Nikolay: Hello, hello.

This is PostgresFM.

I don't remember episode number.

Michael: 74.

Nikolay: Thank you, Michael.

This is Michael, by the way, and
my name is Nikolay, as usual.

Only once we had a guest, right?

Maybe it's time to invite more,
but it's a different story.

So, this is PostgresFM and we do
it since summer 2022, not skipping

any weeks and I'm still surprised
that we didn't skip any weeks.

It's good.

And today this topic was again
suggested from outside.

We didn't invent it, right?

Michael: Yeah, I chose or insisted
on this topic this week.

We're going to go along with the
monthly blogging event, PGSQL

Phriday.

Pavlo from Cybertec has suggested
the topic events or Postgres

events more specifically.

Nikolay: Not event queue, right?

No, not events.

Michael: No, in-person conferences,
online events.

This quite broad topic and we could
go in any direction you want

with this.

I've got a few things I'd love
to talk about and ask you questions

about.
You've hosted meetups in the past.

We've mentioned it briefly on the
podcast before.

So yeah, I'd love to get your perspective
from an organization

point of view as well.

But yeah, speaking, we've both
spoken at in-person and online

events, both attended and listened
to people's talks.

So yeah.

Nikolay: Yeah, this is an interesting
area, which I have a very opinionated

point of view on this topic and
I think for me it started back

to 2006-2007, actually earlier.

So offline events are fun, They
are good, but I don't look at

them the same anymore.

And this is funny, right now we
have an AWS huge event happening,

many Postgres folks are there,

Michael: right?
Re:Invent.

Nikolay: Re:Invent, yeah, they just
announced Limitless Aurora

with Postgres support first and
MySQL coming later.

Yeah.

And people are already joking about
limitless bills, right?

See?

I haven't

Michael: seen that.

That's great.

Nikolay: Yep.

See, I have all news not coming
there, right?

For knowledge, we don't need to
go anywhere anymore because the

Internet is delivering knowledge
better.

I observe also hybrid approach
when we go to offline event, but

you consume knowledge mostly through
computer being at offline

event.

Offline event is good to meet people,
right?

But I think, I still think there
should be some new type of event

invented to support human meetings,
you know, just for like collaborations,

so not in the form of sharing knowledge
like 1 guy on stage with

slide deck, because I think, like
it's my opinion, but let's

start from the past and maybe by
the end of this podcast, some

people will understand me better.

I know not everyone will agree
with me and I also enjoy meeting

with people in person.

But for knowledge it's not efficient.

This is my position anymore.

And meetups are kind of dead almost,
unfortunately.

There are some meetups, well, golden
era of meetups is behind

us, unfortunately.

When, you know, like 200 people
came and company hosting says,

we don't have coffee break supplies
anymore, let's close registration.

I say, forget about coffee, let's
bring more people.

They're coming, coming because
of Postgres's popularity.

It happened many years ago.

So, where to start?

Michael: Well, you mentioned '06,
'07 as a specific point for you.

What was happening then?

Nikolay: It was actually the first
event, the first serious event for

me was when I was speaking.

It was actually, it was about databases
but not about PostgreSQL.

And I was presenting some, I don't
remember, maybe my master

thesis materials in Kyiv, Ukraine,
actually.

It was some scientific conference,
actually.

It's kind of, not scientific,
academic conference.

Yeah, it was kind of fun to
come there to speak.

It was very long ago, 2005, maybe
'04 even.

So, yeah.

Then 2007 was a turning point for
Russia and that part of the world

when somehow a few things started
to happen.

First of all, I chose Postgres
in 2006, and then I participated

a little bit in XML development
and visited PGCon in 2007.

And in the same year, 2007, we
started, not we, one guy actually

started HighLoad conference in
Russia.

His name is Oleg.

Oleg Bunin And another Oleg, Bartonov,
I also met him around that

time because Oleg Bartonov is,
you know, like Teodor Sigaev and

then Alexander Korotkov, these
are big Russian names.

Maybe after Vadim Mikheev, who
was before them, working on WAL.

These guys worked on full-text
search, GIN indexes, GiST indexes,

A lot of such stuff related to
performance and how we retrieve

data faster.

And so it was happening very fast
in 2007.

I also had my first startup and
I used Postgres a lot.

And then this conference series
started, High Load.

And since I just visited PGCon,
I immediately invited Bruce Momjian.

Michael: For anybody that doesn't
know, PGCon is or was hosted

in Canada yearly and is very Postgres
developer-centric, I believe.

So for people working on the Postgres
internals itself, it's

probably the best conference each
year for them, would you say?

Nikolay: It's not kind of best,
but it was the first in 2006,

so I visited the second one in 2007
in Ottawa, and this year I

visited the last one, so no more
PGCon, unfortunately.

Michael: But isn't it being taken
over by new people?

Nikolay: Yeah.

Dan, who was the organizer of PGCon,
said that the successor

of it will be in Vancouver, I guess,
next year.

And a different set of people will
be organizing, and the name

will be changed to PGConf something,
I don't remember.

But anyway, my decision right now,
I'm on the stage, like we're

jumping between past and current
and trade side.

For me, I said no more offline
events for a couple of years.

I don't see the point.

Last three events were terrible in
terms of efficiency for me.

And well, I'm not complaining.

I just, this is my position.

I prefer sharing knowledge online
right now and consuming and

learning online.

I still miss meeting some people,
but I can organize it separately.

So this is my vision.

So back to 2007, we started again,
not we, I just was like a

member of committee from the beginning,
responsible for the database

topic, and I invited Bruce Momjian
right away because I just met

him in Ottawa.

So he visited Moscow, and since
then he visited many, many times,

High Load.

And at the same time, one guy and
I, we were working on our startup

together.

We started a meetup in also 2007,
2008.

In Moscow, we were a Russian user
group.

I later renamed it from Postgres Russia
to RuPostgres when I realized

that half of people speaking Russian
are outside Russia.

So it became RuPostgres.

And this story continued for me
with some ups and downs till

2022, February 24, when the war
started, like the war started

earlier in 2014, but a new phase
invasion started, and we had

a disagreement internally because
my position always was you cannot

split completely politics from
regular work and we need to be

careful.

The main organizer and the absolute
majority of program committee

were against me and we didn't agree
that like they started to

delete my messages internally.

Michael: Sorry to hear that, yeah.

Nikolay: And I exited the program
committee like a few days after

the war started.

So but anyway, the experience I
had running these meetups offline,

again, up to a couple of hundred
people, it's huge.

It was insane, like 2015, maybe
16, when Postgres popularity

started to grow significantly.

Michael: So were these 1 day events?

Were they evening events like meetup
type things?

Nikolay: It never was more than
3 hours.

And sometimes I combined a couple
of topics, 3 topics.

The actually let's cover a few
like 2008 we had very interesting

event.

It was recorded again in Moscow.

After that, I was so exhausted
because I thought I'm a very bad

organizer and I don't understand
technical details and like I

cannot do it anymore.

But later people, many years later
people watched this recording

and said, oh, this was a real amazing
event.

What I did, I invited 3 guys from
the Postgres community and 3 guys

from the MySQL community, including
Peter Zaitsev.

And we had kind of battle, Postgres
versus MySQL.

That was fun.

But I was exhausted.

After that, I took a break, like
a couple of years, no events,

and then only relaunched it.

Not a couple, many years of break.

Then I watched how Josh Berkus
is running meetups in San Francisco,

and I realized that it can be kind
of interesting again.

And then I returned to Moscow for
like a few months.

Then I started visiting, it was
like, due to work and life reasons.

It was already after the invasion of
Crimea, but complicated things.

But I relaunched meetups there
and immediately got huge attention

after many years of...

You know, like before 2014-15 Postgres
was like kind of an outsider.

It was like you need to prove why
you choose it because it's

complicated, it's hard to maintain
and so on.

It's still hard to maintain, that's
why RDS and others have their

value delivery, maintenance headache
is solved.

But in 2014-15, probably because
not only because of JSON, I

think also because of RDS, popularity
started to grow.

Yes, yes.

And I felt like, you know, like
this, I need to choose some,

you know, how to organize meetup
once again.

Well, I have some connections.

I just ask some people who work
at big companies, can you host?

They are ready to host because,
you know, they want to compete

and show how good their offices are,
you know, because for them it's

advertising for HR purposes.

Michael: Anybody hiring is a really
good way of showing off.

Nikolay: Yeah, yes, exactly.

And you can just use it if you
want to organize.

Currently, the situation is different
because of COVID and remote-focused

companies.

This is another reason why my Postgres
meetups are declining.

We can discuss it a little bit
later.

So what happened, I decided, okay,
it will be Yandex.

And I said, okay, how many people?

They said, okay, maximum 80 people
maybe.

I said, okay.

And then they quickly asked me
to close registration because

out of, you know, coffee cups or
something like I said, coffee,

it's not needed.

Like we can do without coffee,
you know.

And then they asked a question I
could not forget they said this

is your event of course but can
we like grab like 10 minutes

and also present something that
will be about Postgres. I said,

well, of course, what will it be?

They said, this is how we migrated
the Yandex.Mail from Oracle

to Postgres.

I started to feel like, wow, this
should be not 10 minutes only.

And then Vladimir Borodin, who
hired Andrei Borodin later, which

is interesting, just coincidence,
last names the same.

And then they presented it at PGCon
later, like it was huge migration

from Oracle to Postgres, and the
reasons were interesting, solutions

were interesting, like super interesting
topic.

So I found the gem right away,
just trying to come to a big company

and use their facilities to host
meetup.

And then I remember, in the first
row I saw new faces and then these

new faces became kind of almost
like friends for me in the following

years because these guys worked
at large companies, large e-commerce

companies and so on.

And all of them used Postgres.

I was like, wow, Postgres is different
now.

This return was super successful.

So I have good memories.

But unfortunately, of course, like
this political situation and

so on, like I hate this, like we
are divided right now and some

people work at companies I don't
appreciate.

Not only appreciate, I consider
like kind of enemies already

because if a company has some,
any connections with the military

Russia, this is an enemy for me.

Michael: Yeah, I think your cat
agrees.

Nikolay: Yes, unfortunately, it's
sad but like anyway, this was

a cat experience.

We had at some point like a couple
of hundred people.

And you know, like,

Michael: Yeah, that's a serious
turnout.

Like I went to a recent London
PG day, which is so we have post,

we have conferences in the Postgres
world and obviously language

specific ones that include some
Postgres talks.

But we also have these PG days
that tend to be 1 day events full

day, not just 3 hours, kind of
like 6, 7, 8 hours full of talks as

well.

So maybe even 6, 7, 8 talks.

And it's London, right?

Like this is a hub for the UK and
possible to get to from quite

a lot of other places in Europe
quite easily.

And I think we had fewer than 100
attendees.

It was a good turnout, but 200
people for a three-hour event

is a serious turnout.

Nikolay: It also
depends on channels you use

to attract people.

But I honestly think the golden
era was when Postgres community

population started to grow, like
15, 16, 17, 18, these years.

Now it's declining and people don't
see big value coming to offline

events because, you know, the
problem with knowledge consumption,

like learning at offline
event, Of course, it's good you can

ask your question definitely, but
honestly, you can ask a question

in a follow-up under YouTube right
now and we will try to answer

as well, but in a synchronous format.

But the problem is if you miss something,
you cannot rewind.

This is problem number 1.

The second problem if it's boring you
cannot speed it up.

For example, or skip.

Right now I'm talking quite slowly,
so you can probably use 1.5

speed to still understand me, right?

This is efficiency, right?

And also tickets cost a lot, right?

Usually.

Like, okay, if it's the same town,
okay, it's good.

But yeah.

Michael: That's a good point.

It wasn't free as well.

So I'm guessing your meetup was
free for attendees?

Nikolay: Always free.

Michael: That's a big difference.

I would say though I think there
are these differences and I would say that meeting people

and connections that you make are
really interesting.

It doesn't always happen in person
events, but I've noticed when

you're involved in some capacity,
like as a speaker, I found

people approached me and it was
a much better level of conversation

with people when I was a speaker
than attendee to attendee type

conversations.

Like it's not always difficult
to have good conversations, but

yeah, I found as a speaker, it
definitely helped.

Nikolay: Some online events which
happen synchronously about

recording and then the ability to asynchronously
rewatch or watch

what you missed.

Some events try to have formats
like moving the speaker and people

who are interested in a separate
room or virtual room and then

you can spend time asking questions
and so on.

Actually, honestly, with online
events, it's also sometimes good,

sometimes bad.

Like I had, for example, a couple
of hours event when more than

500 people watched me.

The key here is that the Postgres community,
the core Postgres community

is quite small.

And if you want more people, you
need to reach backend developers,

like analyst people and others,
like maybe even frontend developers

and so on.

In this case, it can be huge.

But if you're just targeting the
same people all the time, it

won't grow fast, I think.

But the quality of questions will
be also different in this case,

right?

Michael: Well, there are big Postgres-only
events.

The one in Europe I went to, the
one last year, we both, that's the

one time we've met in person actually.

That was over 600 people at one event,
but it was 3 or 4 simultaneous

tracks.

So it wasn't 600 people watching
every talk or watching one talk.

It was split across multiple tracks.

And then, so it was jam-packed.

In fact, I was on the program committee
for that, which was a

lot of work.

In fact, we can cover that a little
bit.

I think it might be interesting
to people how those work.

But my point for that was it was
stacked each.

It was multiple days, 3 or 4 days
of 3 or 4 tracks of talks with

breaks with coffee breaks
and lunch breaks, but no

kind of like nothing else, no other
structure around meeting

people.

And then I've heard this phrase,
is it the hallway track people

call it?

So skipping the odd talk, sometimes
it's good to skip just for

a break and maybe if you run a
business, maybe check your emails,

or if you're still working, if
you're on call, that kind of thing,

just check in.

Or just go for a walk.

But the alternative is to hang around
and chat with other people that

are skipping a talk.

Nikolay: Right.

If you know the area in general,
the field in general, probably

even if you attend offline, what
I think all speakers and conference

organizers, event organizers should
do is collect...

Collecting slide decks in advance
doesn't work, because a lot

of people, including myself, prepare
them until the last minute.

But there should be a way to distribute
SlideDecks, the finalized

version, right away online.

Like when the talk is started,
SlideDeck is ready for sure.

So this version should be distributed.

It's a simple thing, but even organizers
fail here often.

They try to tell me, oh, you need
to send me your slide deck 2 weeks

in advance.

I have done it many times.

I never did.

And they cannot do anything with
it.

It's not possible.

Because there is no slide deck
yet.

But in the last minute, they should
collect and distribute because

others who probably are attending another
track or they're just working,

as you said, right?

When I attend the events, I usually
give my talk and I never

attend anything, honestly.

But it's just not efficient.

I cannot sit for 1 hour and listen
to this thing of which 80% I understand

just from slides.

It's not efficient.

I want to look at slides quickly
and then probably talk to this

guy directly.

Well, sometimes it works, sometimes
it's interesting, but not

often at all.

Well, I mean, my situation is probably
different because at the

same time I found myself attending
person-to-person rehearsals

or PostgresFM open talk series,
and there I enjoy, I can spend

1 hour diving deep and so on.

But regular talks, very often I
just see like this I know, this

I know, this I know.

And it's like, I would just rather
look at slides and catch you

later in the hall, right?

Michael: Yeah, I don't think you're...

to discuss some deep...

I think you're atypical here because
you probably know the contents

like it.

Most good talks start with like
some introductory stuff for people

that don't know the topic well,
that's almost never going to

be useful for you. But it's really
useful for other people and

I personally probably have a higher
tolerance for watching these

talks live partly because I know
less and partly because I think

there's this every now and again
there's a throwaway comment

that the speaker probably didn't
even plan to say but just happened

to come to their mind during the
talk that's not in the slides,

that is the most interesting thing
to me about that talk.

And sometimes 55 minutes of a talk
were not useful, but the 5

minutes that were, were so helpful
to me that it was worth it.

So I think there's...

Nikolay: I agree here.

And if a speaker says, like, you
can interrupt me anytime, I

usually also do it.

And this is good.

Why does

Michael: that not surprise me?

Nikolay: Probably, yeah.

Probably we can just cover some
additional, like, deep, narrow

topic during the talk.

It can be interesting, I agree
here.

But anyway, I think flexibility
should be good here.

People who want to watch it fully
in person, okay.

People who want to watch it later,
because sometimes too...

HighLoad had more than 3,000 attendees
and 10 tracks in parallel.

Almost always I saw some interesting
materials in parallel, like

I cannot attend them if I can't
walk between them, but they also

compiled everything and later,
a couple of months later, you

can watch everything.

They also published books.

I remember reading these books
from time to time.

You just download it.

Today it's easy.

We have a pipeline for our subtitles.

Michael: But a talk's not the same
as writing a book.

It wouldn't be a very good

Nikolay: book.

There are many goals why you want
to attend a talk.

Many goals.

And if it's pure knowledge, if
you want to understand some other

people's experience and some just
knowledge, you want to understand

some method how to do something
or complexities, you probably

don't need it now, but you might
need it later, like in half

a year, for example.

In this case, having a recording,
having books transcribed, it's

easy again, like with AI it's super
easy.

We have it, right?

In this case, this builds value,
this conference delivers, right?

And I remember these books were
helpful.

I was reading some MySQL talks,
which I didn't attend because

I was not interested.

But now I'm interested in this
topic and Postgres and I see MySQL

has something.

I remember there were talks about
it.

I just read these materials with
slides and text and you understand.

Then you can watch the recording if
you want or you can combine.

There are many ways.

Also some people are okay with
listening, some people need to

see it.

There are different people, right?

Some people want to redo, oh, I
don't understand, I want to rewind.

What I'm trying to say, offline
works for some, online works

also for some.

And if there is good material
and a good speaker, today we

can produce many results just from
1 recording.

I mean, 1 time you explain something,
and then we have all possible

variations of this knowledge and
we can help people consume it

asynchronously or synchronously
and in any way, like text, video,

audio, right?

Michael: I think you're right about
different people having ideal

methods of learning, but I think
there's also a point here about

accessibility.

And I mean, I don't mean it in
the kind of technical sense.

I mean, it almost from a, like
being able to afford to do things

and also being able to like traveling
to events is it can be

expensive.

Staying can be expensive.

Tickets can be expensive.

And in the, at least in my experience
in the Postgres world,

unlike some of the bigger language
conferences, we don't tend

to pay.

I don't think I've seen any many
events, if any, that pay speakers

or pay for travel or pay for a
conversation.

Nikolay: How lot always do that?

Michael: That's great.

Nikolay: Not pay, but reimburse
the travel costs and hotel costs

just to make it simpler to bring
better speakers.

Yeah, but not just

Michael: better but also just a
wider, well I guess that doesn't

imply better, does it?

If you're recruiting from a wider
pool of people, that your quality

will improve.

Nikolay: PGCon also I remember that
then, I mean PGCon paid me

in 2007 to go to Ottawa,
and it was amazing that

I was very young and of course
like lack of money obviously and

this helped me.

I always remember this.

Michael: I do apologize.

I think PGCon is the exception
or was the exception.

Nikolay: Right, right.

PGCon did it.

And that's great.

I saw they did it this year again.

Actually, I could use that help.

I just didn't use it because already
I can cover myself.

I don't need it.

But it was cool to see it was still
happening.

And I think it's the right thing
to do to help people come from

various parts of the world.

Michael: Yeah.

And with the sponsor, I think there
is enough money in Postgres.

I haven't tried running an event.

I do realize it's really difficult
to break even.

I think money-wise it is difficult.

But we can get

Nikolay: a lot of satisfaction from it.

It's not difficult.

It's a lot of work, but if you
do it with passion, it's not super

difficult.

And there are many companies who
are good sponsors, right?

For example, I don't know, like
big events, like some I participated

in, 3,000 people and so on, had
budgets like more than $1 million

per event.

There are many companies who are
attracted.

If you go with like banks, like
e-commerce, they come and you

just need to organize a lot.

This is like 24/7 work for a couple
of weeks before the conference

starts and you won't be able to
sleep if you need a huge event.

But it's possible.

Michael: Most of our events are
still run by very hardworking

volunteers, people that aren't
getting paid.

Nikolay: Right, but why we discuss
this?

Once they started accepting money
from companies, they should

stop saying volunteers.

They have money.

Just charge them more and that's
it and deliver better quality.

Just do it or like, I don't know.

For example, here we discuss the
controversial topic which I

have very strong opinion on.

If you have sponsors, you must
provide recording.

At least for people who paid, right?

Otherwise, for speakers, it's not
fair.

They come, they see, okay, you
brought like 20 or 50 people,

but the work is so huge it's better
to go to YouTube and then later

people who are interested will
be able to listen and see your

material.

So of course a conference in person
delivers better connections.

I had very good follow-ups, very
good.

Like they said, we attended your
talk, blah, blah, blah, super.

But I don't understand why I need
to spend a huge effort preparing

slides.

And this happened with the last three conferences
actually.

So you spend a lot of effort.

Probably you pay for your tickets
and hotel to come.

How do you do it as

Michael: a speaker?

Nikolay: I mean, well, usually
it's so, as we discussed.

There are exceptions, but for a
Postgres community, the norm is...

I don't know.

Okay, this is PGCon 2007, that was
an exception.

HighLoad, always paid.

Nobody else paid, usually.

I always come just because I'm
interested in presenting, you

know.

And not anymore.

I'm not coming anymore and I can
tell everyone.

Last year I had a rule, I come
only if you record.

Now I just don't.

No.

I tested once recently with Jung.

Ah, it was online.

I could come to East Coast, but
I decided to present online only

and I confirmed I'm not doing offline
events anymore.

So what I'm trying to say is a
very simple idea.

You do a lot of effort.

You have a break in your work.

Of course, if your employer pays,
this is a different story,

right?

Basically, the employer bribes
you here.

So like you have a vacation, you
visit, it's super fun, cool,

but if we speak purely about professional
efficiency like work

efficiency and so on, like effort
versus result, I don't understand

any reason why this talk cannot
be recorded because you did a

lot of effort and you want the
result to be recorded too.

It's super annoying when you spend
a lot of effort.

You came to some conference, you
found even 50 people, cool.

But you understand, if it was recorded,
maybe a few hundred more

from different parts of the world would
watch it and tell you something

interesting, for example, or follow
up, or like you would find

more like-minded thinking, right?

This is why you're presenting usually.

Michael: Several of my first, so
I was late

Nikolay: for the post-question.

Recording was super expensive.

Michael: That's the normal reason,
yeah.

Nikolay: I only have rude words
here.

It's not a...

Michael: Can I move us?

I think you've got a point, but
also it's difficult to...

All we can do is ask and see what
people do.

If they can afford to...

Some conferences are now saying
that these videos are sponsored

by this sponsor, so the sponsor
paid the money for the recording.

So that's a way of

Nikolay: solving it.

This conference has forgotten that
this is open source and we should

go low-key, like simple approach.

We all have like iPads or something
and we can record on a laptop.

We don't need like high, expensive
production.

Yes.

Well, it will be high quality.

The microphones are good already
enough, or you can buy a $100 microphone.

It's not expensive.

$100 or $200.

The Rode, they're like this, like,
how it's called, I don't remember.

This, you attach them, it's wireless,
super cheap.

I mean $200, what is it?

And then you don't need to record
faces at all.

All you need is just sound and
slides.

Michael: Yeah, good point.

Nikolay: It can be 100% automated.

And then one guy, one student, you
can pay this guy like 300 bucks

to publish these 20 talks or how
many you have.

That's it.

I honestly, like, for me, it's
such a deep problem.

Michael: I want to move us on,
but partly to advertise something

you've been doing, which is PostgresFM's
OpenTalk series on YouTube.

So if any speakers out there, I
know reaching out to them is

quite a lot of work, but if any
speakers out there have given

a talk that they wish was recorded
or was at an event that wasn't

able to record for some reason,
you'll have them on to the YouTube

channel to present and ask a couple
of questions.

And I think that's a really great
service.

So if anybody wants to reach out,
that would be awesome.

We can add to that series.

Nikolay: Yes, thank you.

Indeed, I did it last in this year
a little bit.

I don't know, like maybe 10 or
15 talks.

But I found like I don't have the capacity
for inviting people, but

if someone is interested and this
is a very interesting material,

definitely I will be happy to do
it.

Because for me, it's actually zero
overhead except one hour or how

much time do we need.

We do it online, nothing like no
special production at all.

And now we have a beautiful workflow,
or how is it called, pipelines,

that delivers very good quality
subtitles.

We have it, and it's super cheap
and super fast, and the quality

is super good.

I enjoy it.

Last time I said, in the previous
episode I said, "begin except"

blocks because I spent too much
time with Python recently.

And this pipeline, because it's
based on Whisper and GPT-4 Turbo

models, APIs from OpenAI.

It corrected me and in subtitles
you have "begin exception" as

it should be.

So it's a good example, but it's
very gentle correcting.

So with most corrections, usually
when you write a book, for

example, and the corrector starts
to bring some stupid corrections,

you know, like you stop it, don't
do it.

Here it's very gentle and just
uses the list of terms we've built

and extend from time to time.

And optionally we can have an article
or blog post out of that.

Talk also.

It would require some additional
effort, but I think it makes

sense if it's an interesting material
we can redo the talk if

it was not recorded at some conference
and then we can also have

a blog post out of it with very
low effort.

Michael: If people want, right?

Nikolay: Yeah, if they want to.

If anyone listening to this is
also an event organizer, I also

offer to cooperate here.

I can share.

It's not rocket science.

It's just AI.

Oh, nice.

If you have some recordings already,
we can ask authors if they

are okay with it.

By the way, with my talks, I usually
give in Russian, in the

past, there's a guy who just enjoyed
doing it and just publishing

these posts, transforming recordings
from YouTube to a good blog

post which was discussed on some
platform.

He was just enjoying, like he was
trying to find good materials

and then publish a good article.

It's like translation, but not
from English to some language,

but it's translation from video
audio to text because many people

are better consuming texts.

So we can do it if you have a lot
of video recordings of recent

events, we can transcribe them
and create blog posts out of it.

Michael: Or just get better subtitles,
right?

Like the subtitles alone are great.

Nikolay: Subtitles can be improved
or I still think if you had

a good event, you have a good opportunity
to reach out to a wider

audience if you convert it to text.

It's text with pictures from slides.

Illustrations.

So step by step.

And it's quite easy to read.

I also enjoy reading such like
transcoded talks.

Maybe sometimes better than watching
full video because...

Michael: Well, much more skimmable.

Nikolay: Yeah, skim through is
the keyword here.

Michael: The thing I wanted to
share about the, like for anybody

out there considering submitting
talks to conferences, there

were a couple of things that surprised
me being on a selection

committee.

One was that experienced speakers
often submit multiple talks to

the same, multiple abstracts to
the same event.

That shocked me.

It never occurred to me to submit
multiple.

I thought I've got this one idea
for a talk, I'm gonna submit that.

But the problem is as a selection
committee, you don't really

want to talk of a similar topic
at

Nikolay: the same

Michael: event.

So you happen to submit a talk
to the same event as maybe the

world expert on that specific topic.

Nikolay: Yeah, fun story related
to this.

I also usually submit multiple
because I know that this increases

my conversion.

You submit, you want something
to be selected.

If it's a narrow topic, maybe the
program committee doesn't find

it interesting for the audience
of this conference.

So you submit multiple, because
you have materials or pre-materials,

like maybe you already tried some
of them at some meetups for example,

or maybe you have a new version
of an old talk with better details,

like recently with DjangoCon, which
I presented online, It was

my seamless SQL optimization tutorial,
like a three-hour tutorial,

like insane.

But it was version 3 already of
this talk and they selected it,

although I submitted I think 5
materials as I usually do.

Yeah, it's just experienced speakers
usually have a lot of materials.

Bruce Momjian usually says, check
my website, brucemomjian.us,

I don't know, talks.

There are updated talks, there
are new talks.

You just select what you want.

This is what most experienced people
do.

But let me tell you a short funny
story.

During COVID, I actually decided
to return to PGCon in 2019-20,

but I wanted to visit Ottawa once
again.

And I did it this year, but previous
years due to COVID it was

fully online, unfortunately.

So I was selected in 2020 And then
they selected both my tutorial

and some talk.

And then I remember during online
session, Dan administrated everything,

like he was managing a lot of computers.

And then I remember I connect to
my second talk and Dan says,

oh, like off of recording, he says,
oh, you again?

Like I said, yes, I was.

2 talks were selected.

He said, it should not happen.

We have a rule to select only 1.

We need to fix the SQL which guarantees
maybe a wrong group by, you

know,
Michael: or something.

Nikolay: Constraint.

Yeah, constraint.

Lack of constraint, I guess.

I guess I hacked it and it should
not be so.

Michael: Probably a mistake by
the committee, but that's a good

point.

Nikolay: No, no, no, no.

He mentioned that technically it
shouldn't be possible because

the talks are obviously stored
in Postgres.

And when they select, they select
trying to have only 1 talk

per speaker.

So I guess it's a bug or SQL
and it should be fixed.

That's it.

But not anymore.

No more this event, unfortunately.

Michael: Well yes, so you're talking
about kind of from an experienced

speaker perspective, but I'm talking
about new speakers, people

that don't have experience doing
this, try and come up with a

second one.

No, no, no, so as a

Nikolay: speaker,

Michael: as a new speaker, and
I would encourage people to try

it out because I think it can seriously
improve your like the

number of connections you make
at an event or it definitely helped

my confidence talking to people
at events for some reason.

I did actually also want to make
an offer to anybody that hasn't

ever spoken at a Postgres event
before, I know it can be daunting

and practice is really helpful.

If you want to practice in private,
I'm happy to be a sounding

board.

So do message me.

I'm happy to do

Nikolay: that.

I think, yeah, yeah, yeah.

I also see that selection of talks
just based on abstract is

a very poor-quality event organization.

It should be done with, I think
it should be done almost always

with pre-vetting talks, because
it's mutually helpful.

Of course, it's time-consuming,
but it's mutually helpful for

both sides.

So the speaker has a rehearsal,
feedback, can improve, and the

program committee understands better
and has aligned content.

Because usually bigger events do
this.

Why are you skeptical?

I see your face.

Michael: I'm pulling your face
for anybody listening Well my

experience with the the big Europe
event We had hundreds

of submissions which but what which
I know I was encouraging

people to submit multiple, but
5 or 6 is potentially too many.

If I'm just talking about 2 or
3, and you get hundreds of submissions,

and to vet, maybe you need a two-phase
process.

1 is on abstract.

Nikolay: No, no, no.

It should be a four or five-phase process.

It should be five.

But of course, the process here
requires development.

It requires experience and ideas
and maybe this is for bigger

events.

For meetups, of course, a couple
of talks, let's go, and you

just decide.

But if it's big, okay, 500 submissions,
so what?

You see which won't go for sure,
just remove them, it will be

already 200, and so on and so on.

Then you have shortlist, shortlist probably is 100, depending

on how many slots you have.

Then you need, of course, you need
several.

For some people you need more than
the accent speakers.

I don't know, Peter Zaitsev, you
don't need to vet him at all.

It's really fine.

So automatic, fast track, we call
it fast track.

That's it.

So automatically accepted, you
just need to choose the topic

because always Peter Zaitsev is
submitting 5 talks, always, maybe

more.

But for people you see first time
or you have doubts, sometimes

people have doubts.

Also funny story, I was voting
as program committee member, 1

guy, and then I ask questions,
simple questions.

In the evening of the conference,
the guy says, I won't do it.

These questions are killing me.

I understand I don't understand
this topic well enough, but I

didn't agree with this guy.

I said he's good enough to present
still.

I asked some important questions,
but people can ask them as

well.

It's for preparation and so on.

It's not just to say failed and
they said no, no, no, we still

want you.

So I needed to spend time in the
evening before the event to

convince him to present.

I was like, please, you're still
good.

So we had another call and I just
said, you're good.

I needed to refill his energy and
level of confidence because

I actually destroyed it.

It also requires experience from
program committee members not

to attack.

I'm sitting like an expert,
I'm going to like for self satisfactory,

you're saying so hard questions
just but this is not right.

The goal is not to praise yourself,
right?

You need to help to improve the
quality of material and maybe

to give some idea what kind of
questions people may ask.

Sometimes you say, our audiences,
for example, usually are very

practical oriented.

You're from academia, let's try
to connect better, right?

And you start asking questions
and shift slightly focus here.

There are many nuances here, but
the idea is sometimes if you're

too offensive, it can be harmful.

And this is, I think, the story,
I was not super harmful, it

was just this speaker particularly
was quite fragile.

By that time, in the early days
when I was on the program committee,

for sure I was harmful
sometimes, I know it.

But I fixed it myself.

So yeah, sometimes you need to
put additional energy to fix it.

But if we don't have this process,
we are blind.

Like we don't know, it's like maybe
it will work, maybe not,

and people don't have time to prepare
better.

They're alone.

People always appreciated this
kind of process we had.

Always.
They said, thank you, now I know.

Sometimes it was in person, by
the way, like rehearsals in person,

if it's the same city, some people
say, I'm going to ask this,

you don't need to rehearse the
whole talk, you just say I'm going

to cover this, that, and this sequence.

And you have good feedback from
people who are on the program committee,

usually, the program committee is organized
from people who are CTOs

or something like that from big companies
that have a lot of experience

so their feedback can be very helpful.

Not always but sometimes, right?

But you at least understand what
to expect and improve and so

on.

So I think both conferences should
follow this process.

They don't.

Many of them, they're just lazy.

Good enough, you know.

We don't record, we don't do rehearsals,
we don't have money

for that or we don't have time
for that.

Michael: Or it's different people,
right?

Like I was on the talk selection
committee, but I had no other

like I wasn't paid for it.

And it was like it was free.

Nikolay: I wasn't paid either for
the program committee.

Never I was paid.

Oh, maybe we discussed it, but
no.

Michael: That's the problem, right?

If you're already looking through
hundreds of abstracts, the

idea of also looking at dozens
of talks and giving constructive

feedback on early versions of those
is just so much work for

unpaid people to do.

I think it's a lot to expect.

The problem isn't we shouldn't
do that.

It's that we should work out a
way to fund it and do it.

But on the flip side, I've been
at very few bad talks.

Like what's the worst that happens
if we give a chance to people

that haven't had it?

Nikolay: Super boring talk.

Michael: Yeah, that's the worst.

Nikolay: Audience knows material
much better than the speaker.

It happens.

Yeah, but not often.

Michael: How often?

Nikolay: Quite often.

Okay.

And always when it happened, it
was a mistake from the program

committee.

They didn't pay attention to it.

I had it.

I usually visited some talks for
control level, for high load.

I just visited and then I provided
feedback like, this is super

boring.

How come we selected it in the
first place.

It's super boring.

And we improved over years and
established a very good process.

And other events adopted a similar
approach.

I know it for like 100% and all
like some Java events adopted

and so on.

And they learned from us and we
also developed it further, we

learned from them.

It was a mutual process and it
lasted between 2007 to 2022.

Kind of like 15 years, right?

It was super fun and I think it's
a good experience.

But again, I think, like, the bottom
line from my side, it's

a lot of effort if you want a very
good quality event.

And I think offline events must
be with online components, at

least asynchronously with recordings.

And It's also, by the way, an additional
ad for the conference,

because if you distribute materials
with some delay and people

consume it even for free, they
know this conference exists and

it delivers good materials.

Next time, probably they will convince
their employer to pay

for a trip in person.

So this builds value over time
and brand and so on.

Michael: I agree, I think free
is great but even if you can't

afford to do it for free and it
needs to be some small fee, I

think there's also a potential
thing to be done there.

I think people are willing to pay
for these things.

Nikolay: In my opinion, many more
interesting things are happening

online these days, but okay, I
might be returning to offline

events in a couple of years.

Michael: I think there's something
to be said for local events.

I like going to any that are local
to me in the UK or I'm planning

to go to one in Paris that I can
get the train to quite easily.

Do you think that's quite nice
to meet people in the area that

are also working on similar things?

The more international ones I think
are interesting and the huge

ones I find quite daunting.

But I think there are some examples
that we can learn from.

Like, I think the Rails world recent
conference was organized

by professionals that were set
up and funded by companies that

use Rails specifically to do things
like improve the documentation

and hold a conference and do you
think and they

Nikolay: super helpful

Michael: people that went to that
conference I don't actually

know that it was made available
online.

I don't think it was, but please
correct me if I'm wrong there.

But people that went to that conference
had such a great time

and were telling stories about
it on podcasts I listened to

about, they were raving about how
well organized it was, how

many fun things were put on.

It can be done even for open like,
Rails is open source.

There's no reason you can't have
paid people doing a professional

job in an open source world if
it's a big enough community.

Maybe that's where we suffer a
little bit.

Nikolay: Yeah, well, yeah, I agree.

So, yeah, anyway, a multi-modal approach
is good.

Like, and a multi-channel approach
is good.

So I agree there are big events.

I just maybe happened to have some
special experience.

Michael: Yeah.

Thanks so much for sharing it.

Nikolay: Yeah, sure.

Sorry it took so long.

No, it's because...

Yeah, yeah.

Anyway, good luck to everyone
who is attending.

If speakers, if it's not recorded,
we can redo, definitely.

I just stopped inviting, but it
doesn't mean, just because I

stopped inviting doesn't mean that
we stopped fully.

And now we also have a pipeline to
build blog posts easily from

that.

Michael: Great.

Nikolay: I know blog posting takes
a lot of time, so this is

the way you spend 1 hour redoing
your talk without preparation

and then you have a blog post out
of it.

You fix some issues if any, and
we post it anywhere.

Michael: Although I have noticed
a slight trend towards some

slightly shorter blog posts.

I think some people are getting
back into blogging a little bit

more and it's okay to write short
blog posts.

In fact, this podcast is an entry
into a blogging event that

is encouraging more people to write
blog posts, no matter how

short or long they are or how much
work they are, I would encourage

getting into chat.

Nikolay: Just be careful with ChatGPT
for it, right?

Michael: Please don't, yeah.

Nikolay: Why not?

Okay.

Michael: Like, well, if you use
Chat GPT, we'll do that, but

please, if you use Chat GPT, verify
things that you're saying.

Nikolay: Not each word, but each
token.

A token is a part of a word, you
need to verify it very carefully.

Michael: And please verify it before
you ask others to check

your blog post.

I've been burnt by that, and it's
particularly painful verifying

it for someone else.

So yeah, that would be my advice
to people using Chat GPT.

If you want people to keep reviewing
your blog posts.

Nikolay: Makes sense.

Michael: All right, thanks so much,
Nikolay.

Catch you next week. And thanks,
everyone, for listening.

Nikolay: Bye.