Artem Zhiganov, Harmonica: How to improve collective sense-making with GenAI
Ep. 02

Artem Zhiganov, Harmonica: How to improve collective sense-making with GenAI

Episode description

A conversation with the creator of Harmonica, an open-source tool using AI to support collective sensemaking in online communities. We explore the origins of the project, its design for asynchronous dialogue, and the vision behind building democratic infrastructure without relying on venture capital.

Download transcript (.srt)
0:00

can

0:06

i'd say

0:10

so we are here with from atom

0:12

thick enough

0:15

and today we love we talk about

0:18

the or monica and disease or another

0:21

episode of their democracy innovator podcast

0:25

so thank you hurt them for your

0:28

time and they're very clear for having

0:30

me

0:33

i'm very curious about the or monica

0:36

and so the first question is that

0:38

what is the actually or monica

0:46

i can tell he goes for monica

0:47

but maybe he wants me to give

0:49

you some quick intro like my background

0:54

also your background and whatever like you

0:58

like

1:00

okay so my bedroom and isn't a

1:03

product marketing i used to do like

1:05

a lot of use a resurgence communication

1:10

strategies and

1:13

and eventually product strategies for clients like

1:18

product innovation you know like defense prince

1:22

and customer experience design then i got

1:27

very tired of all that and i

1:29

decided to make like a pivot and

1:35

find something that would be more that

1:38

with more that resonate with me more

1:40

and that sir why i decided to

1:43

or get a liquor some indication that

1:48

like

1:50

a degree and then i went to

1:52

the uk where i studied

1:55

innovation management and

1:59

part of that program was writing a

2:01

thesis and they decided to write my

2:02

thesis about

2:04

cooperatives and then i learned about commons

2:08

elinor ostrom

2:12

and i realized that sir this is

2:14

like super

2:17

profound

2:19

ah theory that she was working on

2:22

and in her research and that took

2:26

forty years maybe and eventually she'll got

2:29

the nobel prize for the researcher was

2:31

the first woman to ever gets our

2:33

nobel prize in economics so elinor ostrom

2:36

was a great great sir political

2:40

our political scientist may be political economists

2:45

anthropologist who believed that sir

2:49

community shared resources or communities with shared

2:53

resources is an alternative to

2:59

and markets and states so basically this

3:02

is like a third

3:04

the economic paradigm that is very different

3:07

from both companies like private companies and

3:11

also the states

3:13

the state the beasts

3:17

management all you know state owned companies

3:20

all planned economy and so on

3:24

so that that was very very inspiring

3:26

for me and that's what how i

3:29

got into web three i was a

3:31

governance lead in a row and i

3:34

consulted a few does

3:36

after that

3:38

and that's how i got into meta

3:41

gov

3:42

because not to golf is like this

3:45

the community that is researching cutting edge

3:47

governance and web three governance in particular

3:49

and i was learning a lot from

3:52

met the gulf

3:54

and the also i got into newspeak

3:57

house which is a very interesting place

3:58

in london where they can have try

4:01

to

4:04

cross-pollinate ideas that are related to governance

4:09

and open source and civic tech and

4:12

the basically like intersection of democracy and

4:16

technology

4:18

i

4:19

that it's one of the most interesting

4:21

places in the world they think in

4:24

for the people who are interested in

4:25

democracy and technology and that's where harmonica

4:30

started so i i came up with

4:32

idea for harmonica while living in you

4:34

speak house in london

4:37

and also working at at as a

4:39

governance lead in a small doubt

4:43

and as the governance lead i was

4:44

first i was really struggling with

4:48

you know getting some input from the

4:49

community

4:51

i like trying to get people comments

4:55

on proposals in the forum right those

4:57

proposals you know join the workshops and

5:01

can contribute during those workshops and fill

5:05

out the surveys that i asked them

5:07

to feel

5:09

and all that was very very difficult

5:11

like people didn't want to do anything

5:13

they didn't want to write proposals they

5:15

didn't want to discuss proposals they didn't

5:17

want to fill out the forms that

5:18

didn't want to join workshops was like

5:20

hurting the cat for so at the

5:23

time i didn't know that but basically

5:25

this is what's or

5:28

i call sense making right now

5:31

so

5:34

oh when there was

5:36

you know did during my residency at

5:39

newspeak house john g p t came

5:41

out and it was like a very

5:44

very big topic like follow a lot

5:46

of conversations that happens in les twenty

5:50

two and twenty three

5:54

and which some of those i was

5:56

part of and so we discussed how

5:59

language models could be used to improve

6:03

governance of communities

6:06

and that's again the towel harmonica started

6:09

like i thought that maybe we could

6:10

use

6:12

generative a i to help with sensemaking

6:15

to help

6:17

the community managers governance leads and everyone

6:23

else who has to act as facilitator

6:27

not necessarily being professional facilitator but everyone

6:32

you know who is managing a community

6:34

that has to facilitate discussions from to

6:37

the time they have to facilitate

6:40

and so how generative a i could

6:42

help them how can we empower those

6:45

people the in those roles and to

6:49

or basically do

6:53

decent the collective sense making

6:57

so that's what we're monica is about

7:04

okay it's and i'll ever see now

7:07

i was looking at about your curriculum

7:08

that is the i will say alpha

7:11

economic and the and of yeah innovation

7:15

management and

7:18

and it's interesting it will be interesting

7:21

to know ah how you develop this

7:26

product i mean i deserve the software

7:30

was in terms of her

7:33

i've seen that he applied for some

7:35

grants and the that was the something

7:39

i think that allow you to build

7:40

the platform

7:44

but

7:46

it it will be interesting i think

7:48

for the people that want to build

7:49

the a platform for certain our little

7:52

bitter about the the backgrounder because the

7:56

ah

7:58

yeah your background in economic i think

8:00

it made the in a way for

8:03

you

8:04

it was may be easier for you

8:06

to find the summer and

8:12

some money to sponsor the project or

8:16

not not know first of all you'll

8:18

have background the the comics i started

8:20

in high school of economics that's the

8:22

name of the university that i studied

8:24

boomer added when i got my bachelors

8:26

degree

8:27

and it was like long time ago

8:29

twenty years ago

8:30

do

8:32

i graduated in two thousand and salad

8:35

but

8:38

i i didn't study economics i started

8:41

management and then a once i i

8:45

asked worked in advertising and then i

8:47

worked in product management and

8:52

i tried to build some startups before

8:55

harmonica but sir

8:57

i never

8:59

you know it was more like experiments

9:02

and i didn't

9:06

raise any like a venture capital for

9:08

those

9:10

sometimes i just invested my own saving

9:12

sometimes was like friends and family and

9:16

with harmonica it's actually the same

9:19

i invested them my savings and i

9:21

ain't got some grunts got some donations

9:24

from friends or through give us which

9:28

is a

9:30

platform for crypto donations and by the

9:33

way we are participating in the quadratic

9:36

funding round on giveth right now and

9:38

we would really appreciate even like one

9:41

dollar donation if anyone can do that

9:44

before fourteenth of february and and harmonica

9:48

is open-source try

9:50

correct

9:51

so

9:53

while we discussed disgusted a few times

9:56

if it's possible to fundraise i mean

9:59

to raise venture capital for open source

10:01

projects and i think there are some

10:03

examples of that i think this but

10:06

basically it's possible but you know after

10:11

reading elinor ostrom and

10:14

he going into this kind of posts

10:16

capitalists the

10:19

and

10:21

thinking was his school of thought not

10:24

only i mean not only thanks to

10:27

all eleanor but also some other

10:30

the researchers and thinkers like michelle bounds

10:33

for instance from peter pi foundation like

10:36

i i became very

10:40

disillusioned in venture capital like i really

10:43

don't want to go that route i

10:45

want harmonica to be like to open

10:47

source

10:49

with without this kind of growth imperative

10:53

or for profit

10:57

imperative like i don't want to to

10:59

build harmonica for product which makes it

11:01

much harder to build it like we

11:05

we won a few grants are actually

11:08

to grunts the through get coin and

11:11

i think one of them was like

11:14

three thousand dollars and other one was

11:16

like seven thousand so that's not a

11:19

lot of money and i hope that

11:22

we will

11:24

be figure out how to make it

11:28

more in a more systemic way and

11:31

make harmony curb self-sustainable sustainable without

11:35

and compromising the open source narrative or

11:39

if this that we are

11:42

committed to so right now it's a

11:44

free tool there is no subscriptions

11:48

and we we might introduce some kind

11:52

of premium features or for b to

11:55

be customers and maybe that would be

12:00

certain

12:02

yeah the to address specific use cases

12:05

that are important for me to be

12:07

customers

12:09

which show would be yeah with which

12:12

could be i think

12:15

available as like additional thing but the

12:18

the core product the court experience of

12:20

harmonica should be free and available to

12:22

everyone

12:24

and so we rely on grandson and

12:26

we rely on donations

12:29

then he yeah it's are important for

12:32

this kind of product i of projects

12:35

maybe is saw in the this word

12:37

fit better and yeah i think to

12:41

be open source because the we are

12:43

talking about the very serious things i

12:46

mean the

12:48

you're talking about sense making and inside

12:52

a nicer way to this group

12:57

what the softer what the software does

13:01

and the yeah it seems quite hard

13:03

the for innovative project to ever some

13:06

funding and the yeah sorry for them

13:08

as understand get misunderstanding about your background

13:12

and the

13:15

and it was thinking about like the

13:17

first question and like what is our

13:21

monica like for someone that has never

13:24

tried the the tool like i've tried

13:26

it and is a is a chat

13:28

no way

13:31

and the user has a conversation with

13:33

an llm about the a certain topic

13:38

right now

13:41

and they're the idea is quick to

13:42

collect the

13:45

all the different positions

13:47

how will you describe it in and

13:50

to someone that the as know the

13:52

right yeah so what's the actual product

13:56

healthworks and

13:58

yes you're at the we decided to

14:02

focus at least for now focus on

14:06

one on one

14:08

dialogues with the participants of each process

14:14

each session or we think that it

14:16

has a lot of benefits people can

14:20

be much more honest or open and

14:25

they can be anonymous and they don't

14:28

have to you know think about what

14:33

what other participants or maybe their team

14:36

members or like people from the organization

14:39

and how they might interpret their words

14:42

and

14:44

they can do that

14:46

at at their own time like he

14:48

is important to note that the harmonica

14:51

is facilitating asynchronous

14:56

discussions or asynchronous workshops asynchronous sir deliberations

15:01

so

15:03

yes of the host creates a session

15:09

and invites the participants and participants joined

15:13

the session and they start talking to

15:15

the

15:17

a i read the language model through

15:20

our web app through the chat interface

15:23

that you mentioned which is very similar

15:25

to charge a pity but unlike tragic

15:28

p t it's a multiplayer experience it's

15:31

multiplayer session which means that even though

15:34

you talk to everyone talks to the

15:37

a i the ai is saving all

15:41

those responses to a vector database

15:46

and it can ask follow-up questions or

15:50

additional questions based on other participants responses

15:55

what we call cross pollination

15:58

so

16:00

basically we once participants to be able

16:03

to

16:05

comment on each other's ideas or build

16:07

upon each other's ideas while talking to

16:11

a i saw like they talked to

16:12

each other but indirectly they talked to

16:15

each other through the ai

16:18

and the the ai synthesizes the statements

16:22

the responses and

16:24

it can identify the clusters of opinions

16:29

in the way like similar to what

16:30

police or how police works i with

16:32

policy you create some since statements and

16:36

participants can react to those statements but

16:37

also they can add their own statements

16:40

and then or the participants can react

16:42

to those additional statements from other from

16:45

from you and so eventually you can

16:48

create this kind of

16:49

cluster raised or like the you can

16:52

map the

16:54

i've been clusters of of your community

16:57

so we wanted to something similar with

16:58

harmonica but with conversational

17:02

in interface though that everyone can just

17:05

to chat with the ai and also

17:07

we are working on an additional modes

17:11

of input like voice so that he

17:13

will he won't have to type you

17:15

can will you will be able to

17:16

talk to her monica and also we

17:18

want your to add like additional like

17:22

elena

17:22

it's like sliders like made me like

17:25

multiple choice and so on so it

17:27

would be much more

17:30

and flexible or dynamic experience for participants

17:34

but most importantly there and responses will

17:40

be saved in this fashion database or

17:43

context window and session context and the

17:49

first of all the

17:50

harmonica will

17:52

identify

17:55

like

17:57

important statements or may be outliers like

18:01

strange ideas that are in it things

18:04

are original and then for instance asked

18:07

participants what they think about those strange

18:11

ideas outliers because what we believe is

18:15

that

18:17

consensus is overrated like consensus of common

18:20

ground is the one you know it

18:24

it can be it is great when

18:27

communities were aligned and like everyone is

18:29

on the same page and you know

18:32

on that common ground but also in

18:36

of i think especially in law

18:37

larger communities like common grounds can be

18:39

quite bland you know like when there's

18:43

a lot of people something that all

18:45

of them agree on can be quite

18:47

small or you know banal like obvious

18:52

you know and then it's it's very

18:53

very important to

18:56

get feedback on less obvious ideas coming

19:01

from some of those members some of

19:03

those groups we want other groups to

19:05

comment on those ideas and then it

19:08

would be hopefully it'll be much easier

19:11

to for instance like develop a strategy

19:13

that will be very inspiring for most

19:16

of the community

19:16

but it wouldn't be boring it wouldn't

19:19

be obvious it would be based on

19:22

some of the outlier ideas coming from

19:24

some of the members which are supported

19:28

by other members

19:30

you know what i'm talking about so

19:31

it's not just finding common ground it's

19:32

not just finding something that everyone agrees

19:34

on it's about amplifying the good ideas

19:40

and then let's i think what sensemaking

19:41

is about

19:43

thought that i will say or do

19:47

a sort of ai mediator mediation and

19:51

also sort of collective of brainstorming weafer

19:55

the llm yep this is very interesting

19:59

i was also thinking about the device

20:01

the input that the could make every

20:04

finger ether

20:07

and the

20:12

about the there is also on an

20:16

idea that the i thought about the

20:19

so collectivist collective utilization of problems like

20:23

someone as a certain kind of problem

20:26

may be in the same town another

20:28

person as this improv

20:30

and the eventually this tool this project

20:34

called the show that many people in

20:38

that place share the same problem and

20:40

so eventually they can also try to

20:42

find a solution to to death

20:46

in them

20:49

you told me about the a little

20:51

bitter about the the idea how you

20:54

had it and but i would like

20:57

to ask you if you were interested

21:01

in the

21:04

in this the topics not maybe nothing

21:07

civic tech but like air in your

21:10

experience i have you had any previous

21:15

to experience about conflicts

21:19

because of you know sometimes it's normally

21:21

we disagree i

21:27

and the some time we understand the

21:29

other person but we not share what

21:31

they say what that person say

21:36

and the i dunno was there for

21:40

you like a a moment when you

21:42

thought okay we need that new tools

21:45

i dunno maybe a personal situation something

21:50

that you thought okay we need the

21:53

new things because they it's it's hard

21:55

to to agree if we use the

21:59

normal and

22:01

if we just communicate in a normal

22:03

way

22:05

while with ai and a

22:09

and they use of llm cooler facilitate

22:14

call sections

22:17

to be honest the conflict resolution wasn't

22:19

something that i thought about when i

22:21

came up with her idea of for

22:23

harmonica i

22:26

like i said it was more about

22:27

my struggle with the

22:31

just getting input from the community like

22:35

it was very you know

22:39

it took me a lot of time

22:41

and also i had to process the

22:44

inputs like for instance you know when

22:47

i did like workshops and miro and

22:51

mira not sure how to planet

22:54

and you know people added some stickers

22:57

to the white board and then after

23:00

the workshop i had to copy all

23:02

those stickers somewhere and like tried to

23:05

cluster them and to synthesize them so

23:08

that it would be like the something

23:10

like actionable make some kind of

23:14

summary of the workshop and the to

23:16

me it was like one of the

23:19

most difficult things with and in terms

23:22

of workshops spiritual workshops so it's a

23:24

little it's also about the time basically

23:26

separate so it's it was a very

23:30

it took a lot of time and

23:32

also it was hot

23:33

to engage people just to get like

23:36

a lot of people to participate and

23:39

finally you know like with for instance

23:41

with them

23:43

like okay cures frederick objectives and key

23:46

results which is very popular framework for

23:48

a setting like objectives for our organization

23:54

like i tried to do that

23:56

with with my team with the other

23:59

teams but first of all i did

24:03

have much experience with it so i

24:05

wasn't entirely sure that i'm doing it

24:06

the right way and secondly other people

24:08

didn't know about ogres until i told

24:10

them about it

24:12

and so for them it was like

24:14

a learning experience and that they they

24:16

didn't quite understand what they need to

24:17

do so

24:20

i think what a harmonica

24:23

one of the benefits of harmonica is

24:25

that it helps you to use the

24:28

right framework and the right way like

24:30

for instance if you want to develop

24:31

a goes like harmonica will know how

24:34

to do it and if you want

24:36

to develop a theory of change for

24:38

the community from one occur will tell

24:40

you or suggest one of the ways

24:42

to to develop a thief

24:43

of change in terms of conflict resolution

24:45

i think that's a very interesting topic

24:50

basically for me conflict resolution is one

24:53

of the key

24:55

and elements of governance

24:58

so a lot of people think there's

25:00

like a very like a popular misconception

25:05

that a lot of people think that

25:07

governess means decision-making and actually that's not

25:10

the case decision making is one part

25:13

of governance that

25:15

it is suddenly decision-making it's also

25:18

no code of conduct for instance like

25:22

certainly principles of how your organization is

25:25

like operating and and that can that

25:29

there's a big overlap between those principles

25:32

and decision-making ideally the decision-making should be

25:36

like synchronized with your values and the

25:38

prince

25:38

apple's

25:41

and there is also a very important

25:43

topic of tools like what does the

25:46

use and who has access to those

25:48

tools like who can actually

25:52

edit or delete you know important documents

25:56

in the cloud that's not so very

25:58

very big part of governance to me

26:00

which some people just don't think about

26:02

the enough some just don't think about

26:05

and then there is conflict resolution which

26:07

is also equally important like how do

26:09

you actually resolve conflicts or disputes

26:12

which is by the way like it's

26:13

one of the design principles of elinor

26:16

ostrom like element western ghats for nobel

26:19

prize basically for developing eight design principles

26:22

for building successful comments like how to

26:26

how to create a successful comments like

26:29

a if you wants to build the

26:30

occur open source project which what

26:32

or eighty wants to maintain a forest

26:36

she was obviously a and forests and

26:40

the like fish in the water but

26:42

the same principles apply to open source

26:44

projects and the other digital commons so

26:47

conflict resolution is one of those eight

26:49

principles like if you want to if

26:52

you

26:52

want your community to maintain or develop

26:55

shared resources

26:58

you need to have conflict resolution in

26:59

a place

27:02

and either i i dunno if

27:09

like how exactly a i could be

27:11

used for that i'm sure there's probably

27:14

like a few different startups that are

27:16

trying to develop to to build a

27:18

tool for conflict resolution with a i

27:21

right now

27:22

the harmonica is

27:25

so i think for conflict resolution facilitation

27:28

is also very important basically facilitation is

27:30

something that is important for sense making

27:35

and for conflict resolution

27:38

and how many trying to understand how

27:40

to

27:43

turn facilitation wisdom until code like how

27:48

to codify facilitation wisdom so that a

27:50

i could facilitate workshops in a specific

27:53

her away like for instance like a

27:55

retrospective road like harmonica can run a

27:59

retrospective for your team or it will

28:02

talk

28:02

to use or participants to the members

28:05

and it will ask them about you

28:07

know like the standard the retrospective questions

28:10

like what was good about the last

28:12

sprint what didn't work so well what

28:15

can be improved and how could we

28:18

actually improve it or how can we

28:20

fix the problems right of there's like

28:22

three

28:22

three main questions in a retrospective and

28:25

it's kind of obvious like yeah like

28:27

he can write a prompt for like

28:29

djibouti and explain all that for the

28:33

language model and then it will this

28:35

prompt will be used to

28:38

like you know talk to each participant

28:44

collect the inputs from them

28:47

and then another prompt which is also

28:49

part of the retrospective session will be

28:51

used to synthesize all those responses and

28:54

generate the summary

28:58

of your respective and that summary would

29:01

be

29:03

pretty good like it will be very

29:05

close to water in a liquor retrospective

29:08

outputs should look like it according to

29:11

like a jail

29:13

the thinking like you know experts in

29:18

agile and management who are we really

29:21

wanted to collaborate with and like we

29:23

want them to give us feedback like

29:24

if the prompts that were using is

29:26

not very not very good then we

29:28

are happy to

29:30

change them you know improve them or

29:33

maybe those experts wants to create their

29:35

own prompts and their own then place

29:38

as we call them and the it

29:39

can be not only about agile it

29:42

can be about anything and so may

29:44

be conflict resolution could also be

29:48

you know what part of that experience

29:52

maybe the the actually can be like

29:54

a lot of different templates a lot

29:55

of different frameworks that help people took

29:58

resolve conflicts we just haven't thought about

30:01

that yet

30:03

but maybe it's actually in important like

30:07

direction for tools like harmonica like how

30:11

how you could do

30:15

yet how the a i should talk

30:17

to people but the what when bit

30:21

skeptical about this is i think conflict

30:23

resolution is an emotional like with when

30:26

you trying to develop like oak yards

30:28

or a roadmap it's basically it is

30:31

just about ideas it can be quite

30:35

rational it doesn't have to be emotional

30:37

with conflicts i think there's much more

30:39

emotions in place and i think it's

30:42

very important not to lose those emotions

30:44

in the process like you need to

30:47

actually be ideal you want to be

30:49

in the one in one room with

30:50

the someone to the resolve the conflict

30:53

like if you just talk about that

30:55

that stuff through zoom or some other

31:00

and you know

31:03

video conferences at all it's already like

31:08

harder to resolve the conflict and if

31:10

you do that through the chats with

31:13

a i

31:14

i'm afraid that the just to you

31:16

know it you just lose all the

31:19

emotions that you need to actually hear

31:22

the other person like not just saw

31:25

with a

31:26

not just the war is would also

31:28

like the by the language and stuff

31:29

like that so

31:32

i dunno having the conflict resolution is

31:33

very big topic that can you give

31:34

you can think about like projects like

31:36

clara's phrase that claris is doing kind

31:41

of decentralized to

31:44

conflict resolution but that's more like a

31:47

court right likely just turf people who

31:49

who

31:51

i tried to resolve conflicts and as

31:54

a service you're using blockchain but i

31:56

don't think they're using ai

31:59

and

32:00

i can correct now

32:02

then and you are made me think

32:05

that the maybe facilitation is the step

32:08

before conflict resolution

32:12

i mean because i just despise listen

32:14

to the mental skill that you need

32:16

for affix resolution as well as sensemaking

32:19

yeah yeah exactly if we ever way

32:22

to understand each other before we have

32:25

a conflict then hopefully we done after

32:28

i have a conflict and so that

32:30

is why of incur

32:31

the to like harmonica could be very

32:34

useful

32:36

and the elect ask you if you

32:38

encountered any kind of problems i dunno

32:40

maybe you are thinking about something okay

32:42

we have to do it in this

32:44

way or in another way and then

32:47

your face some something unexpected i dunno

32:56

well room

32:58

you know

33:00

every startup the has like dozens of

33:05

things like that

33:07

until they get to their black product

33:09

market fit and we we haven't got

33:12

there yet

33:13

so i guess we have something like

33:16

that like every week

33:18

but also like when he asked me

33:19

about this right now i just i

33:21

need some time to like remember something

33:24

that would be interesting to talk about

33:26

and like you know in general this

33:29

is all very fuzzy rent like for

33:31

instance we we started with a different

33:35

but like the first the for the

33:38

type that were built was a bot

33:39

for discord

33:42

and then we realized that and

33:46

it's not that easy to

33:49

like facilitate a workshop in there

33:53

dems on discord through through the ams

33:56

on discord

33:58

and also like a lot of people

34:01

that i talked to about harmonica they

34:03

were not underscore like they had to

34:05

install discord and register just to be

34:08

able to test the harmonica

34:11

so then we decided to move to

34:14

telegram and like last summer i was

34:17

invited by

34:19

as ozil community and and of he

34:21

knows as a little was as allow

34:24

it sur le carr conferences as sponsored

34:28

by ethereum community foundation

34:32

and so the first one was in

34:34

montenegro in twenty three

34:36

and the last year it was in

34:39

georgia

34:41

georgia the country not the state

34:44

and

34:47

so they invited me to help them

34:50

organize the one of the weeks because

34:53

it's a long contracts they call themselves

34:55

pop-up village so i think it took

34:57

like six or seven weeks and total

34:59

and one of those weeks was about

35:01

decentralized governance

35:04

and so i helped them organized that

35:06

weekend also i thought that it it

35:08

was it would be like a good

35:10

opportunity for us to test harmonica

35:13

like a pilot project for harmonica and

35:15

so we built a telegram telegram bot

35:19

ah for that week so we tried

35:22

to use from monica at the conference

35:25

and zoo village georgia

35:28

and invited the participants of the conference

35:30

of that week people who attended the

35:34

the sessions that i organized

35:37

like with stickers and

35:40

and some workshops we invited them to

35:43

use the the telegram bot harmonica on

35:48

telegram

35:50

which also wasn't very successful like

35:54

a lot of people joined those sessions

35:55

but never finish them they didn't respond

35:58

to the questions

36:01

and the i realized that yeah we'll

36:04

probably just it wasn't the right solution

36:06

for that context like people who attend

36:08

the conference they're not members of the

36:10

same team some most of them don't

36:13

know each other so they didn't have

36:14

any like

36:16

skin in the game they didn't want

36:18

to discuss anything with each other they

36:20

didn't want to help us and harmonica

36:24

so

36:28

with all that sir feedback i in

36:30

the nothing that but feel that but

36:32

also like just the kind of

36:37

you know basically that wasn't the very

36:39

successful experiment for us so we decided

36:42

to start using to to build a

36:44

web app like to just to create

36:46

something like judge a pity on on

36:49

the web and i think that this

36:53

currently working pretty well the web app

36:58

but it also has its limitations and

37:02

i think that people

37:04

and a lot of people just don't

37:06

want to open links like if someone

37:08

sends them a link especially if it's

37:10

not the direct messages but as if

37:12

it's like a channel on discord and

37:14

like everyone on the community is invited

37:17

to join the session very very few

37:20

people actually click on the link and

37:21

that's what we've seen recently for instance

37:25

in orinda community like they tried to

37:27

do at a year retrospective with harmonica

37:30

in les les december or january to

37:34

reflect on twenty twenty four as a

37:37

community and i think it was like

37:40

ten people for joined the session

37:45

with the yeah i think like a

37:47

few hundred members and on the discord

37:49

server like maybe a five hundred six

37:51

hundred members in orinda and only ten

37:54

of those users actually joined the session

37:57

in harmonica

37:59

which you know it's it's a problem

38:03

for us and we haven't the

38:07

figured out how to solve this yet

38:09

but one of the hypothesis that we

38:11

have is that we still we we

38:12

need to go back to building a

38:15

discord bot so that we made you

38:17

will make like the full-circle and but

38:20

it will be a new kind of

38:21

bot that would be using a

38:25

that would be connected to an agent

38:28

so it will be basically like harmonica

38:31

agent that would be joining

38:36

certain channels or threads are in a

38:40

committed to server and maybe it will

38:43

be and

38:46

trying to facilitate

38:48

discussions happening there like so that people

38:50

wouldn't need to leave discord they wouldn't

38:52

need to join any sessions it will

38:55

be it won't have any set it

38:56

will be like specialists sense experience just

39:01

happening on the community platform that the

39:04

community the so they both inside the

39:06

the group that is core group

39:08

telegram group or whatever instead of further

39:11

use that as to fried to the

39:13

bot privately

39:15

the oil joining a session on the

39:18

web which is to the current solution

39:21

exactly and the i think about these

39:24

the problem i mean someone build up

39:27

a project our platform that can be

39:29

very useful and then people are may

39:32

be they are yet as you said

39:34

afraid of opening a lie

39:35

the or maybe they are just lazy

39:37

i dunno they ever things through their

39:40

their their their life and the may

39:44

be gimme vacation systems called the help

39:47

in a way to

39:51

yet to stimulate use earth to to

39:54

do a certain thing to reply but

39:57

also as you say like they're out

39:59

you could be

40:03

ether so you just the the is

40:05

just say something to there

40:09

cell phone a year he absolutely i

40:11

think people can

40:14

sheer like ten times more information if

40:16

they can talk

40:18

instead of typing

40:20

and we basically we want to increase

40:23

the bandwidth we want to for sensemaking

40:26

we need to get as much information

40:27

as possible

40:29

from the community or from the organization

40:32

so we really need to on the

40:35

one hand be able to use different

40:38

modes of like except different kinds of

40:41

input including voice and on the other

40:44

hand we need to be able to

40:46

join the conversation

40:51

that are happening already and you know

40:54

that could be a boss that is

40:56

deployed to and discord server or maybe

40:59

it's just the something like as call

41:03

maybe we should be able to join

41:04

some calls so we could build the

41:08

zoom app rights that

41:09

would be invited to those calls or

41:12

we could use the transcripts that are

41:14

already generated by tools like autor and

41:17

fireflies you know those a i know

41:20

taking tools so that's something that we

41:23

are working on right now we want

41:24

to be able to add those transcripts

41:27

as the context to the sessions so

41:32

it's not just what people

41:35

i say or what people tell harmonica

41:39

in our sessions but also being able

41:42

to import conversations that happened elsewhere

41:47

may be on zoom may be in

41:50

discord we want to be able to

41:52

kind of

41:55

combine or merge all those different to

41:59

conversations to form this kind of a

42:03

you know

42:03

collective intelligence or

42:06

like a context for a i and

42:10

then the next question and then on

42:12

the question but the something that we

42:14

are working on day i want to

42:16

share with you is that if we

42:18

managed to like form that collective intelligence

42:21

or she context sir based on all

42:24

those conversations that can happen in

42:26

harmonica or outside of harmonica then we

42:29

also want to be able to turn

42:31

that into

42:34

and agents itself

42:38

and that agents would be able to

42:39

represent the opinions or the perspectives of

42:45

that group of people

42:49

ah somewhere else

42:51

so this is what the we call

42:53

a i avatars

42:56

and i think that could be like

42:58

a

42:59

really fundamental breakthrough for

43:02

like online governance when groups of people

43:06

will be able to deliberate

43:10

on certain topics and then create an

43:15

ai agent that would er

43:18

represent them based on all those deliberation

43:23

the based on all those conversations with

43:25

and all those responses that people shared

43:29

and you know what they agreed on

43:32

and disagreed on so i think it

43:34

will be very powerful away to scale

43:37

will do

43:39

deliberative democracy because you know it's very

43:43

hard to deliberate and stuff all the

43:45

time like people have other things to

43:47

do they don't want to join like

43:49

citizen assembly as every week

43:52

but if we managed to create something

43:55

like and governance surface for the team

43:58

or for the community

44:02

and that would enable them to create

44:05

an agent that would accurately represent their

44:08

perspectives and values and like girl intentions

44:13

like what are they really once maybe

44:14

within the company like maybe it team

44:17

wants to create an agents that would

44:20

represent it

44:22

on like higher level company

44:27

the strategy meeting rights of that the

44:31

interests of that team would be represented

44:33

the when the department is making decisions

44:36

and then maybe the agent of the

44:39

department will be represented at the company

44:41

level the this is making the same

44:43

applies to politics like maybe a neighborhood

44:47

can

44:47

discuss lot of different stuff like deliberate

44:50

for like maybe a few months and

44:52

then that context would be enough for

44:55

this neighborhood to create an agents that

44:57

would represent this neighborhood on city level

45:01

deliberation and then maybe the city consisting

45:04

of all those different neighborhoods would be

45:05

able to create an agent

45:07

with represents the city on national national

45:11

level

45:13

deliberation so this is i think very

45:15

very important because it allows you to

45:17

scale a deliberative democracy

45:21

absolutely it's very charming them

45:26

the the way the the a agents

45:29

can represent people

45:32

and the that's also a water was

45:34

talking about the when i said about

45:36

the collective is collectivization of problems so

45:40

i went to a laker some people

45:42

that ever the same problem can be

45:45

represented by a certain the agent to

45:48

them

45:49

i dunno the state level town level

45:52

whatever here

45:54

and the i'm thinking i mean now

45:59

the plus from now is the i

46:02

mean is developed and a you're talking

46:04

about her a lot of for improvement

46:07

that are very cool

46:09

and the

46:13

which kind offer

46:15

ah i dunno like up

46:19

problems you think you will face or

46:23

skill you will require that you do

46:25

not have now i mean a mythical

46:27

of the said is a big community

46:30

and luckily there are also other communities

46:33

and the people are collaborating saw

46:37

like it is there any out that

46:39

you need the

46:41

the

46:43

had another

46:45

okay thank you for asking

46:48

yes we need it

46:50

a lot of different kinds of help

46:53

that's for sure we are a small

46:54

team there is only five numbers in

47:00

tim harmonica including myself we have been

47:03

bootstrapping like we don't have external funding

47:07

apart from the get coin grandson the

47:10

giveth the nation's

47:12

so we definitely need the you know

47:16

we need to fundraise basically somehow for

47:19

the to make it more the able

47:21

to work on this full i think

47:23

i'm the only to remember that works

47:25

on harmonica full-time are currently and

47:29

but that's kind of obvious right like

47:31

fung funding also we definitely need like

47:35

developers who would be aligned with what

47:39

we're doing who really like wants to

47:42

fix the democracy and make the planet

47:47

more happy place

47:50

and

47:51

so developers is another thing but i

47:54

guess what we really needs right now

47:58

is the facilitation wisdom because like i

48:03

said i think facilitation is the core

48:07

value that we are trying to deliver

48:11

with a

48:11

a i

48:13

though it's harmonica is basically a i

48:16

facilitation tool

48:19

in a way like a high-powered since

48:21

making tool and and the facilitation as

48:24

be part of that so i wanted

48:26

to use this opportunity to share some

48:30

and updates from the

48:34

an interoperable deliberative tools program that's i

48:39

participated in with the harmonica i think

48:43

you are also part of that the

48:44

program and so

48:48

a few weeks ago

48:49

we decided to collaborate with some of

48:52

other members of this a program including

48:57

a i object as institute and the

49:00

eyes

49:01

and so i hope that other projects

49:04

will also join us and may be

49:06

built benefit from what we want to

49:07

do it and what they want to

49:09

do is open source library of facilitation

49:14

patterns so facilitation patterns facilitation tactics

49:19

basically we want to turn the difference

49:23

facilitation traditions like

49:28

socio chrissy or nonviolent communication or maybe

49:33

canadian or you know there is a

49:35

lot of different stuff that is very

49:37

very interesting that has been developed over

49:40

like decades liberating structures in and is

49:43

another great example and those are like

49:46

different traditions the a focus on different

49:49

things like obviously like agile is more

49:52

about teamwork and like companies that companies

49:55

in particular and mobile communications just about

49:58

how you can facilitate discussion in the

50:01

room and how do you can how

50:03

can you make people

50:05

you know open and the criticized safe

50:08

space for more open discussions so those

50:12

are they like different things different aspects

50:15

of facilitation maybe like difference even like

50:18

levels of facilitation but we want to

50:21

try to turn all that into code

50:23

so that's tools like her mom

50:25

annika what dr the city or eyes

50:28

or maybe some other deliberative tools which

50:33

i think her yeah hopefully will be

50:37

open source as well

50:40

would be able to benefit from this

50:43

open source library which we call open

50:45

facilitation library so are we really want

50:49

to engage as many facilitators as possible

50:52

to share their wisdom with us

50:55

you know it for salvation experts from

50:58

those all those different

51:00

the facilitation traditions we really want to

51:04

collaborate with them to create this kind

51:06

of shared facilitation

51:10

library

51:14

this is critical in mean that a

51:16

library means that the i will help

51:20

her yeah new developers new people that

51:24

have some new ideas to develop their

51:27

project yeah

51:30

i say that

51:32

you know there are some examples of

51:33

libraries like this for example there is

51:35

a session lab a library and the

51:40

result the wise democracy project that also

51:43

tried to describe this kind of butter

51:45

language of democracy and there's like a

51:48

big overlap with what what would we

51:50

want to deal with what we also

51:51

want to do

51:52

not just the describe those facilitation or

51:55

tactics or patterns but also tried to

51:58

turn them into a machine readable form

52:00

so maybe like typescript

52:03

so that we could basically teach a

52:08

i how to use those things how

52:11

to use those patterns

52:13

the you know the it will could

52:15

just like copy those things to the

52:17

prompts or turn them into a gigantic

52:21

wore clothes and stuff like that so

52:23

it's also about the cold in a

52:26

big way sauce a sort of machine

52:29

learning on there

52:31

on all the possible ways there yeah

52:34

humans used to mediate yeah and the

52:36

so it seems like that the what

52:39

are needed now are people that are

52:41

not from the tech side from the

52:43

tech field but they are yes studying

52:47

the ways that people use to ah

52:51

yet resolve conflicts or or to facilitate

52:54

the conversations was citizen assemblies yet

52:59

yeah i felt like the same in

53:01

the i mean there are a lot

53:04

of people's that the have some i

53:09

mean technical knowledge and so they are

53:12

developing new tools like a could be

53:16

or monica could be other tools and

53:19

and the and a lot of people

53:20

that have a lot of knowledge like

53:22

a people that may be studied the

53:26

facilitation techniques and the still are not

53:29

in this field

53:30

but i really hope that in their

53:32

next future they will come and teach

53:34

household their the knowledge that they have

53:38

and them

53:40

about term

53:44

yeah questions about other tools that the

53:47

maybe you have seen around that they

53:50

looked very interesting to you and the

53:54

that may be share something with harmony

53:56

color may be there are very different

53:58

to but they were you were inspired

54:01

by them

54:02

i am as you cited bodies before

54:07

yeah

54:10

yeah there's definitely a lot of tools

54:12

that the

54:15

ah

54:18

yeah

54:22

to trying to do similar things

54:26

as her monica

54:28

and there is also a lot of

54:30

tools that we are using

54:34

the to build her harmonica but yeah

54:37

i guess you're asking me about the

54:40

the former

54:43

so yes obviously policy is a big

54:45

inspiration and

54:48

unfortunately

54:50

it hasn't been

54:53

updated for a long time i think

54:54

it's just like like a public good

54:58

now and nobody is extracts to develop

55:00

it and like improve the wax so

55:04

it's pretty like old call us and

55:08

you could say that we are trying

55:09

to

55:13

yeah like do a similar thing but

55:15

with a different tx so that is

55:16

more modern and conversational and a i

55:20

powered but harmonic is not the only

55:23

tool that is trying to do that

55:25

so i mentioned the talk to the

55:26

city i think the city is a

55:28

great project

55:31

also part of the sir interoperable collaborative

55:33

tools program

55:35

and

55:37

there is a rematch which is a

55:39

very mature project

55:43

which is also using a i too

55:47

run collective dialogues so he could sculptor

55:50

collect dialogue system

55:53

and by the way rematch is not

55:57

open source and it's quite expensive it

55:59

costs i think are like thirty five

56:02

thousand a year

56:05

thirty five thousand dollars a year i

56:07

think

56:08

yeah it's a lot of money i

56:12

dunno at lot yeah i guess they

56:15

maybe they i i know that they're

56:16

working with the un and maybe u

56:18

n is like a pro bono customer

56:21

for them

56:23

but the definitely work with like large

56:26

companies tool can spend that much money

56:29

on their collective dialogue system which i

56:33

find interesting there is also another american

56:38

projects are called think scape built by

56:42

a company called

56:43

the unanimous a i

56:45

and i think it costs a about

56:47

the same and i made me be

56:49

mistaken but i think it costs like

56:51

three thousand a month

56:53

and then escape is also using a

56:56

i agents to organize a deliberative process

57:01

is or

57:03

some kind of yeah

57:07

online discussions and they have like a

57:09

very interesting design so as the they

57:12

want the create small rooms for people

57:16

and people discuss different topics in those

57:18

rooms and then there are agents who

57:19

listened to those discussions and the and

57:22

then they have other agents who represents

57:26

are the groups in your group so

57:28

that you could the

57:31

respond to

57:33

ideas from other groups that are collected

57:37

by one agent and like reproduced in

57:41

your group by another agent something like

57:43

that they might be

57:45

you know i help him that sir

57:48

wrong but maybe it's not very accurate

57:50

description but i hope for that's more

57:52

or less how they work

57:54

so that escape is interesting and

57:57

yeah i dunno like you know there's

58:00

is so we're building a collective the

58:02

alex system of i've been talking will

58:06

have been met i mentioned if you

58:07

collect of alex systems but there is

58:09

also very interesting start happening like and

58:12

but bridging systems like how to identify

58:15

different

58:17

opinions and the

58:21

like with what they do with committed

58:23

to notes and i think committed to

58:25

notes is just an early example of

58:28

bridging systems and they will be much

58:29

more interesting bridging systems in the future

58:32

and so i think in a i

58:35

think it's interesting to do

58:38

like discuss not right now i am

58:41

afraid to jump offers soon but i

58:44

think it's interesting to discuss a sense

58:46

making stack like what since making stack

58:49

the consists of

58:52

and i think the hurdles like harmonica

58:54

is just one part of that stack

58:56

so what what else do we need

58:59

to like enable

59:04

a good the collective sense ah

59:07

at the me the is like

59:11

or of open question that i don't

59:13

know the answer for right now but

59:15

like i think that the yeah sensemaking

59:21

he is definitely something that we need

59:24

to work on their civilization to address

59:27

the matter crisis

59:29

that's not my thoughts that sir i

59:31

think something that i heard from daniel

59:33

martin burger which eyes one of my

59:37

but which had been like a big

59:39

influence on myself and the last few

59:41

years

59:43

and yeah i think it's important to

59:46

think about the collective sense making and

59:48

how a i can

59:51

enable it i think that there's some

59:53

irony that on the one hand a

59:56

i is dangerous technology like it could

59:58

potentially kill as kill everyone but also

1:00:03

basically we can solve the political problems

1:00:06

that we have without a i

1:00:08

so i think that's a very interesting

1:00:10

moment in history because i'm the one

1:00:13

hand we are building the tools that

1:00:14

can kill as but on the other

1:00:16

hand there is this growing political problems

1:00:22

that can be sold only with the

1:00:23

help of ai

1:00:26

and basically we have this window of

1:00:27

opportunity that we we need a i

1:00:29

but also it's dangerous and we need

1:00:32

to build a i that would help

1:00:34

us do

1:00:37

like fixed the democracy our otherwise it

1:00:40

will be very dystopic yeah it seems

1:00:43

like a sort of paradox or

1:00:47

but yeah i completely agree that the

1:00:49

a i n technology can be used

1:00:52

the

1:00:54

yang the song good ways to move

1:00:57

the to make the war the a

1:00:59

better place

1:01:00

him

1:01:02

another small question i mean is there

1:01:06

i mean there are all the snoopy

1:01:07

all these people i mean you're part

1:01:09

of her and i would say i

1:01:12

am also part of

1:01:14

do that are trying to find new

1:01:17

ways for people to collaborate to to

1:01:20

discuss and the do you think that

1:01:23

the all these people are collaborating in

1:01:26

a good way right now i mean

1:01:28

that that the intent of and i

1:01:31

mean

1:01:33

it is to facilitate dialogue and the

1:01:37

and for for other people like for

1:01:40

dollars for companies from i dunno institutions

1:01:44

but at the same time i think

1:01:46

it's important for people like and their

1:01:48

civic decker space to collaborate and the

1:01:53

do you think that this is happening

1:01:55

right now do you think that something

1:01:57

should improved do you have a message

1:01:59

for older people in into the civic

1:02:02

tech while you know it's a it's

1:02:07

a good question because i think there

1:02:09

is a systemic problem here are which

1:02:13

is funding

1:02:15

this stuff is not very profitable

1:02:20

and if it is built for profit

1:02:23

it becomes much harder to use it

1:02:27

at scale

1:02:29

i mean

1:02:31

basically you want this stuff to be

1:02:33

open source

1:02:35

i like palace

1:02:37

but if it's open source then it's

1:02:39

very hard to

1:02:42

of raise funding

1:02:44

it's very hard to hire like engineers

1:02:48

because engineers are very expensive

1:02:51

so all of this project they get

1:02:54

into this

1:02:57

conundrum or

1:03:00

you know like a difficult the situation

1:03:03

where are you on the one hand

1:03:05

it's

1:03:07

yeah like you want to build something

1:03:08

that would be very like he useful

1:03:11

for the humidity but also like is

1:03:16

not easy to fund it so it's

1:03:18

very hard to the build it

1:03:23

even if you have some kind of

1:03:24

a fear of change and like impact

1:03:27

model and you know they just like

1:03:30

get coin that is trying to solve

1:03:31

this problem with grunts radek with quadratic

1:03:33

funding and grunts and i guess get

1:03:36

coin is like a great example and

1:03:38

garlic there should be more

1:03:41

projects like bitcoin and thankfully get coin

1:03:43

is also developing a protocol that should

1:03:45

enable other communities and other organizations to

1:03:48

run similar like grand rounds using their

1:03:51

protocol of allow

1:03:54

and so let him get quite as

1:03:55

doing like an amazing job and i

1:04:00

want more people to know about get

1:04:01

coin

1:04:04

but also i think it's just the

1:04:07

way

1:04:09

the capital is allocated in in the

1:04:11

economy and maybe we need to just

1:04:16

i think of a new institutions that

1:04:18

would be focused on

1:04:22

like

1:04:24

funding this kind of projects because

1:04:28

yeah it's it's hot it's hard to

1:04:30

fund them with venture capital because venture

1:04:34

capitalists wants to like the next a

1:04:37

growth rate and the

1:04:40

exits and it's also not they're easy

1:04:43

to funded with grunts like even though

1:04:45

get on is doing a good job

1:04:47

i think it still like you can

1:04:49

only get when so many grunts a

1:04:52

year and those grants are not very

1:04:54

big so you can't hire a lot

1:04:57

of for experienced developers sir if you

1:04:59

just the

1:05:00

if your funding is coming from get

1:05:02

coin

1:05:04

so one of the things that i

1:05:06

was really inspired by the or excited

1:05:09

about his

1:05:11

the idea of a guilds as

1:05:17

basically like communities or networks that can

1:05:21

fundraise

1:05:23

and then allocate those funds among their

1:05:26

members

1:05:29

based on much deeper relationships with those

1:05:33

projects than the get coin the can

1:05:36

have rights of because get coin doesn't

1:05:38

care about your project get coin is

1:05:40

using like this kind of mathematical models

1:05:42

to identify projects that get the letter

1:05:45

like support from their communities like if

1:05:47

you can get a lot of people

1:05:49

donate then you get some funding from

1:05:51

gift coin

1:05:52

with guilds i think it's a much

1:05:54

more it could be much more basically

1:05:59

it could allow it much more substantial

1:06:05

funding because if a guild can find

1:06:11

like an institution that is aligned with

1:06:14

the mission of the guild it could

1:06:16

get a lot of money and then

1:06:18

it would distribute that money among members

1:06:21

of the guild like projects that

1:06:25

you know part of that guilds

1:06:29

network or portfolio so i guess it

1:06:33

it this kind of similar to venture

1:06:34

studio but the venture studio is it's

1:06:39

basically accompany you know it's a company

1:06:41

that is part of this kind of

1:06:42

capital capitalistic her

1:06:46

paradigm and i think the adventure split

1:06:48

is are amazing i think it's one

1:06:49

of the most interesting things happening

1:06:52

in terms of like startups and the

1:06:54

of venture capital so like you know

1:06:57

companies that build startups read venture builders

1:07:01

that's a super interesting phenomenon

1:07:05

you know there the community called art

1:07:07

thou that is trying to do something

1:07:10

similar but to leave the like web

1:07:12

three projects and it's like come into

1:07:14

dream and like they have discord server

1:07:16

and like everyone can join the server

1:07:18

and started contributing to those projects and

1:07:21

then hopefully our in there will be

1:07:24

able to raise for

1:07:25

funding itself to distribute it among those

1:07:28

projects but still i think it's it's

1:07:30

not a guilt and maybe met the

1:07:33

gulf could be

1:07:36

the that guild may be met with

1:07:38

medical is doing right now he could

1:07:40

be something similar to what's a guild

1:07:43

is supposed to do and i hope

1:07:45

that the with all the reputation of

1:07:47

the goodwill that medical has built over

1:07:49

the years it will it will be

1:07:50

much easier for medical to fundraise than

1:07:53

for her harmonica or

1:07:56

or the projects that are part of

1:07:58

the thicket of met the gulf eco-system

1:08:01

so i hope that

1:08:04

do you know we can yeah we

1:08:07

can design and implement this kind of

1:08:10

new approaches to for funding or which

1:08:13

wouldn't be capitalistic but also would enable

1:08:17

to hire engineers

1:08:23

yeah absolutely seems like a i mean

1:08:25

as as we said the before like

1:08:27

funding is a problem and the but

1:08:31

the really hope that in the next

1:08:32

future because of the desert ai yi

1:08:35

hyper

1:08:38

that there would be summer

1:08:41

new kind of findings and it's an

1:08:43

interesting to him

1:08:45

the thinks about gills and the is

1:08:48

there anything else that you'd like to

1:08:50

add the

1:08:53

that i have tasked okay no i

1:08:55

think her we have covered a lot

1:08:57

of important topics and the yeah i'm

1:09:03

grateful for the invitation this is the

1:09:06

first time that i to talk to

1:09:08

about for monica with the someone like

1:09:11

this like an hour interview format so

1:09:14

it was great for me just do

1:09:17

you have this kind of experience sense

1:09:21

ah yeah i hope that it wasn't

1:09:23

to that i wasn't talking to slowly

1:09:26

because that's one of my problems i

1:09:27

talk slowly

1:09:30

nah it's i i took in are

1:09:32

horrible way so i it was perfect

1:09:35

and it was the was very interesting

1:09:37

for me so thank you a lot

1:09:41

thank you for the opportunity to

1:09:46

what about the

1:09:51

so i