 I'm just writing more feedback stuff here. So yeah, hopefully everyone is doing well. So remember after the course or after we're done here, we will go to the learner zoom and get feedback on how it was. And actually, can we post that on the HackMD so anyone else who hasn't registered can see it? If so, can you send it to me by the other chat? Oh, let's see, were there any other notes, feedback? Maybe we can talk about what our... Yeah, so we're sort of thinking one of our plans now would be to run this course once a year in the summer. And then in the winters run an advanced course that's more on workflows and how people actually use things. Like examples of, okay, here's three ways people use GPUs and let's see. Or here's the way one of us configures our shells and stuff like that. So that's our plan for the future. Yeah, like if you have any suggestions, what would you like to hear and what kind of things you feel uncertain about? What kind of things, like just put it out there, write it out there, if you feel like, okay, this is the thing I don't yet know about and I would like to know about, put it there so that we can then try to in our future courses discuss about it. Because, yeah, like Richard said about the GPUs, they're not magic, but they are really complicated. And that applies to many other things because like in the end, many of these things are designed by companies, designed by fields that are very much like engineering fields. And through that, they end up sometimes like you have a lot of like standards and like APIs and all kinds of like complications that basically you have to learn about and usually like it can get tiresome to learn about things. And yeah, it's good idea to just like ask what kind of problems you have, what kind of things you would like to learn about. Then we can like use our like what we have learned throughout our job helping other people like to point you to easier solutions. Yeah. And your general thoughts on the live stream course idea overall, like what kind of courses is this suitable for, what kind is it not suitable for? Although in the future, if it had been easier to get rooms and we had prepared a bit more, our plan would have been to have a in-person room where you can go and watch the stream together. And then there's food and people there to help you live with your own computer. So yeah, that could make things work a bit more like that can sort of combine the passive attendance with the active attendance. I agree with this more how to run your actual program. So do you think we could simplify the real examples and do some at the beginning and then go back to the end and do them again? Yeah, probably, yeah. Like one of the complications like with this is like in the past, we had had a few examples on how to run like a math lab program, how to run an art program, how to run a Python program that were basically like, okay, you have the exact same Slurm requirements, you just switch which module you want to use, and then you just like launch it. But yeah, it would be like it's like this course serves two purposes basically, like we have this, we need to like try to explain how the queue works. Like there's some things that are like inherent in the HPC system, like the queue and stuff like that, that is like you won't encounter it outside of HPC system. Like when you're running at your own laptop, you don't encounter a queue. So we need to explain it sort of and it requires, it has many of these subcategories. So we need to like try to explain it and it's, it uses a lot of time. And if we include some of the actual program running there, it can get messy, like the analogies, they don't match anymore, it can get muddied. But at the same time, if we don't include actual programs, like the queue seems to abstract. So it's like this kind of like, how do we balance the both sides of the equation? I don't know, like maybe this time we went to abstract, that might be, maybe next time we do this, maybe we can try to figure out some examples that would steal, like the queue stuff and stuff like that. But at the same time, would like scratch that each of seeing like actual programs being run. Does anyone know of people who didn't attend the later days because they didn't like the course? Because maybe we have a biased example here, because of course, everyone who's answering now has been here the whole time. Yeah, we should have like this, like when you uninstall a program, there's always like, why did you uninstall it and choose a reason? We should have that done, of course. Why did you leave the course? Did anyone who stayed the whole time not like the live stream and hack and de-format? That's what I really am curious about. Really, I'm curious to people who didn't stay, did they not like the format, but should we? Yeah, definitely on the third day, we need to figure out better examples, like that's obvious to me at least, like it's very hard to explain the concepts without having some good examples, but maybe we should figure out a way to describe them. Maybe we need to Google a bit, like how other sites do it. Should we split bash basics? We have some bash basics course with the similar kind of... Yeah, I mean, so this is a good point. So some of our courses use Zoom, some do the live stream format, and the live stream format, well, there's a lot of setup here in my apartment to make this work well. I think it could be simpler, but basically the smaller courses don't do live stream yet, but I'm wondering if we should do fewer courses, but try to make them bigger. Yeah, it's this kind of like... this bash thing is a good example of... Sometimes in this, in order to use a machine, like try to know this kind of a cluster, you need to first learn how to do a bash, and then you have to first learn how to use Linux, and then it creates this kind of barrier of entry that isn't nice. Of course, it's understandable that because the systems are so complicated and these are the tools that make them manageable for us and for people in general, these are the tools that have to be used, but at the same time, the training part of how to get to use these systems, it's usually like... Yeah, there are some hidden requests in this pipeline that you don't know about, and that's not something we want to be there. We would want to solve these kinds of problems, so yeah, I think it would be prudent to probably have a big bash course. Also, if you like these courses and you like this and you want to come work with us, then yeah, we would like to, in the autumn, get some sort of junior assistant worker who can help with things like our courses and some basic Triton maintenance tasks, which would be a great way to start building your skills and these kinds of things. It would sort of be a different path from research, but still might be interesting for the right people and certainly compatible with a research career. Okay, should we go to the Zoom Roman? Yeah, for my part, at least, thanks for everybody for participating and asking great questions in the HackMD. It really makes the course a lot better for us and a lot better probably for you if there's a good rapport with chat and we can actually answer questions that you have because if we're just shouting into the internet over here, it doesn't mean that you get stuff done faster and better and if you're asking the questions, it really makes a world of difference for the course itself. Okay, so I will get the Zoom that we're going to for discussion and paste it in the Twitch chat and also in HackMD and then we can talk to you there. Okay, thank you everyone for being here, commenting on HackMD, talking with us, staying here until the end. At least, I've enjoyed giving the course as stressful as it might be. So yeah, thanks and see you next time. Bye. Bye.