Best Practices for Driving Productivity with Noteable ChatGPT Plugin

[Music] foreign [Music] and thanks for joining us uh we're doing this live stream on chat GPT with the notable plugin I'm Elijah Meeks here at notable uh I'm being joined [Music] Chad Skelton and Matty or maybe not maybe we'll start with Matt I think Chad might have oh sorry sorry oh did I did I freeze up yeah just just spectacular work for a live stream let's let's start out with the Frozen silent host um several of us from notable Kyle Kelly Chief Architect myself and Noel Catherine who's a front-end Dev um Chad Skelton Mackie are joining us you've been using the notable plug in a bit so uh why don't I'll let you two introduce yourself we'll start with Chad hi thanks Elijah so my name is Chad Skelton um I'm a former data journalist from the Vancouver Sun and uh now a teacher I teach journalism and data visualization at my home University Quantum Polytechnic University near Vancouver and online through the University of Florida and I've just been playing around with um the notable plug-in having a lot of fun with it um put a few things on YouTube my kids finally think I'm cool because I'm a YouTuber um and yeah so happy to talk about uh what I've found neat about uh notable yeah and Matt uh hi everyone uh I'm Matt gee I'm the CEO and uh Chief data nerd at bright Hive uh and you know uh have been we've been using notable both in our hiring process I'm going to share a little bit about that um as well as uh working with the plugin um on our core platform uh itself so uh just really love the experience I've had so far and have learned a lot and excited to talk with other members of the community that's great so before we jump into examples and talking about why we love the plug-in I want to give a little bit of room for Kyle because Kyle was the first one who saw this coming like I saw chat GPT and I thought well this is fun I'll go play with this AI find out what it thinks about theory of knowledge but Kyle how was it that you said no no this is this is the tool for making data-driven documents oh yeah yeah so um yeah so I'm Kyle I'm a notable as well and you know I've worked with notebooks for like roughly 11 years as in like literally developing on them but I think before that it was um you know through the lens of sage and Mathematica and kind of all the precursors that led to today and um I remember last year I was playing with the with the models that were originally out the gbd3 models before we had chat 2bt uh and I was like these are these are kind of neat there's like something that's going to be there but it was expensive at least I thought for for what you could do and fairly slow I was like I think there's some neat stuff that's going to come out of this we should experiment with it and look at it later um and then opening I released first this like they released chat gbt which was great and it was like pretty quick and they released that as a as an API endpoint we could work with um and then like it went further when they said oh we would allow a plug-in to do things and when I saw that I I immediately like the first thing I did was how do I hook this thing in to write notebook code like how do we make it do notebooks because it probably knows notebooks it is seeing a corpus of them it's seen people going through tutorials I was like I bet it could like automate running through a tutorial basically like what is its reasoning capability um and so I think like the as soon as as soon as we had plug-in access I was like oh we have to do this and it was like within within moments I was like this this has to be done we have to make this we have like a headless execution platform and notable so like in Jupiter you normally uh don't have a way to update your document um in automation behind it behind the scenes and you know some of the same people at work at a notable worked on paper mill which we used for scheduled jobs and we wanted to take that but make it part of an interactive compute and so we have this real-time user interface that we're basically letting chat gbt be a contributor a collaborator on the the document that you're working with and when I first saw it I just was amazed when I finally sat down and wrote a prompt and it just built a notebook Alessandro I'm glad you could join us do you want to introduce yourselves yeah uh nice to meet you I'm Alex I'm an ml engineer a physic and yeah I'll be using uh multiple plugins since I found it out on the plugin store I think some weeks ago yeah I mean has been actually I feel a key component of my on my on my workflow looking at data and mm models you know uh for the past weeks yeah I wonder like what was your first thought Matt when you saw it in operation like when you first uh played with it what did you think yeah so first I mean I I mean I didn't want to build it myself but I mean I had in mind as soon as I heard about code interpreters okay how do I plant is it into my notebook as soon as I saw something that's okay I probably know what this is I'm happy I see it um yeah and uh how I've been using it basically I would have really liked is that um it give me the output of notebooks uh so it's like a code interpreter with way more functionalities and I can plug in into other parts of the of the of the infrastructure that's for why for me is really helpful so I actually set up my notebook with all the files all the API I use and then I can you know basically automate away many things you know yeah that's helping New Zealand I think that it's it's super exciting that you're left with a notebook and you're left therefore with all the infrastructure and workflows that Jupiter's established right like publishing and scheduling things like that Matt Matt what what's your uh what's what was your first experience like sure uh so it was actually delightful and I you don't get to use that very often in technology but it actually was um uh and and so I had we'd been playing around at bright Hive with um using uh just uh opening apis for generating kind of um SQL code that we could run uh in a secure environment so not needing to pass data back and forth and I'd been kind of on the code interpreter um uh wait list for a while when seeing the uh all the tweets come out from people doing awesome stuff with code interpreter and I've been kind of frustrated I wasn't getting bumped off the the wait list and so when I saw the uh notable um plug-in get announced I immediately added it uh and the I this is my the very first thing I did with it um I the next day I had a set of interviews that I was doing for some fellows for uh in a position at a Bright House and I needed to come up with a an exercise for them uh that we could do on a live coding session just to get a sense for you know some of their skills and so I I just asked the uh the uh plug in this I said you know you're a hiring manager uh for the support engineering team and Education data startup uh create a notable notebook that can serve as a live coding test that tests the candidates basic abilities with python data profiling data analysis uh inference and writing custom functions and uh three things I was impressed with right away one is um the do you guys have done a great job in in making the setup pretty easy like very dead simple setup um and giving you instructions on how to do it if you haven't done the setup yet um so all I had to do is paste in the project and the first version of The Notebook was actually pretty good uh it came up with a structure that had you know sections for Task 1 data profiling data analysis inference and then writing some custom functions uh the only thing I didn't like was that it was using the iris data set it was kind of generic and I wanted to the case study we wanted to collaborate with them on um was going to use we wanted them to use actual education domain data so I said can you do the same thing but use this particular data set from the college scorecard which is put out by the department of energy or Department of Education and it did it updated the whole notebook um and and created a pretty good collaborative test so my first experience was was really fun well I think that what's what's particularly interesting Matt is uh we've just rolled out the feature where it can read The Notebook and rewrite it and so you could even have a notebook that was like your base uh interview notebook and then you could say create this notebook but for this kind of position right like with this kind of specification so you have your your base notes book that you're pretty happy with and then you're using it to create these custom notebooks that are designed specifically for the the role that you're evaluating and maybe even for the candidate right like speak in a language that this candidate or speaking uh use the metaphors that's going to most resonate with this candidate Noel you've been on the inside um here at notable and seen all of this stuff happening uh as it's been happening what's what's been your experience with using the plugin yeah I think um maybe I'm unique amongst software Engineers that I was very hesitant and like skeptical about why anybody would want to use chat GPT to do anything you know so Kyle's been Kyle's been on the train pushing it for a while and I've kind of just been in the background but once we got this plug-in right away one of the first things I did was like I'm coding something in the UI that deals with our outputs and how they display and so I'm asking chat GPT make me some test notebooks so it's already making my job easier where it's making tests artifacts for me so that I can test my work at notable so that's kind of like a meta use of it and then as kind of a beginner in data science I've found it just revolutionary for learning how to do data science how what are the possibilities so I'm obsessed with color what you see now is my screen um and I'm like I wanna I wanna analyze color data but I don't have any clue where to even start with that and so I really loved chat gpt's five things that I could possibly analyze and I never even thought of some of these things like color in nature right up my alley so from a learning perspective it's just fast-tracked my learning in data science and my ability to just work at a slightly higher level than I would normally be and get unstuck when I'm not sure what to do next I just asked chat gbt it tells me edits my notebook and it's I've made entire you know artifacts with things that I can go back to let me see if I could show an example while I have you I have this um this notebook I should have loaded it ahead of time that I got from this git repo Wiki art vectors where they took they created some vectors to do styling color data for artworks on Wiki art and the notebook itself when I first got it oh it's not loading I'm looking at this and I'm not a math person and I'm just like I have no idea what all of these functions are for and what are they doing and what are they telling me and so I sent the link to chatgpt had them edit it a version for me and they um told me what each of the function definitions is about and that's just one kind of really fun example of not sure what I'm doing get a Little Help From My AI tutor and now I'm on my way to analyzing some pretty Advanced color data I think that I mean I was just doing an example we just released a video I guess uh Chad and I being the YouTube stars in the room I have to I have to advertise our new YouTube video because what we have heard over and over again is how do I get my data into chat GPT so we released a video showing you how to connect to Json CSV bigquery snowflake Excel postgres raw text uh and uh when I was editing the raw text which was the the text of Hamlet I told it after it had run through a few things to visualize it for me and it said hey you can't visualize some of this stuff like some of this stuff isn't visualization which I I didn't appreciate hearing this data visualization uh person but also I understood that and I love that moment when chechi b t becomes opinionated right because it does help me to understand things better and also you know I think that I like it when it pushes back a bit because it does it does uh teach me or at least challenge me to to think more deeply about things Chad what was what was your first experience with the plug and what was that like well I feel like I'm coming this from from a very different perspective because I would describe myself as a failed programmer um I tried to learn a little bit of python my last few years when I was at the Vancouver Sun I think I built like three web scrapers and a couple of Twitter Bots and that's about all that I've ever done and I've never really gone back to it and so um I come at this with sort of not being not really even knowing what a notebook was or having only a vague idea of what a notebook was and what I've been sort of most impressed with is that like I've sort of been playing dumb with Chachi PT like literally just giving it a data set and saying please do some data analysis please do some charts um and I've been sort of surprised at how good of a job it does at sort of picking out some of the most interesting Trends um in the data set like looking at when bikes are most likely to be stolen and and things like that um this was an interesting one here where um the pattern here sort of shows that most bikes are stolen kind of around noon and dinner time and there's this weird Spike at midnight and I actually asked Chachi PT and this is something I had to describe in the video like um why do you think there are so many bike thefts right after midnight um and it gives a variety of different theories but one of them it says is like well maybe it's a data entry error like maybe it's because when the police don't know what time the bike is stolen they leave it blank and it codes it at zero zero zero zero and then I said well can you look at that and then it looks at this distribution of all of them and finds that almost all of these are right right at midnight uh and that's something that I do in class to try to see if my students can real to make that jump of sort of maybe the data is actually wrong and most of them can't and so I was pretty impressed that that Chachi DT caught that um the other thing that I'm finding useful again as a failed programmer is that like I sort of know enough about this space to know about some of the things that are out there but that I don't know enough how to use and so I've been sort of playing around with things like fuzz matching and stuff like that to see if it can kind of clean up data sets which is something that I would not have the confidence to do myself in Python but I'm able to do it kind of um through chatgpt and these notable notebooks and that and then I really like that I've got these code Snippets because I can then you know this is maybe a hope of mine that I'll sort of slowly learn a little bit of python because I can kind of look at it and I know enough of it to kind of start to decode what it's doing and it's really been delightful for me because I think what it really remember is that the reason I kind of gave up on coding is that I I couldn't stand the syntax I would get one little comma or semicolon wrong and the whole thing would break and it took me forever to figure out why but I quite enjoyed the problem solving aspect of of coding and so this has been quite delightful for me because it's allowed me to kind of play around with it again without worrying so much about the details although I will say it does worry me a little bit that um you know I may not sometimes know when it's making mistakes because I can't fully understand uh the code but I think it's really powerful and I and I've yeah I think it has real potential um even as sort of as the first step if you're going to do your work in Tableau or power bi or something like that just to sort of say hey like look at my data and and it might surface things that you might have missed otherwise right like so yeah I've been I've been pretty impressed at what it can do again with often with very very minimal props like literally I think my first YouTube video is 20 minutes and all I'm doing is please analyze this data and make some charts and then when it stops I'm like more charts more analysis and like I'm not telling anything more than that and it and it can get to a they can do a fair bit of of pretty useful stuff so so I've been pretty impressed at what it can do well one of the things I loved hearing you say was that you weren't familiar with notebooks I think that one of the exciting things for this product and for a company is to expose uh folks to data-driven documents that are in these computational notebooks um computational notebooks have been around they've been used by data scientists there's a highly associated with doing data science we've got Alex here he's a data scientist doing ml he's like a classic notebook user but what we've been hoping for is that is that more people could pick up notebooks and use notebooks for more tasks for data exploration data communication and so on Kyle we have a question from the chat that's probably best answered by you how many people together created this plugin and first before he Kyle gets into the full answer I want to say there were cast of thousands right because it's trained on all of the notebooks that have ever been written that were publicly accessible pick out yeah so I think I think uh and I'll say this is like a general statement on like plug-in development it's not like the plug-in is one piece that is the rest of what makes like the notable platform like how do you use notebooks how do you expose data connections all of that like um the plugin itself started with a tiny team of like two people that were like huh would this be a good you know you start off in the Prototype and then you start productionizing and devops leans in and helps with support and everybody's kind of adding on and so like to me the whole company created this plugin uh over over time like we we created a platform for doing headless execution of Jupiter notebooks um and then exposed it to the world and then the plug-in was just like a perfect like oh this this meshes exactly with that that's the plug-in aspect but I mean obviously you know when you get in and you're asking these questions and the reason that it knows so much is that there's all this open source that's been developed on over years um from you know the num Focus slash Pi data stack of pandas and matplotlib and and all the python syntax and all the tutorial everything that's been out there about what you can do with open source tools for for Python and visualization and data cleaning and all of that it's all out there um and that that's I mean that's really the beauty of it is that the world's knowledge about how do you work with these open source tools is is encoded and worked in here and all we're doing is trying to create the extension point that allows chat TBT to know how to use our platform to enable you to do this this data analysis or to create reports or you know refactor your notebook and kind of everything else whatever we can do to make you more productive you know it's kind of built on this base of all the Scientific Python tools that have been out there and it's really amazing um to answer the question yeah just a lot um Alex what what's been your experience using it uh for real work how's that how's that yeah actually I feel quite different from what phones have been saying here in the call uh in my case my UC is pretty firmware I mean I'm I feel pretty pretty familiar with both Python and tools so I don't use it to like write code for me it's more how I think of it it's like it's my co-pilot so usually is Duty for as a co-pilot to do better things in this case it gets a like a co-pilot with tools so it's just someone that's smarter to reason about I I never actually use the output uh like I never use the output that it gives me you know but it's really helpful to reason the actual most common use case I have I was actually seeing if I could share something by the way uh yeah I'll check anyway how I use it often I've reasoned with the plugin around how to visualize the data I want to visualize so it first starts okay here's the data I have here's what I want to communicate with my data you know here is the kind of analysis I want to do how can I go about it you know and usually I iterate a lot uh with example data on the kind of charts I want uh an example is it usually you know when we train ml models we have accuracy gain on this model you know and most of my job is to look at how the accuracy on different segments of the traffic and there is a lot of nuance in what what you are communicate and visualize the plugin really helped me when I just give some example data and I say okay let's try to look at this thing and give me back a chart okay that's not actually what I mean it doesn't suggest what I wanna what if we try like a scatter plot you know any give me a scatter plot okay actually make the size of the dots correlated with the x-axis so it's still a copilot but it's a copilot that can access any python library that can access my own data and I can even set up to use my own apis you know and um so actually I mean it's pretty indispensable because I mean yeah I was talking to somebody about it and I said it's like it's like co-pilot if copilot also built your whole airplane right like there's not a good I've been trying to think about what the uh what the metaphor is like it's your it's your contractor that doesn't sound as ambitious it's your construction companies it's they come back and help um where do you folks you know met I mean you started this by talking about like using it for to me like just a radical new use case that I didn't imagine myself of the of job interviews um where where do all of you think this is going like where do you see things in in six months I'll start with you Matt but like six months five years does this does this really fundamentally change things or is it going to be more of an inflection um yeah great uh great question and uh I can tell you that for us it's changed already how we hire um for this position that we are hiring for um and and very particularly one of the things we decided early on is to as part of the interview process allow them uh allow the candidates in our collaborative interview to use the plugin itself um because uh we want to assess for the skills that um that are in combination with the co-pilot we've been talking about like how how agile is this person in being able to work with um with uh the their notable plug-in in helping solve an analytics problem for a customer and so I think it the fact that it um within a relatively short period of time has changed the way we as a small company are thinking about a pretty important position um uh in addition to uh you know the dozens of my friends who are who've either have changed their business models uh or just change the the nature of how they hire or how the people work within their company um as a result of plugins like this I think that that pretends a lot um that the next six months are are going to be um we're going to see a lot of uh stories like the ones we've heard today um in domains that none of us uh have you know fully anticipated Elijah you might still be on mute yeah technical problems it's just lovely I'm just having a great time not using sound properly uh maybe chat CPT can help us out with that too um we've got a comment about uh transitioning to data science and they're using the plug-in with uh kaggling you see data UCI data request and it feels like cheating right like this idea of we had to actually talked about do we create like uh uh do we enter into every kaggle competition with this what do we you know what does this mean for those kind of competitions now it plays off of what you're saying Matt about the uh about the interview process but also sort of the changes in work practice Chad do you see much of this affecting you know data journalism do you think that that's going to get deeply into into that realm soon or have you heard anybody using this or things like it like code interpreter in data journalism practice um not a lot yet although uh one of um the Canadian data journals I know Roberto Rocha early on in Chachi PT was using it to try to sort of clean up some data sets like having a list of companies and having it kind of standardize them and things like that um I honestly at this point come at this more from a teacher perspective and and and I think the reason I kind of decided to get chat GPT plus was because I was slightly terrified about how this was going to change my job and what we're doing and that's sort of how I feel about a lot of this is that it's sort of I'm sort of simultaneously excited and terrified because in some ways I'm kind of Blown Away at how good this is already like not even six months from now like like like tap GPT and this plugin are making mistakes but they're actually making fewer mistakes than my students make uh in their first semester usually they're it's catching things that they miss what I worry about is those students then hopefully become better and can catch those mistakes and and I do worry about what's going to happen if this is kind of integrated into Power bi and Tableau and all these kind of things in Excel and people never get to the point where they actually know how to do it themselves or know how to catch the errors because it's so easy to just let the AI do it and I don't know how we're going to deal with that or or if or if it'll completely change workplaces and that you'll have a bunch of people making the basic things with AI and then a fewer people number of people that are reviewing it to make sure it's not filled with errors so so it's hard to sort of see how this is going but I do think it's really going to fundamentally change things um two other points that I'll just make um I think we talked about before the chat began but I'll just mention one of the things that I really like about notable is a very weird technical thing about Chachi PT is that the images of charts in your chats don't work when you go back there so like like I've got one open here um on bus pass-ups like when I go back to this data set I was scrolling properly but oh there it is like like the chart that used to be here is now a broken uh image thing right and so and so it's helpful that I can go back to the notable and it's not broken there right so I've got kind of a permanent useful uh chart there and then the other thing that that I found really neat about Chachi BT and I think is very powerful about it and Elijah hinted at it is that is that Chachi BT knows a lot about everything right and so and so yes it can analyze your data but then when you can you can also ask your questions like well why do you think that is so this was pass-ups of buses so when buses are so full um they leave uh passengers behind at the bus stop and these are the 10 routes that had the most pass-ups these are all in Metro Vancouver and I said you know why why do you think that might be and Chachi PT knows enough about Vancouver to say well this this is connecting metro town and UBC metro town is one of the largest shopping centers in Canada UBC is one of the largest universities in the country it's not surprising that that route would be really busy right so so even being able to make that leap of like we're analyzing the data but then what is this whole world that's around the data it knows more about that than most data analysis like well any data analysis tools I've ever used Tableau never knows anything about the data that I'm working with and so there's something quite powerful about a tool that actually can start to answer real world questions about what that data might mean so that that I found quite useful as well well we're coming we're coming up at the uh the end of the scheduled time I wanted to give Noelle a chance to talk about you know as because you're a professional software developer now you're in this profession as this is happening at a company that's building it do you feel any do you feel any um responsibility for destroying your profession and replacing it with a with a with a bot or several Bots I mean how do you on in to be more serious how do you how do you think that this is going to affect things oh that it's such a good question and I was just looking at a a tweet on Twitter earlier where a guy was comparing uh a programmer who does 100 code versus a another uh developer they had that was doing no code plus AI plus a little code and just how much faster it was for the the second person to finish the job and how much cheaper in the end and and how that's such a big benefit for uh releasing a product you know the the time to Market is such a big opportunity and yeah I think in my experience like having chat GPT or copilot helped me with code help me with data science especially me I think my experience as a data scientist not really is the best example where I can't just know zero data science and go get a job with Matt over here and pass his data science interview like I'm not going to get hired just using chat gbt and I think that's the key is you still need domain knowledge you still need the expertise still hallucinating and lying to me and so I don't think my job's gonna get taken anytime soon but I'm also a senior developer and I look at maybe these entry-level jobs where you're just getting started I think are going to be harder but at the same time it's going to be easier to learn it's our it's already helping me learn new things and and get better at my job so I'm hoping that it it leans more towards like a tool to accelerate learning and we've got an artist go ahead yeah oh yeah we've got a comment from the uh from the chat saying that this is like a Visa calculator Excel right that doesn't get rid of accountants or the calculator Should students be allowed to use calculators on tests is is that where you think we're going now yes and it's similar in art you know like photography from the days when we're using uh completely big mechanical photography to now digital photography it's changed a lot but it didn't get rid of photographers it just changed what was important to become a photographer and it's the same with programming I I think it's changing what you have to focus on when you're starting or when you're learning but you still you still have to have your brain your human brain to really analyze and make sure that your answers are correct hey Kyle we've got a question um does chat TPT also hallucinate with code less so or more noticeable I can tell you it hallucinates with data in a quite interesting pattern but we can get into that in a different conversation yep yeah I think so first I'm gonna I'm gonna oversell it for a second and say that like the plug-in and the plug-in model that attribute is providing here gives a store of knowledge like large language models kind of on accident are a store of knowledge when you know in reality they're they're really trying to understand the structure and semantics of like how do humans put together language from from text and it has all this knowledge and then it just tries to fake it you know it's going to guess what's the next token that I'm going to predict here or once you put the plugin in you're basically saying like this is real data and of course you know some of your data might be fake too like whatever you know all the all the normal problems even as a human are still there um and we're going to keep running into token limit issues or like you know how much data can we actually send chat TPT to give contacts um because when it gets overwhelmed it's just like you know I'm just gonna make up stuff I don't know what the rest of our conversation was I'm just gonna make it up so we will hallucinate and it and it means that you still have to be there as a practitioner you know a domain either a domain expert or as you know somebody that you know is used to working with notebooks whatever it is you still have to be the human in the loop the AI doesn't get like full control over what's going to happen there to like assist you help you be more creative um and it will sometimes do weird stuff so you need to you need to validate uh after after the fact um it's pretty good and like you know it can use the plugin as a store of knowledge but but that's that's definitely there um I want to add on to something from earlier before where we talked about like job replacing and everything else I think it's important to note that but I think that could only happen if humans were really good at describing their requirements for what they need in software or data science like the whole thing is operating off of what you describe and so like you know you can make better and better prompts and do that but that is the work like describing describing the thing you want to solve is part of the work and if more people get to practice that like I think Society will get better if people are like oh I needed some writing to describe what I'm trying to do is something that's like a software engineering skill that everyone needs to learn anyways is a data science skill everybody needs to learn like so I think I think there's plenty of opportunity for humans to be kind of enhanced with this I just want to add to that if that's it's all right uh I think that's such an important Point Kyle I we we work with a lot of organizations that I um that I think some may term like lower capacity lower data capacity organizations right where you don't have a deep bench of data analysts um and yet they work they have data that uh has the ability to significantly improve people's lives if they get their right questions answered and one of the things that I I've seen that I'm really excited about in this whole area is that a much broader set of folks uh are are able to pose the questions themselves and uh get answers back that are rigorous that are based on the data that they have that they need to have skepticism about and have be able to do a smell test and say that just doesn't seem right but they can all they have to do is be really good at posing the question and have a good enough filter to be able to say I can tell when something's BS or not um and those two skills I think a much broader set of folks can develop those skills and as a result benefit from um the uh the the data that they have on hand I agree I agree I think I want to go back to something else I want to uh bring this bring this to a bit of a conclusion you know when we were we had the comment about Excel people didn't it wasn't just accountants who used Excel once Excel really got good right I use Excel for all sorts of little tasks um it's it's used by I think half a billion people and so that accessibility that you're talking about Matt is is incredibly important because it shouldn't be that only those who know how to code get to make certain decisions in the book what was it to you know learn to program or be programmed that kind of uh ethic that came out of the late 90s and early 2000s and the other bit of it is is another Microsoft product and we're not saying this because we're getting added to things plugins when it releases um but clippy people constantly bring up clippy and say they don't want clippy but what I'd like to point out is that we all remember clippy we don't remember any of the other failed products that came up from Microsoft for 30 years we remember clippy because we did want Clippy we did want that we just wanted it that worked we wanted it not to be some kind of annoying thing that didn't give good suggestions it never really worked we wanted something like this that that can help us to do our work and to allow more people to do um more things with with these tools that we have access to and with that I'm gonna I'm gonna call it here you gotta say like okay I was gonna say like the difference with clippy clippy was annoying because it was being pushed on you and here users are asking for it they want it I disagree Kyle I think people remember clippy because they wanted they wanted I don't think because lots of Microsoft puts pushed lots of products and lots of features on people and we don't remember any of them we just remember clippy and clippy we remember clippy I I'm gonna die on this hill we remember clippy because we wanted clippy to work and so we all actually sadly remembered what clippy could have been and I think that we're seeing that now because everybody is in love with this uh this or not maybe not in love with it but we're seeing huge adoption about these uh AI tools so thanks everyone for joining us if you would be interested in uh talking about the work you're doing notable plug-in we'd be happy to have you on a live stream like this just uh get a hold of us and thanks everybody including the notable folks for taking the time to talk about this notable plugin thanks for the invite

Channel: Noteable

Views: 808

Keywords:

Id: 2IEMKXskOvQ

Length: 39min 24sec (2364 seconds)

Published: Fri Jun 02 2023