Getting Started with MotherDuck Webinar

0:00[Music]

0:23hello everyone welcome to this webinar our uh you know live stream we'll call it for those Geeks out there who are allergic to the term webinar um we are talking about making analytics fun frictionless and ducking awesome uh AKA this is our getting started with mother duck uh stream and this is in alignment

0:47with our GA release last week so we're super excited about the ga release um

0:55and uh let's just see here last week

1:01June 11th we went GA um and what does that mean well uh

1:08the ga released means that easy data is finally here big data is dead easy data is here uh we have a production ready duck DB 1.0 release that came out two weeks ago uh unmatched efficiency of pricing and execution you know we strongly believe in in how amazing duck DB is at executing your SQL analytics queries um and we are backing that up

1:35with our pricing model uh which really takes advantage of that efficiency of duck DB we have sock 2 certification for those of you who need that level of of security and compliance uh we have org level sharing and of course backed by a worldclass team uh so we have two sections here today uh one is the introduction which I

1:59will be giv on how and why we built mother duck uh differentiating a little bit between duck DB and mother duck and an overview of the product and then we're going to follow it up with the second section which is a live demo uh which will be given by Nathaniel Thompson who I'm G to add to here so he can wave hi to

2:20everyone uh and Nathaniel's going to talk about the UI he's going to talk about how mother duck and duck DB integrate into what we call the modern duck stack uh and then programmatic access how you access uh mother duck from various languages including Python and uh web assembly so uh we have a lot of exciting things in store for you here

2:43today now bye Nathaniel we'll see you again soon uh all right love love doing that um and uh why mother duck why did we build mother duck um you know it turns

2:58out that we we build things based off of

3:02uh what we see in the world and with mother duck uh we saw duck DB first we saw how popular duck DB was getting how great the SQL language was how much folks were excited about duck DB's usability uh and the way that they're improving SQL um and we decided hey what what can we do here can we build a uh

3:28cloud data warehouse based off of this um and so we take we looked at Duck DB we said all right the creators of duck DB we really do want to support open source um and so we have Hest and Mark here who are the creators of duck DB it's built and maintained by the duck DB community and Foundation as well as duck

3:50DB Labs which is a consulting company that hanis and Mark run um and it came out of the CWI which is the Dutch national computer science research Institute and we talked to them uh this is Jordan tagani our CEO uh chatting with Mark and hanis in their first meeting and we said hey we're interested in building a cloud-based version of

4:14this uh are you guys in uh and so it's a little bit of unique model here where uh we we do Consulting uh or we pay them for Consulting uh through duck DB Labs uh and they have a owner share of uh of mother duck but you know we've been working together collaboratively to build uh duck DB in the cloud as a cloud

4:36data warehouse and and the founding team you know came from a lot of different organizations uh throughout the data industry and we all noticed that uh you know Hardware was growing but software was staying the same um and we said hey how can we take advantage of that and and duck DB really is the answer super lightweight efficient yet powerful SQL

5:02engine uh that we can do scale up instead of scale out um and take advantage of of the power of modern-day machines um and we'll see a little bit more of that later today but cloud machines are 400x bigger uh are M1 M2 M3

5:19M4 if we're up to that yet laptops are uh you know super super powerful there's a lot more memory available um you know when I worked on Google big query in 2013 uh the 60 and a half gigabytes of RAM was the max size of RAM on an ec2 machine uh now it's 25 terabytes so huge differences there so the founding team

5:41got together and said hey let's do this uh and now the team as we have a ga product uh is a much larger team uh and we brought up brought in Industry veterans from outside uh you know those initial companies and uh including you know people from companies like elastic looker confluent uh Microsoft

6:03and uh some more folks from all those other companies as well and we said hey we're going to build a best-in-class cloud data warehouse based off of Duck

6:13DB and you know really it's it's about these ideas which I started to talk about you know we have huge servers uh we have fast laptops we have uh the idea

6:25that data isn't quite as big as we always thought it would be uh and we should focus on usability and then of course duck DB slaps so let's supercharge it so mother duck is a serverless collaborative ducking simple data warehouse that extends the efficiency of duck DB to unlock multiplayer Cloud analytics at scale uh so what does this

6:48mean well it means it's a ducking simple data warehouse essentially uh we're just expanding on that a little bit here but it does all the capabilities that you expect in terms of sharing data within organizations and organ organizing that data um and it does it in a way that is

7:04uh really easy to uh build maintain and

7:09uh and use on a day-to-day basis we build it for two different purposes uh for two different audiences I should say um there's on the right here the classic data teams that that use data warehouses um the data teams you know are folks you know maybe one to two person data teams small midsize organization you know up through larger

7:31organization departmental groups who are really just focused on data so data scientists data analysts data Engineers uh you know who have this responsibility across the organization or across a department to focus on data uh and getting value out of that it's a classic data warehouse for that purpose but we've also found this other audience uh that is really in need of of of this

7:54type of technology which is app developers people who are building uh sass applications to uh allow exposing

8:03users data back to themselves so if you think about on the consumer side something like straa where you know you're doing your runs you're doing your cycling and you're taking that data and exposing that to the user so the user can see how things are are going with their uh exercise plan and uh you know

8:23the same analogy can be made to a lot of B2B apps so SAS works one of our big customers is uh exposing Revenue data to CFOs um but there's plenty of other use cases use cases like that so appdev data teams uh you know both of these audiences are really loving mother duck uh and the reason is because it's easy

8:46uh allows you to query your data across both your local machine and the cloud we call this dual execution uh allows you to share and update your data across your organization and uh then control your cost uh with fine grained pricing uh in a way that is is truly personal so every user of of mother duck uh ends up

9:09getting their own resources so the users aren't fighting with other users uh and this really eliminates a problem that's very common amongst both the appdev and and the data teams and then the Dual execution is really the idea that you can have one query access data that is both local and machine uh as well as data that is in

9:31the cloud you can also materialize data back and forth between the cloud and your local machine and all this works is because duck DB the full duct DB engine it's only a 20 megabyte executable it actually fits you know on your local machine in the client in addition to running on the server in mother duck um so you have access to that full power

9:54locally on your client uh as well as in the cloud and this can be you know the client can be a Python program or the client can be uh in web assembly in your browser uh but either way uh you get to experience more processing power by taking advantage of that local compute um and you can uh experience low latency

10:17of being able to access data uh and really drill down to your data on your on your local machine even though the data originates from the the cloud data warehouse uh the pricing is pretty simple right now um we we have a forever free plan um and then there is a free trial of the standard plan available uh

10:39but the standard plan is only $25 uh per month per organization um with up to you know 100 gigabytes of storage and 100 compute hours uh there's a great calculator on our pricing page you can figure out what the cost would be based off of your data and how much data you're adding and things like that uh and of course uh like I

11:00mentioned earlier we have the modern duck stack we are integrated uh with over 50 Technologies I think at this point across the data ecosystem with the ga launch uh we announced several new ones uh which are FIV Tran Tableau Monte

11:19Carlo um and you know these really allow

11:23you to do everything from orchestrating data loading into your data warehouse to visualizing your data uh to you know maintaining compliance and uh control over your data and I forgot one other bi tool that we announced as part of GA which is powerbi so all of the top Technologies here uh check it out mother duck will integrate with them um and

11:49it's been a real pleasure seeing how easy it has been for all of these these vendors uh across the data ecosystem or the modern data stack as as they call it it should be the modern duck stack but they call it the modern data stack uh you it's been great to see how easy it is for those folks to integrate all of

12:06uh duck DB into it so now I'm GNA say on to Nathaniel uh and it is time for the demo and I'll add him to the stage here and myself so

12:20Nathaniel take it over awesome thanks a ton well thank you all for being here and let's go ahead and dive into some real world real time examples of working with mother duck so the first thing to get started here is I'm here in the mother duck UI you can see here on the left I've got uh my sort

12:42of navigation within my world of of mother duck so that's going to be all of my notebooks which contain my SQL cells and then also that's going to be the databases I have accessible in my mother dock account and in our case we're going to work a lot with the sample data database today and everybody's mother duck account here you can sign up for a

13:05mother duck account for free at mdu.com and you'll have that sample data available to play around with so anything you see here today you can go ahead and replicate yourself on your end in your account for free so let's go ahead and get started by just seeing High Lev what kind of like scale are we able to work with in mother duck at kind

13:27of a day-to-day level uh in this case I'm just going to run a select star on this uh table of data from that's public

13:37data from New York about ride share we can see just that fast with the power of duck DB and mother duck I was able to get 18 million rows returned and so the

13:48the scale here is really really impressive and is one where we're able to take advantage of both the cloud and your local laptop to do some really interesting things to the point on your local laptop if we head here we can open up the column Explorer what the column Explorer does is it takes that result set we got here from this query and lets

14:12us ask follow-up questions on it and we can see it's got some little uh charts in here that we can poke around with um some interesting things we can see for example 28% of this originating Bas num column is null and so that might tell us something about data quality something we if that's unexpected we might want to

14:32dig into that further but that's a question we got the answer to without needing to run another query so we can see here we can click around we get additional info for example in this request datetime field we can see this split out over time and see a count and then down here we can even see for

14:51example a histogram of the different pickup location IDs uh over time as and look at these as well some something like trip miles we can see a similar thing and then we can see stats about the distribution here now the really cool thing about all of this is all of these statistics were calculated locally using my machine's power and so the

15:17colum Explorer uses mother Duck's dual execution so I get my Fallout questions answered really quickly because I'm using the compute power of my laptop to run additional query rather than always going back and forth and doing a round trip to the cloud mother duck is going to intelligently say do I have the data locally to make this happen and use uh what I've got

15:40already to do this without needing to grab additional data from the cloud or run that run that processing on cloud compute and if it's yes it's going to do that locally for me so that's why running across again 18 million rows here I'm able to do some pretty impressive uh analysis so let's kind of move on down here and

16:03we can see how aggregations do in mother duck here I'm taking an average of trip miles uh across specifically uh Uber

16:13ride so that's what this uh this wear Clause is going to do for us here and we can see that was incredibly quick uh and that's because duct DB is a columnar database and so I get results on my aggregations very very fast and that is one where we were able to you know take advantage of that hybrid

16:35querying uh and then also that that really really fast aggregation happening in the browser using mother duck wasm so we're using the power of your compute in the browser uh using web assembly in rui

16:51that's something we'll touch on later as well with some additional really cool examples I'm excited to share uh another thing we have here is duct TB SQL uh duct TB SQL is a worldclass SQL dialect I may be personally biased but having worked with a number of SQL dialects over time uh it is definitely my favorite uh and it just

17:15has a lot of things to make working with data ergonomic as ergonomic as possible and a lot of considerations so taking a look at some examples here I can say something like give me all the columns in a particular you know data set and let's say for example for something like New York I for whatever reason don't need to worry

17:37about the congestion surcharge column in my table anymore I can just select all the rest of the columns and we can run that here and we say give me everything except that congestion surcharge column and suddenly I get my results so I don't need to oh if I can't select star anymore I need to go and list out every

17:56single column I can just write it uh uh exclude it right there uh some other ergonomic things that are really nice about duck DB SQL I can Group by all when I'm doing aggregation so here um any number of Dimensions I have selected non-aggregate columns that is going to group by all of them and I don't need to

18:17go and either type out the column name aund all the different column names in my group by or do the one two three four five six seven and keep tracking oh I took one away and just Group by all of them I know I want uh I'll Group by give me my aggregation so very convenient another thing you

18:33might notice is we do have like trailing columns here excuse me tra trailing commas here and so I can you know add in additional columns I can you know it's a lot more orgonomic to make iterations on my queries without needing to go back and be like ah so many errors what am I what's what's going on with the syntax

18:52so it's here to help uh which for me has been really convenient uh another quick little example here was this first one right I didn't write a select star I said I just want everything in the table so from this table give me everything I want and here it is here are all the columns no problem and uh it's just another example

19:13and there are many more uh there are some great blogs uh on the duct DB blog about how they've built this ergonomic SQL and some other really cool examples if especially if you're interested in SQL syntax uh there are some pretty pretty impressive ones that make life a lot easier for kind of working with SQL day today so let's go ahead and head down

19:38here a bit and we can take a look at some other features of the mother duck UI so let's say uh I have a query here

19:48where I'm using a case one to assign the appropriate uh label in this case it's Uber and lyt again we're working with ride Shar data from New York and I'm looking at the total mile of the ride share trips and I'm splitting those out by the date and the ride operator so you can see I've got that that sum of trip

20:07miles and then I'm doing a date trunk on my request date time to just get the day so that we can Group by again I'm doing a group by all here so I don't need to go and say like oh we need to make sure we Group by by these non aate columns it's Just Ducky be will handle it for me

20:23so I'm gonna go ahead and run this and oh right so here's the thing looks like I messed up the syntax on this one very easy mistake that I have absolutely done myself this comes from Real World Experience where I've I've put the the

20:42time frame in the wrong spot thankfully uh mother duck has fix it which is our AI powered uh SQL fixer and so what this is going to do is fix it takes a look at any errors I get from my query and looks at the query itself but not the underlying data and so what we're able to do is we can say hey let's see uh

21:06based on uh that error here's what we think the fix is going to be and we the idea is we don't want to be intrusive right so uh if this looks like something that is helpful or is a correct answer to kind of what you were aiming for then we can just click accept and run in this

21:24case this is what I was going for so I'm going to go ahead and hit accept and run here and I get my results just like that the other thing I can do in the UI is I can head in and open this menu which will give me some options to do some additional analysis on my results in the

21:42table so in our case we can for example

21:46aggregate these by date without writing any additional SQL so the way we can do that is we just can hide that date TR that date column and we'll say we pivot it and now closing this menu I can go and I can see split out by lift and Uber right that's what we selected in our case when column I now have the total

22:05and then everything is just aggregated under that so I can get at some additional answers really easily without needing to write any more SQL on that super convenient for kind of just like little data exploration things uh another thing that's convenient about working in mother duck is I can work with data wherever it lives so in this case case I might have

22:30let's say some data on uh cab rides that

22:34is a little bit fresher than what I have uh loaded in my database already and that I've got stored in S3 in a parquet file so that's totally fine I can right here hit run and go grab and query that data in the paret file in S3 from inside of mother Dock and not only can I go in

22:56and query it I can go in and join that with data in mother duck or even data I have locally which we'll take a look at an example for a little bit later uh so that's a highlevel overview

23:10of the mother duck UI I can you know uh work with other people in my organization so all the data I have here I can go and I can share that with others that I have in my organization and say hey you should be able to to query this data as well that's the benefit of my duck is I'm not limited to

23:31just the duck DB on my machine duck DB the power of duck DB is available to everybody in my organization and I can add people in even if they don't know about duck DB but they know SQL I can have them query my tables and work with the data I've got uh surfaced for them so the the other nice thing is they've

23:50also got their own compute and so I don't have to worry about them overloading you know a a job I have running to load data into mother let's say uh every individual user gets their own compute whereas everybody queries against the same storage so we can avoid the need for you know uh creating a bunch of replicas let's say uh taking up

24:12a bunch of space so let's uh go into some

24:18additional examples uh out kind of outside of the UI mother duck is you know a cloud data warehouse so we're not just going to be using it in the UI uh highle before we dive into kind of like some specific examples we can use uh all sorts of different tools as Ryan mentioned we can orchestrate with mother

24:37duck so that's going to be I can use it with airflow I could dagster prefect Mage kestra any of those and and many more I can integrate with mother duck uh I can also load and transform the data however I'm already familiar with so whether that's DBT or airite or FIV Tran

24:58able to get data into mother duck the same way I would any other database and transform it in there and so let's go ahead and dive into uh an example on the bi front uh mother duck uh let me grab this switch screen here over to our next

25:23one there we go

25:30all right so here we are in Tableau because mother Duck's a cloud data warehouse I can share this data with other people in the organization right and so they get answers wherever is most convenient for them uh for a lot of people working with data uh bi tools are going to be a really important part of that workflow

25:51really important of that modern duck stack right and so mother duck supports Apache suet Tableau evidence Omni hex and a number of other bi tools as well uh and uh today we're going to kind of poke around a little bit with an example in Tableau desktop so here I am in in Tableau desktop and we can see

26:13looking at the data all of this if we look over here at the data source these are the tables we just saw in in mother duck right these are this this is the same data I'm connected live to my mother duck database over jdbc and so I can go in and I can do all the stuff I'm

26:31familiar with doing in Tableau mess with my measures um and you know grab my

26:39let's grab our drop off date time there we go and again this is live

26:46querying over jdbc against mother duck so another thing uh so this is works again in any bi tool I'm that works with mother duck and the list is just growing from here uh definitely check out the uh mother duck ecosystem on mother duck.com to see if uh the bi tool you're most familiar with or excited about is on

27:10there there's some really really cool ones if you're interested in like exploring what what new options are are out there there's a lot of a lot of innovation happening in bi which is really exciting so not all of us are going to be working in bi a lot of us are going to be working with data programmatically

27:28and so for working with mother duck programmatically there's going to be a lot of different options uh working with duck DV programmatically is really easy and therefore the same is true for mother duck uh we have a lot of language options whether it be CLI python Java

27:44jdbc web assembly no. JS go and that's

27:49just kind of the start of the list uh but today I was thinking we'd go ahead and see how easy it is to kind of go from scratch go do a whole a whole uh run through of what it's like to get started with mother duck in Python so I'm going to go ahead and switch over to vs code here so we can

28:08see

28:14that all right

28:23so let's get the we'll get that new new screen shared here

28:33awesome so here we are in vs code I don't have

28:40anything related to mother duck installed this is just my sort of standard python environment right so I'm

28:48going to go ahead and we'll get started here and we'll do a standard like import OS and uh here I'm am gon to I want duck

28:59DB so all we need to do to get started with duck DB and mother duck is run pip

29:07install duct DB that was really quick and that should address this for us there we go so we just run we have import duct DV and here I'm just

29:23grabbing uh an environment variable in

29:28my case I'm I've gone ahead and brought in an access token for mother duck that is available in the mother duck web UI so you can go into your account settings and grab your your access token uh that's something just to sort of save time and make sure we're not leaking anything here uh I've saved that as an

29:47environment variable already and so you can either save that as an environment variable or another option is uh if it's

29:55not saved as an environment variable uh we will automatically open up a browser window for you to log in with mother duck and save sort of a temporary token uh in your terminal context you can work either in the CLI or in python or whereever wherever you are without necessarily needing to go out and set set variables in advance if that's not

30:14what you're looking for so similar to the Food Network I have an example set up over here uh taking a look I've got

30:23loading in my my token and I'm going to

30:28start by connecting to mother duck and just saying hey show me the databases I've got right I'm just in a regular python file and I'm just using the this is the standard duct TV package all I have here is I'm just saying MD colon with an argument of my mother duck token in this string is where in the duct DB

30:49context I might uh work with like a local file or either a dctb file a parquet whatever it might be uh to work with mother duck literally all I type is mdon if I if I don't have that uh that token already saved and I just want to use hey go out to the browser let me authenticate that way uh I quite

31:09literally just type MD colon and then uh a quote right there so let's see the databases we have available to us by running this right now awesome so we've got the mydb and

31:26the sample data which is is what we're looking for and now what I can do is

31:33because again this is running on my machine I'm not like working in the cloud necessarily I can do normal dctb things so I'm going to go ahead and I'll uh attach a local duct TV database in this case I'm creating this local duct TV file which we can see right here and now if we show we should

31:55see the additional database there we go we see our local file showed up appearing here so in this duct DB context right I'm running the local duct DB package I can see this local dctb file right here that's the dctb database I can see these two other databases one of which is stored in my mother duck account another which

32:20another of which is my mother duck share right so this is data that's shared with me and mother duck both of these two data sets uh those live in the cloud but I'm able to see them and work with them inside of the duck DV package locally in Python uh taking advantage of the power of the cloud when it makes sense and my local

32:41machine when I'm able to do that so I can have a really efficient workflow and reduce costs uh while also getting a really really fast experience especially in follow-up questions so let's go

32:54ahead and run a query on on some data in mother duck so here we're using that same uh

33:04we're using that same New York data set except now we're going to look at some service requests they have for their public agencies we'll go ahead and run that query and we can see the results right here in our terminal duct DB has some really nice different kind of table layouts that you can play around with if

33:21you're interested this is the the default one that uh does some nice truncation and takes into account the size of of the window uh which is pretty cool so this is this is data that was living in the cloud right so kind of like I was doing earlier I also have a parquet file that I've got local here

33:41and I can go ahead and run another query and lo and behold here are my results this is a query on a paret file I use regular SQL right this is the power of duct TB I was used regular SQL to query that file I have stored locally which is great so next let's go ahead and like we did

34:05earlier again now we're in the programmatic python context I'll go ahead and query some data in S3 this is where things can kind of get interesting with different workflows and maybe some you know ETL and you know there's all sorts of interesting ideas here uh once you kind of bring the different pieces together so again I'm getting results

34:25and these are coming from S3 the nice thing about working with mother duck in the uh context of S3 is mother duck runs in AWS and so when I'm working

34:38with data in S3 I'm actually closer to the data than my local machine is when I'm using mother duck and so I'm able to get really fast uh transfer and compute between them uh which is very convenient when when that makes sense for my use case so going down a little here what I can do is we can take a look and see it

35:01how does dual execution work uh in a programmatic context so we saw it kind of built out in a really nice UI sort of user forward way uh with the column Explorer but here's a way that I can use it when I'm poking around with data in Python so I have a dual execution where I'm joining a query between a local file

35:26in this case I have a CSV that uh shows the conversion rates

35:32so using the uh that that taxy data we were looking at earlier we can use my local CSV right over here and we're

35:43going to use sources uh for currency conversion uh we use that as a source for currency conversion and then take a look at that data join it with data that's living in the cloud in mother duck so let's go ahead and run this and what's awesome is just how fast all of this happens so we run and

36:04there's our results and so we see here these are results using data stored in mother duck right in the cloud joined with data locally and mother duck is doing a query plan where it's deciding where it makes sense to run the compute whether it's like okay download the data and run that compute locally on your your local processor or let's get let's

36:26the dataa is already mother dock let's go ahead and do the computer mother Dock and send those results down it's all Dynamic based on where it's going to be fastest and most efficient for any given query so from here we can go ahead

36:44and take a final look at some quick hits let me go ahead and go back over

36:54here swap screen one more time for it

37:04all right so remember I can use mother duck like this across a huge variety of languages so that's node.js Java go web assembly python like we saw and many others uh there's one more thing that I mentioned earlier hinted at that I want to share with you uh this is web assembly in particular we're going to dig into a bit and it's a

37:32little taste of what's possible with the mother duck wasum which is web assembly mother duck mother Duck's web assembly SDK and Mosaic which is a framework for data visualization and exploration this is a really interesting kind of taste of the kinds of data apps you can build Ryan mentioned data app Builders and people you know if you're building data

37:56experiences this is extremely cool stuff so these examples that we're going to walk through are running live on millions of rows using a combination of my laptop compute and storage in mother duck uh to to build these the these

38:15experiences so let's go take a look at I don't know let's say uh let's go for Seattle weather let's say so I get results pretty quickly that was a pretty fast query uh again this is querying mother deck this is using SQL under the hood right so I just go ahead and let's say I want to look at a particular part of this

38:38scatter plot all I do is filter like this and we can see the data updating in real time this is the power of dual execution where I'm using the combination of results and compute storage in the cloud with the uh processing and storage on my local machine to make really really really

39:01interactive data apps another thing we can take a look at is 10 million rows of flights data uh in our case let's go ahead and filter like this no problem easy 60 frames per second no worries just filtering across 10 million rows cross filtering between these

39:23visualizations so you really can create true 60 fr per second data apps with mother duck uh with mother duck wum and Mosaic so it's a really exciting opportunity that um we're we're excited to keep investing in and uh see a lot of Partners kind of building out really interesting ideas here and uh we'll hopefully see more soon so I'll go ahead and yeah so I'd

39:52say uh to if any of this seemed interesting to you and you don't have a mother account yet yet definitely go ahead and sign up for a mother duck account it's completely free at mother duck.com you can uh get started with your with your own data and we'll also have we have a forever- free plan so you

40:10don't need to worry you know if you're just like toying around with stuff it's a really really helpful tool I use it personally for some hobby projects here and there uh which is really convenient uh I'd also encourage you to uh join our community slack uh you can ask questions there you can learn what others are using mother duck for and spend a ton of

40:30time with with mother duck experts if you have if you're just interested or you have questions um you know we have people from all across the company answering questions and helping out there um which is really exciting so you can check that out at slack. mother duck.com uh and I can go ahead and post that here as

40:50well all right that's in the chat there

40:54so thank you so much for joining this kind of highle overview of mother duck and for joining us to celebrate mother duck being GA we're really really excited to see what all of you are going to to build with mother duck and again we're really really hungry for feedback so if you have ideas followup questions anything definitely feel free to reach

41:21out to us in the slack or there's also links on our website where you can you can reach out but jump in and get started uh it's completely free so I'm

41:31yeah I can't wait to hopefully get to talk with many of you and yeah thank you so much for joining