 Hei, wrth gwrs! Hello! OK, awesome, awesome. Attention, everyone. So we're going to be starting the lightening sessions now, and I'm going to call our first speaker to come on board. Let's welcome him with our class of hands. Darlene. Hi, everyone. Thank you. Some of you have seen this already, but I was asked to do it. I want to tell you about being nice. I do a lot of hiring at Canonical where I work, and one question I'm often asked by candidates that I'm interviewing is, is this a nice place to work? And, um, interesting question. Now, you've probably heard this advice that it costs you absolutely nothing to be nice, and it's correct. It's true. Do you know why it's true? It's true because being nice costs nothing because it's worthless. It's a gift without value, and it's time that we stopped firing to be nice. It's a word that's come on a very long journey from quite different meanings to the ones you might associate with it today. It has origins in concepts around ignorance or stupidity, but now it's just become this kind of blah word that carries and says and commits and implies pretty much nothing at all. It's like an ineffectual ruler that's everywhere that we can't get rid of, and it rules us still. Only sadists actively want to hurt people, but we almost sometimes do things that can't be done nicely because they're going to hurt people. Sometimes we have to say things that we know will cause pain and distress. If you're a doctor, a teacher, a colleague or a partner, sometimes you're going to have to say things that have no nice way to be said. What we tend to do is put them off for the fear of hurting the other and then we do them badly. Our reluctance further deepens the harm to the person that we didn't want to hurt, and it deprives them also of the opportunity to do something about it sometimes. Never hurting people being nice is actually impossible, but we can be kind and kindness is what those people who ask me in those interviews are actually asking about. Not whether it's a nice place, but whether it's a kind place and being kind is completely different from being nice. The good doctor, the good teacher, the manager has to deliver bad news. Make tough decisions and do it kindly and with compassion, gentleness and empathy and they have to face up to the hurt and distress or the fear that no they're going to bring about by maybe saying something. So being nice costs us nothing, but being kind can actually cost us a lot. It hurts us to see somebody else's pain and it's worse when we're the ones who are bringing it about. It takes courage to look at somebody else's pain and saving people from all pain is simply an impossible task. Being nice doesn't help them, but being kind can help them because they get at least some truth, some comfort and the kindness in itself is an affirmation of some kind of humanity between us and them. And it's not that it's bad to be nice but the risk of trying to be nice is that you'll be a bad doctor, teacher, manager or partner or bad person to be in a relationship with. And those people who try to be nice and avoid it all costs the harm can do more harm because they want it to avoid bringing pain. And perhaps the horror of avoiding somebody else's pain so much that you don't do the right thing by them is not actually quite so admirable. So I think the word nice it's kind of decayed into this blah sound. I hope it might be swallowed up by something like the nice guy concept and we can get rid of it all together and it's a really horrible word. Maybe we could take it back to meaning precise or subtle but I think it's had its day as a word of approval. In the meantime I do my best to avoid being or saying nice because it's no gift at all to tell someone that they are nice. There's no commitment in it. There's nothing in there for them to live up to. You haven't given them something to hold on to. But if you tell them that you were kind, you were generous, friendly, encouraging, those are commitments. You're putting something on them a label that they then have to live up to. You had to dig deeper for those words and they will have to dig deeper to respond. You've given them a standard. So when you're tempted to be nice ask yourself if you're just flinching from the cost of somebody else's pain. Consider paying the price of being kind because when I'm asked if it's a nice place to work I can say no, it's something that means a lot more. I'm back again for a few moments because we're talking about some of the forthcoming conferences and so on so I've got a few seconds more. I want to tell you about DjangoCon Africa which is going to take place in Zanzibar in Tanzania in November this year. Website is DjangoCon.Africa If you love Python and Django and you love travelling this will be a chance to be part of something absolutely amazing. So come and talk to me. Thank you very much everyone. All right, thank you so much. Please don't forget it's good and it's not bad to be nice. See you in Africa, DjangoCon Africa. Okay so our next speaker is Philippa from Pi Ladies and after this session we are going to be having Sarah. I'm going to the front here so you are ready for your session. Thank you. It's there. Hello everyone. I'm seeing all kind of other things in my desktop. Yeah I know but I don't. Okay where is my mouse? Okay so my name is Philippa. I'm going to tell you why I'm biased. It's an unconscious bias like Nintok. This topic has been already talked about in this conference which is great. So let's go. So unconscious bias I asked chatGPT to help me to put a nice sentence. So these are biases that we have without being conscious about and they affect our thoughts about other people, how we interact with other people and they can lead to unintentional discrimination from our side. So why am I I'm biased, why did I chose this title? So this is me and my friend we are talking about jobs and we both work in tech and she was very frustrated about her job and after a while I just told her you know what if in my company they have a junior position I will let you know. And then there was this awkward silence and she told me Philippa I'm not a junior. And this is when I felt really bad that my bias just affect like the relationship with my friend that I said and I'm actually grateful that she felt safe to tell me that so but it could be another person that would not feel safe to tell me that. So are you biased? So I have a little game for you. There are pictures of people and you will think about you don't need to say anything but you can think about who do you think is what. So who is a doctor? A nurse. Electrician Or a secretary. Who is a teacher? The president? A driver? A maid. And who is smart? We hear this a lot here. Sensitive. Who is a data engineer? Last one. Who is a programmer? So I hope it triggered some thoughts in you. So types of biases that we have just from visuals just by looking at people there are three more or less. So AG we judge people by their age the perceived age of the person and we judge what they can do for that appearance. This includes everything from how tall you are, your skin color your face characteristics everything your clothes and there is the gender bias which is one of the most famous ones like the perceived gender of the person might make you think something about what they can do. Okay, takeaways and conscious biases is real and it might hurt people how you can control that be aware of your biases so you can stop them and avoid you from saying things you really didn't want to. Our Harvard has this project implicit online you can go there and make some tests and check your biases. One thing I really like that also was mentioned here is slow thinking don't jump into conclusions don't make assumptions just slow thinking take a step back don't think what people can do without knowing the people. Okay, be respectful be kind to each other I had the wrong one. All images were generated with me journey I didn't break any GDPR laws here so if you are curious about the prompts I used for those images let me know Stay in touch, I'm in Pilates Bratislava those are my contacts, see you around. Thank you. Thank you so much for that. Alright, you have the next speaker now. After which we are going to ask here Thank you. Hello everyone, my name is Rodrigo and I'm here to share with you what APL taught me about Python Now, by round of applause who knows Python here? Pretty good, pretty good and who knows APL? Weaker, which is good so let's start with a quote by Alan Perlis that says that a language that doesn't affect the way you think about programming is not worth knowing and Alan J. Perlis is not some random dude from the internet it's the guy that won the first quote-unquote Nobel prize of computer science which is called the Turingaworth and so I'm here to share with you how APL changed the way I think about programming because there's two things to programming there's knowing how to think about programming and then there's the syntax of whatever language you're about to use and APL impacted apologies, APL impacted the way I think about solving problems and about writing programs so what's APL? APL stands for A programming language A is a terrible name, I agree however it's important to realize that when APL was invented it was not a programming language it was an alternative mathematical notation and that's why APL is so infamous for its weird looks now dismissing this language just because it looks weird and not the same thing as biases against humans but the same line of reasoning applies it makes no sense it's as reasonable as dismissing Chinese because it looks weird because the characters are different so this is a very interesting language and I got to program with it for a couple of years and at some point I realized weights, I was writing Python code and I realized I'm doing something different and then I thought about it a lot and I understood it had been my experience with APL that actually changed the way I write Python code and I thought that was interesting and I shared a very concrete example now the main takeaway for this talk should be explore other languages play around with them and the more different they are from Python the more likely they are to have an impact and then it's up to you to filter out the good and the bad impact and try to learn from the good things so very specific example this is APL code if ages is a vector of ages of different people this is me counting how many people are 18 or over and the straightforward, basic, simpler way of doing this in Python is with a loop like this and what I'm about to show you is a couple of transformations that step by step should make a lot of sense but that will lead to something that may look surprising and it looks surprising to me when I wrote it and then I realized that I did that whole process in my mind because of my exposure to APL so first things first I'll try to create some symmetry here and I'm either adding one or I'm adding zero so that I have some symmetry I know that adding zero is the same as doing nothing but having it up on the screen makes the next step much easier which is to realize that you're always adding something what changes is the amount you have one or zero that's fine but now what you can do is if you can if you know that booleans and integers are tied together you can get rid of the ifs and you just have the condition I'm not saying this is excellent python code I'm just saying this works and then well you can take it one step further and you can do a some over age generator expression now again I'm not saying this is excellent python code I'm not saying it's going to pass the code review I'm just saying that for some reason I wrote this down and I realized it was because of the connection with the APL code which stands one to one I don't care about the for loop in APL but the remainder of the things are one to one and my exposure to APL made me actually very comfortable with least comprehension because suddenly I understood what to me least comprehension are about or the way I interpret them my mental model of how they work and so this is summing and just one other thing sum and APL now there's a bunch of built-ins and their APL counterparts and if you see the pattern on the right because on the left they're all related and I've talked about this in the past they're all reduced and I think I need to go so thank you very much thank you so much thank you for that okay let's up there why we are hi you're welcome we're going to Alvaro okay Ricardo are you there okay Alvaro are you there sorry if I didn't pronounce that well are you in the we'll skip the next one Florian Florian are you okay alright are you so hello everyone I'm Sarah is this yes so I'm going to talk to you very quickly about what is federated learning so that's going to be like a two minute introduction first a little bit of context what do everyone want well we want to have like these really great machine learning models either because we are data scientists or because we are using them but the problem is that all of this data machine learning models that requires data quite a lot of it and not just any data that requires high quality data and you want to have data that is representative of what you are going to do so you want to have like varied data and it brings a lot of problems it can bring some technical issues some legal issues depending on the kind of data you're working with but the most important part for me it's bringing a privacy issue because as a data scientist I want to have a lot of data but as an individual I want my privacy to be protected so how can you like solve this conundrum so there is a bunch of so-called privacy enhancing techniques which include you might have heard about it some differential privacy or maleomorphic encryption this kind of things and federated learning is one solution to like one solution or what part of the solution on how you can solve this that I privacy issue what's happening on the typical machine learning pipeline is what's on the left side it is like you are taking data from a bit of everywhere and you are putting it in a centralized location and then you are training your model on it what's federating learning is proposing to do is roughly the opposite is say okay the data is going to say what's generated and then you are going to make your model or part of your model travel to the data do rounds like that until you get an aggregating model which is generating on all of the data so as I said the field is growing a lot right now so especially for federated learning there are a lot of open source libraries which are getting quite mature and ready to be used federated learning obviously it's quite vast and so depending if you need some things to iterate for simulation very quickly or if you need something which is more industry grade because you want to deploy it in production depending if you are doing federated learning with a lot of mobile phones for example or if you are doing federated learning in the most industrial context where you only have like a few data centers you are going to have different libraries fitting more or less your need so I am the mentioner of one of the mention of SUBSTRA so I can talk to you a bit more about that we've got a hugging face space if you want to play out of it and have a demo to see how federated learning might be working also you know classical github, we've got a slack and all of that and more importantly I'm really looking forward to meet with other people who are working on this data privacy issue I know that there was some like data anonymisation talk or Lyon but really like please say hi if you're interested in these topics and thank you thank you so much that was really fun next we have Ricardo let's appreciate are you ready yes you're time starts already okay you got this hi hi a different travel program I'm Ricardo and I'm 11 and I play football and piano I'm learning python for example this program makes a square I came by train with my father we want to say the planet this is the several stops Latina, Roma, Bolsano Munchen and Praga children love aeropiton so bring them thank you for your time awesome, thank you so much you did so well thank you do we have the pie community Floriam Floriam first then we have the pie community nest pie community rep are you there? okay, alright your time starts now thank you almost so aftering.help or how I bought yet another domain create your hand if you know about pieformers.info it's like half of you or so so it's a website which shows you when dotformers was new style string formatting how to do formatting with like person formatting and then with string.formers well it would make sense to have f-strings there right? that's what someone thought in 2015 then someone seemed to agree in 2017 and the same thing happened again in 2019 I also looked kind of into that like maybe opening a pull request but there were some discussions around whether how the examples would really fit in and it looked like the project was created that unfortunately with lost activity dying around 2017 I took a look at the commit list and saw oh there's a version 2 okay let's check the version 2 well looks like that was that as well unfortunately so I thought okay why not take matters into my own hands and start aftering.dev cool little domain right I didn't do it for three years until one day before the pike on Germany lightning talks because hey you could tell people about this at the lightning talks so let's do it let's do it well apparently the domain was gone by then just a few days after I came up with the project idea and the domain was still free when I did gone to contacts pricing customers 7151 something if you're here in the audience I'd love to talk to you because fstring.dev is like very well used apparently well done I had to come up with a new idea for a domain so what about f3ung well kind of cool but 100 euro for like a small side project just isn't quite worth it but hey let's see like maybe there are some promotions going on for cool new top level domains fstring.cat why not 5 euro cool well apparently to get a cat domain which is actually a country domain of catalan you need to belong to the catalan linguistic and cultural community on the internet you know obviously like nyan.cat for example so yeah in the end I came up with fstring.help and it's exactly what it says it's a small overview of how fstring syntax works but maybe some things you didn't know about fstrings and a nice overview of the rest sadly not quite mobile ready if someone knows how to make a nice mobile ready website out of a Jupiter notebook let's talk about it at the sprints maybe I also teamed up with Trey Hunter who had this nice even more compact cheat sheets of fstrings which you can now reach with the cheat cheat button on the page I want to close with some news Python 3.7.3.7 was end of life which means all currently supported Python versions now support self-documenting expressions where you can use a form of the variable inside an fstring with an equal sign and in the outputs you get both variable name or expression it can be an expression here and the result also with upcoming Python 3.12 any Python expression syntax is valid syntax inside fstrings so despite my syntax highlighting here really not liking it yet you can do things like comments inside Python code inside fstrings if that's something you find useful last few seconds with another shameless plug I'm the maintainer of Qt browser a Wimlike browser basically like Wimium but without all the web extension limitations I might do some stuff on it at the sprints and if you want a sticker or a Python sticker feel free to talk to me thank you thank you so much that was right on time alright so let's have Alex from the py community and we also have some other communities joining you're welcome on stage is it on? yeah that's on okay we need your follow attention we're going to confront it with at least 15 questions you have to basically answer immediately and let's start with the first question who likes waffles whoop okay so py data is coming in September from partin to 16 who likes puzzle speaking of puzzle a euro side pie will be in puzzle august 14 to 18 this time didn't work so much conference it's like two days of tutorials, two days of talks one day maintain the spin tracks tickets are online please come so please join us at pycon CZ in September in Prague we have a beautiful venue which is an old monastery so it's going to be beautiful as you can see the tickets are on sale so I hope that you can all join us also one of the tracks is py data oriented so I think it will be really interesting for all of you so please come in September who have tried potatoes with mojo pecan we are in the pycon S and canary island is the best place in Europe you know and we will join us in October 6, 7, 8 this year is online and also in person so we expect you on canary island who likes puzzle not together with mountains of chocolates okay geopies 2024 will be in May 27 to 29 in Basel Switzerland link is there okay thank you we won't have chocolate for USA pie by the way who likes ketchup alright yeah so pycon US is in Pittsburgh which is apparently the home of Heinz ketchup in addition to also hosting pycon US for the next two years we would love to see you there I know it's really far from here to get to pycon US we have a travel grant program which you should contact us about but you know give us a little bit more time because it's all the way in May next year this maybe is a weird question but who like homemade food attention attention first time py ladies corn when December we're going to have keynotes talk panel and much more 24 hours conference more info soon who knows what this is Brindzol a haluski the national food of Soraka and you're all invited to pycon Soraka which will be in March next year who likes barbecue okay you know there's a big pie data barbecue coming up where is pie data barbecue happening because pie data is not focused this is all in Austin so it should be happening in Austin Texas right no it's happening actually here in this beautiful place this is Heidelberg Germany it's in southwest Germany and I'm happy to announce a fifth pie data barbecue with keynotes speakers you've probably heard of and yeah thank you who likes place trees so we're going to have pycon Portugal from the 7th of September it's the second edition show up and just today for EuroPyton participants you have a 25% off voucher it's just this slide so if you want it now or it's gone forever three, two, one five fish and chips woo I thought we don't have good food but pie data London 2024 in June and there will be a father's day we'll bring our kids as well we'll have coding things for kids who likes muscle next year we'll happen in Spain Vigo join us who likes dinner kebab pie data piecon is coming back to Berlin it was invented there so in April from 22 to 24 who loves Barbie soup taste it in Vilnaus Lithuania in April 2024 yeah I love communities please a round of applause for them thank you so much this is just the last one who likes pizza come on piecon US has been announced already see this cannot be piecon US this is genuine Italian pizza by the way and this is a less genuine Italian pie than pizza we don't have it so we have piecon Italy 2024 already we forgot to update this light anyway it will be May next year in Florence essentially the very same venue where this T-shirt Europe Python 2011 was collected and this is some reaction if you ask pineapple pizza in Italy so please don't do it and who likes spam no, nobody does nobody please let's clap for them thank you so much I love community and I love the action and the morale that they brought on stage please come up so after you we have Medell from yeah the next person oh okay this says are you there tell me when does my time are we starting okay can I start now oh we're already gone oh okay sorry hi everyone I'm Chin this isn't really quite as fun as all that food I'm really sorry but I'll manage your expectations okay so the first thing is that there is an open source educational resource with free material you don't need that sorry technical problems this is when I say I'm a data scientist not a hardware specialist okay yeah yeah yeah okay great so anyway there's this program thing which I oh no that's still not working yes great okay oh crap so this is sports python oh no this way which I authored with a seeded by a grant from the European Python Society this is basically an outreach program teaching python and data science to especially people who are from marginalized groups so this is me teaching people I've taught that's a github okay number two if you're a python he cares about sustainability I am working on a new open source project to democratize python for environmental advocacy and it's called pick a side plan or profit what is it so currently it's a not very good github page because I am a data scientist not a web dev you can obviously contribute to that at some point so what is this actually about it's about developing free code that anyone can use for data journalism so this is not a new concept you know putting stuff on a map so people can sort of see what's happening to the planet and it's dying and burning but I think what is more useful or interesting about this is that my this project it will basically have reproducible data science workflows so people can basically try and analyzing themselves so for example this is the college colleges in the US well in Boston in this particular case who have power plants which are excessive carbon emitters so polluters anyway so that's what it is you can look that up why so why do we need environmental advocacy and hence why do I think this project is necessary to democratize data journalism citizen science and education that is because there is a ton of denial and disinformation there's a lot of interested parties like big oil who are basically making a lot of money in the short term by effectively exploiting the planet AI generated junk is not helping nor is the free press under attack so where does the idea come from so I don't know what people think about me but actually I actually used to be the first head of data science at the foreign office and I actually investigated environmental issues when I was supporting UK foreign policy so because enough time has passed I did a cheetah at Pycon US oops this year you can look that up as well no go away so the kind of content which was in there was so this for example is deforestation in the Amazon basin this is actually really interesting this is the border between Bolivia and in this case the Ron Hondonia state in Brazil which tells you a lot about different governments governance in different regions anyway the point was there was a lot of ways of sharing how python open source software was great for democratising effectively you know citizens slash individual access to exploring topical issues had really good interest in feedback sorry by these kind of people and if you're from the geospatial community the fact that Esri was there at my tutorial was really interesting so anyway can you help to pick a side project by collaborating contributing stuff doesn't have to be technical can be design stuff anything spreading the word potential users really easy one all those people who just came on the stage invite me to speak to your meetup and also remember the US tutorials you have to pay 150 bucks to get to that so on top of your travel cost so this is going to be a good bang for your buck sourcing ethical funding so for example I say ethical because for example if you're from the UK and you're over energy they are kind of greenwashing so I don't really want money from them it's not to say that I want dollars it's actually because the funding which for example this Washington Post journalist who's actually someone involved in the PSF used to be a board member a lot of those funding are US centric so they're only for investigating US issues and unfortunately environmental issues are not US only and or you could help by building a community environmentalist tech fork and these are some of my contact details please get in touch or you can find me at the social I'll try not to be too drunk thank you very much and that was right on time please thank you okay up next we have Mida Mida you can come on stage and after which we have our last talk for this year visit yes and the time start now alright can you hear me? okay so at our team we ran into a ticket where we had to run Python environments in an isolated environments Python run sorry we had to run Python in isolated environments okay so environments have to be programmable so Python has to be able to define the environment define the machine type and run it the way we want it and we need to be able to choose different back ends to run those environments and the back ends should include virtual end, condom, mamba py and everything okay so we naturally built it we were happy with it and we made an open source project out of it and yeah this is the link to it and we have the QR code okay so how does it look the isolate we call it isolate isolate works with the back ends so here we have a back end which is a virtual Python environment back end and all you have to do is you have to create the environment you pass it some requirements you have to create it and then you open a connection to it and you send a function to it and it will run it in the environment that you defined this works so well actually that we built a product around it so now we have a product called false serverless and from where we wrap isolated around it and we have an isolated decorator where you define the requirements you define a machine type and you define a function called this function it does not run on your computer it runs on false serverless you don't have to think about infrastructure anything like that and how far can you take it you can take it pretty dang far we have an open source repository here called edit anything app it's a full stack application that has Python as a back end so here you see we have a simple flask application that has this isolated decorator with a machine and each endpoint is calling a Python function and in the function we are doing some ML inference and the repository also has some front end code that interacts with the deployed application so now you have an end to end full stack Python application so this is the front end of the application and as you can see we have a simple UI it's also open source so the QR code there is the link to the GitHub repository here we're calling one endpoint to create a mask of the bus so it's created I click on the desired mask I write a prompt for the replacing this is another endpoint that replaces an object inside the picture this is using control net and so we run it this is starting up a machine running the inference, killing the machine and here we have the picture so in summary isolate your isolated environments and edit anything is the app that we built on top of our platform and our platform called FAL this is our team, I'm very happy to be here thank you everyone and see ya that was awesome, thank you so much ok so let's have our last speaker for today we said you're welcome right one more oh great so I wanted to talk to you guys about terminals because we like them especially modern ones for one thing they support unicode specifically I wanted to show you two unicode characters upper half block and lower half block if you print them you get this nice cross pattern and if you print more of them you get a chess board now you can bring colors into it by using unicode so that would look like this but we can do more we can draw stuff so let's draw a bird so that's a bird it's a huge string but it's a bird, if you print it you get a bird it's the bird from Celeste but it's only a string so if it reaches your terminal then it gets printed which means a server an SSH server could send it to you so if you were to SSH to send me a bird maybe the SSH server would send it to you and it would get displayed but I think we can do better than that so what would happen if you were to SSH into a server which runs a gameboy color emulator would look like this so do you know what game that is? yeah you don't get sound over SSH that's too sad it's the legend of Zelda Link's Awakening DX thank you you can actually play that so what you just saw is a project of mine it's called Gambatte terminal and it's a terminal front end for an existing gameboy color emulator which is called Gambatte it's written in Python it's optimized with Cython and it's powered by two amazing libraries called SNK SSH and PromToolkit you can trade out you can either SSH into the server or you can install it locally using pip and run it directly with your own very legit ROM oops sorry and that's it you can check out the Repository and I'll see you at the social event the lightning talk for this for today for those that are going for the social event later this evening see you there and if you're not going to see you tomorrow thank you so much everyone and special thanks to the speakers thank you