be complete without

53:45

a huge shout-out and a thank

53:47

you to our producer Josh Cohurst

53:49

for everything he's doing to make

53:51

the show possible so thank you

53:54

Josh And I am sure I

53:56

speak for both my co-hosts, Tim

53:58

and Julie. Thanks listening. Let's

54:00

matter what conversation what

54:03

you're measuring, just

54:05

remember, and questions on Twitter. Thanks

54:07

for listening. for listening. the Let's

54:09

keep the conversation going

54:11

with your comments, suggestions and

54:13

questions on questions on Twitter at

54:16

on the web on the web at

54:18

.io, Hour. Our Linked In Group, and and

54:20

the Measure ChatSlat Group. Music for

54:22

the podcast by by Josh Crowhurst.

54:24

So smart guys wanted to

54:26

fit in, so they

54:29

made up a term called

54:31

up a term don't work. Do

54:34

the analytics say go

54:36

for it, no matter who's

54:38

going for it? So

54:40

if you and I were

54:42

on the field, the

54:44

analytics it for it. It's

54:46

the stupidest, it? So if you and

54:48

thing I've ever heard for

54:50

reasoning the competition. say go for it

54:52

It's the stupidest laziest lamest thing I've heard for

54:54

keep analyzing in competition Music

55:02

podcast.