Ep 39 - Generative AI: Practical Strategies for Thinkydoers with Guest Dvorah Graeser

Sara: 00:00:01

Welcome to the Thinkydoers podcast.

Sara: 00:00:04

Thinkydoers are those of us drawn to deep work, where thinking is working.

Sara: 00:00:09

But we don't stop there.

Sara: 00:00:11

We're compelled to move the work from insight to idea, through the messy

Sara: 00:00:16

middle, to find courage and confidence to put our thoughts into action.

Sara: 00:00:21

I'm your host, Sara Lobkovich.

Sara: 00:00:23

I'm a strategy coach, a huge goal setting and attainment nerd, and board-certified

Sara: 00:00:30

health and wellness coach, working at the overlap of work, life well-being.

Sara: 00:00:35

I'm also a Thinkydoer.

Sara: 00:00:38

I'm here to help others find more satisfaction, less frustration, less

Sara: 00:00:43

friction, and more flow in our work.

Sara: 00:00:46

My mission is to help changemakers like you transform our workplaces and world.

Sara: 00:00:52

So let's get started.

Sara: 00:00:57

Welcome back to part two of my conversation with Dvorah Graeser.

Sara: 00:01:01

In our last episode, we explored some of the ethics and

Sara: 00:01:04

implications of AI at work today.

Sara: 00:01:07

Now, we're getting practical.

Sara: 00:01:10

We'll discuss some specific and maybe surprising ways to leverage AI tools,

Sara: 00:01:15

particularly for deep specialists and those of us wired strategically.

Sara: 00:01:21

If you're ready to move from anxiety to agency with AI while

Sara: 00:01:25

maintaining your ethics and humanity, this episode's for you.

Sara: 00:01:33

Dvorah, you mentioned kind of the shift from anxiety about AI to agency and

Sara: 00:01:39

how people can leverage AI around their creative vision, regardless of what

Sara: 00:01:45

their technical level of proficiency is.

Sara: 00:01:48

Tell me a bit more about how you see the use of AI enabling

Sara: 00:01:53

creative and multi-talented humans.

Dvorah: 00:01:56

Well, one way that I personally like to use AI—I don't know if

Dvorah: 00:02:00

you've gotten to this point when creating, written content, but I'll

Dvorah: 00:02:03

have a whole pile of thoughts, and they are all over the darn place.

Dvorah: 00:02:06

If they were sticky notes, they would be like Mount Everest of sticky notes.

Dvorah: 00:02:09

And I don't know what the heck to do with them, and I get very frustrated by this.

Dvorah: 00:02:13

One of the things that I have found that can be great to do is I shove the

Dvorah: 00:02:17

whole thing into, let's say, Claude.

Dvorah: 00:02:19

Now, Claude has a feature where you can copy and paste, but you can also upload.

Dvorah: 00:02:25

So, let's say you have a PDF and you have some written documents

Dvorah: 00:02:28

and some other stuff, and you just dump the whole thing into Claude.

Dvorah: 00:02:31

And I just say Claude exactly what I'm feeling — I don't know how to organize

Dvorah: 00:02:35

this, like, if I can't make sense of this, or I have one thought, I will write the

Dvorah: 00:02:40

thought, but the rest of it's not clear.

Dvorah: 00:02:41

What do I do?

Dvorah: 00:02:42

And then Claude very calmly chugs along and proposes something.

Dvorah: 00:02:47

They propose an outline.

Dvorah: 00:02:49

It can be helpful if you ask it for something specific, like an

Dvorah: 00:02:51

outline or whatever, but even if you don't, it'll still do its best.

Dvorah: 00:02:55

And then you start a dialogue.

Dvorah: 00:02:57

And this is one of the things I really do like about generative AI.

Dvorah: 00:03:00

We have a dialogue, we go back and forth.

Dvorah: 00:03:02

I ask for constructive feedback because otherwise, it'll only say nice things.

Dvorah: 00:03:06

It's something about the generative AI.

Dvorah: 00:03:08

If you ask for constructive feedback, it'll also say the mean things.

Dvorah: 00:03:10

But I'm going back and forth, and as I'm going back and forth, in my

Dvorah: 00:03:14

own mind, I'm realizing, I don't like that bit, I do like this bit.

Dvorah: 00:03:18

Oh, I just thought of something completely different.

Dvorah: 00:03:20

Now, what went from a blizzard of ideas I couldn't organize will go to an

Dvorah: 00:03:25

organization that maybe isn't exactly what I want, but helps clear my mind,

Dvorah: 00:03:29

reduce my own anxiety, because it's often my anxiety that's leading to the

Dvorah: 00:03:33

pile-o'-sticky-notes problem, at least in my experience — and enables me to get to

Dvorah: 00:03:36

the point where I can do the organization.

Dvorah: 00:03:39

Even if it turns out that it's not Claude that has made the

Dvorah: 00:03:42

good proposal by interacting with it, I get to where I need to go.

Sara: 00:03:46

That is the use case that has been the most transformative for me.

Sara: 00:03:51

And that's where I went from being a skeptic who was not likely to integrate

Sara: 00:03:56

these tools to being a hooked user.

Sara: 00:04:00

They don't replace the human brain.

Sara: 00:04:02

You mentioned hallucinations in our first episode.

Sara: 00:04:06

I don't rely on models for facts or information, but they are really good

Sara: 00:04:12

at organizing big messes of information.

Sara: 00:04:17

The thing that frustrates me is sometimes they're poor at actually organizing data.

Sara: 00:04:23

Even when I train it on data and want data analysis or data, you know,

Sara: 00:04:29

stuff, that's been a big frustration for me with the current tools.

Sara: 00:04:33

Even when I train on data, it can still hallucinate, which is a bummer.

Sara: 00:04:38

That's the kind of stuff that I would think the models would be

Sara: 00:04:42

built to do really well, and with as little hallucination as possible.

Sara: 00:04:46

But maybe there's always going to be some hallucination.

Dvorah: 00:04:50

There's a couple of ways around hallucination, right?

Dvorah: 00:04:53

And this again gets into process.

Dvorah: 00:04:54

We talked about it in the first, episode, but it's still important to bring up.

Dvorah: 00:04:58

You need to have a process for dealing with this.

Dvorah: 00:05:00

So let us suppose you uploaded a bunch of data to, I don't

Dvorah: 00:05:03

know, let's say it was ChatGPT.

Dvorah: 00:05:04

It doesn't matter.

Dvorah: 00:05:05

You uploaded it, and it's still hallucinated.

Dvorah: 00:05:07

So what are ways to check this?

Dvorah: 00:05:09

One way to check this is to ask ChatGPT in the same session to go back, and to check

Dvorah: 00:05:16

its answers in comparison to the data and show where it got each answer from.

Dvorah: 00:05:20

This is also great for summarization, by the way.

Dvorah: 00:05:22

If it summarizes and it comes up with something, you say, "Okay,

Dvorah: 00:05:25

where did that really wonderful, great quote, where is it?

Dvorah: 00:05:28

Show me where it is." And they're like, "oh yeah, you're right, I

Dvorah: 00:05:30

kind of made that up." Okay, no problems, let's just move on.

Dvorah: 00:05:33

A third way for doing it is then to actually go to another tool like

Dvorah: 00:05:41

Perplexity and ask it to do research.

Dvorah: 00:05:43

So maybe you don't want to upload your sensitive data to Perplexity,

Dvorah: 00:05:46

but you can still ask it to do research on the area and give you some

Dvorah: 00:05:50

ideas of what the boundaries of the area that can help corral ChatGPT.

Dvorah: 00:05:54

Or you can go to a completely different tool.

Dvorah: 00:05:56

I guess the completely different tool is the third way of doing it.

Dvorah: 00:05:58

Let's say you go from ChatGPT to Claude.

Dvorah: 00:06:00

You can go to Claude, and then you can say, "Hey, can you please check this?

Dvorah: 00:06:03

And what is good and what is bad, and please give me constructive feedback."

Sara: 00:06:07

That's wild.

Sara: 00:06:08

In my early experimentation, when I was working on my book, there was some

Sara: 00:06:12

research I had, like actual research.

Sara: 00:06:15

I needed academic journal research.

Sara: 00:06:17

that was one of the things that was the most frustrating.

Sara: 00:06:20

I would think that models would be good at knowing that truthful data.

Sara: 00:06:26

Hopefully, this has changed since then.

Sara: 00:06:29

But I would ask for research, like, "What are the leading papers or

Sara: 00:06:33

core papers on a specific topic?" And it would make up papers.

Sara: 00:06:38

You've got ProQuest.

Sara: 00:06:39

It's all there.

Sara: 00:06:41

This you should get right.

Dvorah: 00:06:43

So, for example, Perplexity AI, I like because it gives links.

Dvorah: 00:06:46

And you can go check them.

Dvorah: 00:06:47

ChatGPT sometimes does, sometimes doesn't, sometimes hallucinates,

Dvorah: 00:06:50

sometimes it's accurate.

Dvorah: 00:06:51

I'm never quite sure what it's going to do next, so I find that a little unreliable.

Dvorah: 00:06:55

There are some specialized GPTs for papers.

Dvorah: 00:06:58

There's one called Scholar AI, and that's a good one, but

Dvorah: 00:07:01

there are other ones as well.

Dvorah: 00:07:02

So here, I think it's good to join with others to understand what they've

Dvorah: 00:07:07

done and what's worked for them.

Dvorah: 00:07:08

And that can change over time.

Dvorah: 00:07:09

So, for example, I like Scholar AI, but it's possible in a few months,

Dvorah: 00:07:12

someone's going to come back.

Dvorah: 00:07:13

Maybe they'll listen to this again.

Dvorah: 00:07:15

They'll be like, "Oh, but it's no longer good." Possible.

Dvorah: 00:07:17

Things change really quickly.

Dvorah: 00:07:19

So I think it's helpful to try out different specialty tools, to talk with

Dvorah: 00:07:22

other people about this on LinkedIn or another social media that you like.

Dvorah: 00:07:26

Mastodon, if you're into that.

Dvorah: 00:07:28

And try to get information about what is working right now and what is the process

Dvorah: 00:07:32

for working around some of these issues.

Dvorah: 00:07:35

So if I get anything from ChatGPT, because I have found it to be

Dvorah: 00:07:38

the most erratic when it comes to hallucinating, I will copy and paste

Dvorah: 00:07:41

the whole answer to Perplexity AI.

Dvorah: 00:07:43

And I'll say, "Is this right or wrong?

Dvorah: 00:07:45

Please give me feedback.

Dvorah: 00:07:47

Please find me references.

Dvorah: 00:07:48

What do you think?" And it will go through the whole thing and it'll say,

Dvorah: 00:07:51

"Well, this looks like it's correct.

Dvorah: 00:07:52

Couldn't find anything for this.

Dvorah: 00:07:53

I really don't know about that bit over there.

Dvorah: 00:07:55

And this is what I would do instead." And so Perplex AI will actually

Dvorah: 00:07:58

give you a long, reasoned answer.

Dvorah: 00:08:01

This was without going into their deep research mode, which is awesome.

Dvorah: 00:08:04

But that's like a whole nother kettle of fish.

Dvorah: 00:08:05

It's just regular Perplexity AI.

Dvorah: 00:08:07

It'll actually go through these things.

Dvorah: 00:08:09

So I guess it's unlike Google in the sense where if you get a link and you

Dvorah: 00:08:13

click on the link, the link exists.

Dvorah: 00:08:15

If you get a summary with the link, does the link exist?

Dvorah: 00:08:17

Does it not exist?

Dvorah: 00:08:18

Does the summary say what the link says?

Dvorah: 00:08:20

Not always.

Dvorah: 00:08:21

So here we need to have more of a process for checking for these issues.

Dvorah: 00:08:27

It can still, however, save a lot of time because if you have a multi-tool

Dvorah: 00:08:31

process and you leverage the strengths of one tool to shore up the weaknesses

Dvorah: 00:08:36

of another tool, then you can save a lot of time, but also a lot of mental energy,

Dvorah: 00:08:40

which you can then spend on things like having deep human thoughts, not corralling

Dvorah: 00:08:44

the facts, but understanding what they mean from a human perspective, which

Dvorah: 00:08:48

no AI will really ever be able to do.

Sara: 00:08:50

Yeah, you mentioned in our first conversation the benefits of these tools

Sara: 00:08:58

for people with deep specialization.

Sara: 00:09:02

And that's my listeners, like, we're the ones that go deep.

Sara: 00:09:05

And I definitely see that in my own work, that using these

Sara: 00:09:11

tools helps me make my work more accessible and more understandable

Sara: 00:09:16

to people, even as a deep specialist.

Sara: 00:09:20

They help me, if I need to zoom out and I can't do that on my own then

Sara: 00:09:25

that's something the tools are really good at — helping me figure out how

Sara: 00:09:29

to communicate with fill in the blank.

Sara: 00:09:31

Like, if I need to communicate with a general business audience that

Sara: 00:09:34

isn't my ICP or my ideal customer profile, so an audience that is

Sara: 00:09:41

majority neurotypical, for example.

Sara: 00:09:43

Most of my people are neurodivergent.

Sara: 00:09:45

And that's who I communicate pretty well with.

Sara: 00:09:48

But if I have to communicate with large groups of neurotypical people or

Sara: 00:09:52

just general population, then I have to adapt my communication style and

Sara: 00:09:58

fidelity and altitude to that audience.

Sara: 00:10:01

And that's where I've found the tools super useful.

Sara: 00:10:06

But you mentioned how powerful these tools are for people with deep specialties.

Sara: 00:10:12

I'd love to hear more about that.

Dvorah: 00:10:14

Well, just quickly going back to your idea of the persona.

Dvorah: 00:10:18

So let's suppose you want to talk to a persona that isn't naturally you, or that

Dvorah: 00:10:22

you don't often talk to, and you want to be able to communicate better with them,

Dvorah: 00:10:26

or maybe you just want to check what you're doing against a potential persona.

Dvorah: 00:10:30

There's a lot of work that's been done on buyer personas.

Dvorah: 00:10:33

HubSpot has some templates, no affiliation, just, templates.

Dvorah: 00:10:36

There's lots of good books on that.

Dvorah: 00:10:38

What I would actually recommend is learn about buyer personas

Dvorah: 00:10:41

from the marketing people.

Dvorah: 00:10:43

And you can literally upload your buyer persona into ChatGPT or

Dvorah: 00:10:47

Claude and say, "This is my persona.

Dvorah: 00:10:50

I need to communicate with this person.

Dvorah: 00:10:52

This is what I wrote." And if you have some examples— "And here's some examples

Dvorah: 00:10:57

of the kind of tone I'm trying." Not so much in terms of the tone, but you

Dvorah: 00:11:01

could actually copy-paste articles in.

Dvorah: 00:11:03

"I like these articles, I like their tone. Gee, I wish I could write like

Dvorah: 00:11:06

that. Can you help me with that?" Learning about buyer personas from the

Dvorah: 00:11:10

marketing folks and then putting that into the AI can really be helpful.

Dvorah: 00:11:14

And this can also help deep specialists if they have to speak

Dvorah: 00:11:17

with the not deep specialists.

Dvorah: 00:11:18

When people are super knowledgeable about something and they have to talk with

Dvorah: 00:11:21

someone who is maybe not knowledgeable at all about that, but that person needs

Dvorah: 00:11:25

to know and they need to communicate, that can be really tough, because

Dvorah: 00:11:28

it's really easy to make assumptions.

Dvorah: 00:11:29

I'm totally guilty of this, done it a lot.

Dvorah: 00:11:31

So there.

Dvorah: 00:11:33

Again, having the persona, putting it into the ChatGPT, and then making

Dvorah: 00:11:38

clear that you want to have this be easily readable at, say, like a high

Dvorah: 00:11:44

school level, but with words that this persona would understand, would

Dvorah: 00:11:49

want to use, that is hugely helpful.

Dvorah: 00:11:51

And then, you can actually start with one tool, Or one session, let's say, of

Dvorah: 00:11:57

Claude, and bring it into another one.

Dvorah: 00:11:59

So let's say I use Claude once, then I'll copy-paste everything into

Dvorah: 00:12:02

another Claude session, I'll say, "Please give me feedback. Constructive

Dvorah: 00:12:05

feedback." Otherwise, again, it won't give you the mean stuff.

Dvorah: 00:12:08

But yet another way that it can be helpful is, let's suppose you have done this work,

Dvorah: 00:12:12

and you now have a session in Claude, where it is really helping you produce

Dvorah: 00:12:18

what you need to take that next step.

Dvorah: 00:12:21

Keep that session going.

Dvorah: 00:12:23

Stay in the same session.

Dvorah: 00:12:24

Now, Claude will keep on saying things like, "This is getting long!"

Dvorah: 00:12:27

I hit "Continue." "This is really long, I really mean it." Continue,

Dvorah: 00:12:31

continue, continue, continue.

Dvorah: 00:12:32

Claude and I have a long time argument about that.

Dvorah: 00:12:35

And I know why that is, because it's hard for them to keep the whole session in.

Dvorah: 00:12:37

But once I have actually Claude communicating with me in a way that helps

Dvorah: 00:12:41

me, with wordsmithing or whatever it is.

Dvorah: 00:12:44

Keep going in that same session and then you will find that over time

Dvorah: 00:12:48

it becomes closer and closer to your ideal output of whatever type that is.

Dvorah: 00:12:54

Is it a good summary?

Dvorah: 00:12:55

Is it a way of wording things that is more better or more to the point

Dvorah: 00:12:59

of what you want in your opinion than what you were doing by yourself?

Dvorah: 00:13:02

Whatever it is, keep going in that same session and don't let it

Dvorah: 00:13:06

distract you with another session.

Dvorah: 00:13:07

Say "Continue" all the time, and then you will find that you and the AI

Dvorah: 00:13:13

are really moving together as deep specialists, but with the output that

Dvorah: 00:13:18

is focused in the way that you want.

Sara: 00:13:21

Yeah, I am also a Claude user.

Sara: 00:13:23

That was the one that, for me, felt the most acceptable in

Sara: 00:13:28

terms of policies and ethics.

Sara: 00:13:31

I see a lot of the attention being paid to have these tools, write

Sara: 00:13:36

your marketing copy for you, or have these tools, God forbid, write a

Sara: 00:13:41

book like, you know, those things.

Sara: 00:13:44

And I don't use them for that at all.

Sara: 00:13:47

I use them for better understanding my thinking and the thinking of the people

Sara: 00:13:55

I interact with who give me permission to add their, verbal processing or

Sara: 00:13:59

meeting transcripts to my models.

Sara: 00:14:02

I think partly too, because I am neurodivergent, I'm aware I have

Sara: 00:14:05

gaps in my executive functions.

Sara: 00:14:08

I think of Claude as almost an adaptive assistive technology for

Sara: 00:14:12

me, in helping me work with some of those executive functioning gaps that

Sara: 00:14:18

I'm aware of and that I know I have.

Sara: 00:14:20

So, are there any ways that you use the tools that you're surprised about?

Dvorah: 00:14:26

I would actually say, first of all, I discussed the

Dvorah: 00:14:30

pile of sticky notes thing.

Dvorah: 00:14:31

I never expected that to work.

Dvorah: 00:14:32

I just simply thought it would fall down.

Dvorah: 00:14:34

It works really, really great for that.

Dvorah: 00:14:36

When it comes to having the AI do something like write your marketing

Dvorah: 00:14:40

copy, the problem is people mean different things by that.

Dvorah: 00:14:44

So for example, I would not have it write my marketing copy, but why is that?

Dvorah: 00:14:48

Well, first of all, I'm curious, who are my people?

Dvorah: 00:14:51

How do I want to talk with them?

Dvorah: 00:14:52

How do they talk amongst themselves?

Dvorah: 00:14:53

What do they say to each other?

Dvorah: 00:14:55

What are their concerns?

Dvorah: 00:14:56

I really want copy that is going to communicate with them in a way that

Dvorah: 00:15:00

they want to be communicated with.

Dvorah: 00:15:02

One of the things that generative AI has helped me with is understanding

Dvorah: 00:15:05

how to break that down into processes.

Dvorah: 00:15:08

With verbal communication, I have less of an issue.

Dvorah: 00:15:11

I try to listen to the person, don't I?

Dvorah: 00:15:12

But try to like have it be an exchange of thoughts and all that.

Dvorah: 00:15:15

But when it comes to writing, I tend to default to me.

Dvorah: 00:15:18

I am the audience.

Dvorah: 00:15:19

Nooo.

Dvorah: 00:15:20

Now, ChatGPT and Claude will do the same thing.

Dvorah: 00:15:23

They default to corporate speak.

Dvorah: 00:15:24

Because that's what they think the default should be.

Dvorah: 00:15:27

So if you don't let them know about what you want, they're going to default.

Dvorah: 00:15:31

But then again, we do that as humans.

Dvorah: 00:15:32

So I don't think we can really blame the generative AI for this.

Dvorah: 00:15:34

Everyone's going to default to their comfort zone.

Dvorah: 00:15:37

What's great about generative AI is it helped me explore my comfort zone

Dvorah: 00:15:41

and break down generating marketing content into more pieces, which I then

Dvorah: 00:15:46

have an easier time tackling one on one and bringing it together for something

Dvorah: 00:15:51

which will appeal to the people I want to appeal to who are not me.

Sara: 00:15:55

Yeah, for me, this is a bit of a disclosure, but for me, the

Sara: 00:15:59

other use that I never would have imagined is I use it when I'm over a

Sara: 00:16:06

threshold, like when I'm under stress.

Sara: 00:16:09

I never imagined I would talk to one of these things about anything other

Sara: 00:16:13

than work, but it's almost become part of my mindfulness practice.

Sara: 00:16:17

Now, one of my tools is when something happens and I'm upset, or when something

Sara: 00:16:22

happens and I'm over threshold, or especially when I'm having conflict with

Sara: 00:16:26

someone else, I can sit down and say, "Okay, tool, here's what's happening.

Sara: 00:16:30

Here's how I feel." And they're remarkably empathetic, the ones that I use.

Sara: 00:16:36

Not a substitute for therapy, not a substitute for a coach or speaking with

Sara: 00:16:40

another human being, but that in the moment dysregulation that I experienced,

Sara: 00:16:47

they are helpful with me finding a way to regulate myself through that conversation.

Sara: 00:16:53

And that's a use I never would have imagined.

Dvorah: 00:16:55

I agree with that point.

Dvorah: 00:16:57

I would divide the into two different aspects.

Dvorah: 00:16:59

So when it's something like Claude, it's kind of like, I will talk things

Dvorah: 00:17:03

through with myself internally.

Dvorah: 00:17:04

So, Claude is really just simply mirroring back.

Dvorah: 00:17:07

In that case, I'm having an internal conversation with an external device

Dvorah: 00:17:11

that is kind of mirroring back.

Dvorah: 00:17:13

There are different kinds of LLMs out there who do try to do

Dvorah: 00:17:18

coaching or more therapeutic things.

Dvorah: 00:17:21

I'm always a bit nervous about that because I feel

Dvorah: 00:17:25

that expectations are higher.

Dvorah: 00:17:27

And even tools which are designed for this can still get themselves into trouble.

Dvorah: 00:17:33

Character AI, people were supposed to have these tools to be like AI companions.

Dvorah: 00:17:36

And there've been some really sad cases, often involving minor children,

Dvorah: 00:17:41

but quite frankly, some of these sad cases often involve adults.

Dvorah: 00:17:43

It's just when it's an adult, you can't do anything about it, right?

Dvorah: 00:17:45

They say, "Oh, it's just the way it is," Which I don't agree with.

Dvorah: 00:17:48

But when it's minor children, then people certainly do something about it.

Dvorah: 00:17:51

And there's some very sad stories about suicides and other issues

Dvorah: 00:17:54

involving the use of these.

Dvorah: 00:17:56

With children, it's one thing, I personally just wouldn't let

Dvorah: 00:17:58

children anywhere near them.

Dvorah: 00:17:59

That's what I think.

Dvorah: 00:18:01

But even with adults, I don't think we already always realize that

Dvorah: 00:18:05

these are probability machines.

Dvorah: 00:18:07

If we think about it as having an internal conversation with

Dvorah: 00:18:10

ourselves, I think that's one thing.

Dvorah: 00:18:12

But if you're talking to a probability machine that's setting itself up to

Dvorah: 00:18:15

be a therapist, I get very nervous.

Dvorah: 00:18:18

And the reason why is it may not pick the right thing to say in that context.

Dvorah: 00:18:22

And, you know, humans in some ways are kind of delicate systems.

Dvorah: 00:18:26

We have all these little bits, moving parts and all these things going on.

Dvorah: 00:18:28

We should be nice to ourselves and we should be kind to ourselves.

Dvorah: 00:18:32

And when we're doing that, we have to think a little bit if we're interacting

Dvorah: 00:18:36

with this tool, is that being kind to ourself and are we doing it in a kind way?

Sara: 00:18:40

I think you're spot on.

Sara: 00:18:41

And the other thing that comes to mind as you're talking is, I'm

Sara: 00:18:44

using the tool to talk to myself, kind of like you described, it's

Sara: 00:18:47

like, it's an external thing I'm talking to, but it's an extension

Sara: 00:18:52

of my own awareness and mindfulness.

Sara: 00:18:55

When we work with therapists, there's a reason even with therapists, we do this

Sara: 00:19:02

with video, is that then they can observe all of the nonverbal communication and

Sara: 00:19:08

the tools at this point can't observe the nonverbal communication and make

Sara: 00:19:12

decisions about how to make that kind of calculation of how to respond.

Sara: 00:19:17

So I'm with you.

Sara: 00:19:18

I bring it up, you know, helpful in terms of exactly what you said, like kind of

Sara: 00:19:24

having a conversation with your better self when you're in a time of need.

Sara: 00:19:29

Dvorah, I don't want to end this, but we've got to wrap it up.

Sara: 00:19:32

or If you were going to tell someone who is anxious about generative

Sara: 00:19:37

AI tools, that shift from anxiety to agency, what would you say?

Sara: 00:19:43

What would you coach them as a place to start?

Dvorah: 00:19:46

As a place to start, I would actually begin by

Dvorah: 00:19:49

looking at how I feel about AI.

Dvorah: 00:19:51

I actually had to do that.

Dvorah: 00:19:53

I had to get emotional about AI before I could use it well.

Dvorah: 00:19:55

And it's simply because it is a very different tool.

Dvorah: 00:19:58

It is not going to do what you always expect.

Dvorah: 00:20:00

And so, for that reason, you need to get your own emotional house in order

Dvorah: 00:20:03

with it before you can start using it.

Dvorah: 00:20:05

And that is actually what I would recommend.

Dvorah: 00:20:08

Now, I actually sent you a link to something, which I think you're going

Dvorah: 00:20:12

to put in the show notes, which has a PDF where you can actually do that.

Dvorah: 00:20:16

Anyone listening to this, hit me up.

Dvorah: 00:20:19

We can hop on a call.

Dvorah: 00:20:20

We can talk about this.

Dvorah: 00:20:21

You can look at my PDF.

Dvorah: 00:20:22

There are resources out there.

Dvorah: 00:20:24

They're not as many that about emotions in AI.

Dvorah: 00:20:27

And I think that's a mistake.

Dvorah: 00:20:28

I think that people see it only as a technology, but if we want

Dvorah: 00:20:32

to use it well, it can't only be a technology like Microsoft Word.

Dvorah: 00:20:36

We have to be a little more vulnerable ourselves, and a little

Dvorah: 00:20:39

more open ourselves to using it, which means that we need to think

Dvorah: 00:20:41

about ourselves as we're using it.

Sara: 00:20:44

Awesome answer.

Sara: 00:20:45

Well, Dvorah, speaking of, if people want to learn more about

Sara: 00:20:50

what you do or get connected with you, where can they find you?

Dvorah: 00:20:53

They can find me on LinkedIn.

Dvorah: 00:20:55

D Graeser, G R A E S E R, I'm used to spelling my name.

Dvorah: 00:20:59

But I'm also one of the few Dvorahs on LinkedIn, so you can definitely find me.

Dvorah: 00:21:03

There'll be the link in the show notes.

Dvorah: 00:21:04

So you can definitely hit me up there.

Dvorah: 00:21:06

Those are the best places to look for me because I am not really on

Dvorah: 00:21:10

any other form of social media.

Dvorah: 00:21:11

Although I know you're trying to get me on Mastodon.

Dvorah: 00:21:13

So maybe I'll add that.

Sara: 00:21:15

I think that you would fit in just perfectly there.

Sara: 00:21:20

It really is the only thing I do that feels like old internet.

Sara: 00:21:23

It makes me want to type a zine on my typewriter and send it

Sara: 00:21:26

out by mail, like for real.

Sara: 00:21:29

well, Dvorah, thank you so much.

Sara: 00:21:30

This has been a wonderful conversation, and this is when

Sara: 00:21:33

Dvorah is going to be back.

Sara: 00:21:34

So, if there are questions that you have that we didn't answer, send them in.

Sara: 00:21:39

I'd love to hear them and we'll schedule a follow-up.

Sara: 00:21:41

With Dvorah to get more of your questions answered.

Sara: 00:21:44

Thank you so much, Dvorah.

Dvorah: 00:21:46

Thanks.

Dvorah: 00:21:47

Have a great day.

Dvorah: 00:21:48

Bye.

Sara: 00:21:48

Bye.

Sara: 00:21:51

Thank you for joining us for this exploration of practical AI

Sara: 00:21:54

applications with guest Dvorah Graeser.

Sara: 00:21:57

You can find Dvorah on LinkedIn at D G R A E S E R. And we'll include that

Sara: 00:22:05

link in our show notes at findrc.co/pod.

Sara: 00:22:09

We've also got a link to the resource on emotions in AI that

Sara: 00:22:13

she mentioned in the show notes.

Sara: 00:22:15

I know this episode is going to spark questions.

Sara: 00:22:17

We would love to hear them.

Sara: 00:22:19

So if you have questions about AI use or LLMs or other things having to do with

Sara: 00:22:26

generative AI, please send them to me.

Sara: 00:22:28

Please share them with me.

Sara: 00:22:30

We would love to have Dvorah back to answer them in a future episode.

Sara: 00:22:33

And she and I talked about it after we stopped recording.

Sara: 00:22:36

We welcome questions of any kind.

Sara: 00:22:38

So, they don't all have to be pro-AI.

Sara: 00:22:41

As long as you're asking curiously and in good faith, we'll do our

Sara: 00:22:45

best to provide a helpful answer.

Sara: 00:22:47

And don't forget, if you haven't already, I'd love to have you on the

Sara: 00:22:50

launch squad for my upcoming books.

Sara: 00:22:52

I need advanced readers, people willing to write reviews if you like the book,

Sara: 00:22:58

social media amplifiers, and really Just for folks to be excited, to help

Sara: 00:23:03

me get over the days where I don't feel so excited, or I feel scared

Sara: 00:23:08

to get this book into the world.

Sara: 00:23:10

If that might be you, join the squad at findrc.co/launchsquad, and I cannot

Sara: 00:23:16

wait to get this book into your hands.

Sara: 00:23:19

Thank you for listening.

Sara: 00:23:24

All right, friends, That's it for today.

Sara: 00:23:27

stay in the loop with everything going on around here by

Sara: 00:23:30

visiting findrc.co/newsletter and joining my mailing list.

Sara: 00:23:36

Got questions?

Sara: 00:23:37

My email addresses are too hard to spell, so visit findrc.co/contact

Sara: 00:23:43

and shoot me a note that way.

Sara: 00:23:45

You'll also find me at @saralobkovich on most of your

Sara: 00:23:49

favorite social media platforms.

Sara: 00:23:50

For today's show notes, visit findrc.co/thinkydoers if there's

Sara: 00:23:57

someone you'd like featured on this podcast, drop me a note.

Sara: 00:24:00

And if you know other Thinkydoers who'd benefit from this episode, please share.

Sara: 00:24:05

Your referrals, your word of mouth, and your reviews are much appreciated.

Sara: 00:24:09

I'm looking forward to the questions this episode sparks for you, and I

Sara: 00:24:13

look forward to seeing you next time.