Tech Radar Roundtable: What’s in store for the tech industry? | TLH #42

Speaker A: This is Tech Leaders Hub bringing you the voices of the tech industry brightest leaders. I'm your host, Natasha Mikojczyk and let's get into it. Hello and welcome everybody. Welcome back to Tech Leaders Hub. Um, today I am joined not by one, not by two, but by three different guests. And they have all, well, more than one thing in common. First of all, they all work at SDX Next, so they are my work colleagues. But don't worry, I will grill them just as much as I would my usual guests. And second of all, they were all co creators of Python Tech Radar. If you still haven't heard about Python Tech Reader, I have no idea where you live, probably under a rock. But Python TechRadar is the big project we were working on for the past few months. It has plenty of good insights about Python, about tech and about the future of uh, software development. So that's why today Jan, Krzysztof and Marcin all joined me to talk about their work, uh, when it comes to Python Tech Radar, what they did, what they plan on doing and what is in store for all of us. So welcome you guys. Hello.

Speaker B: Hi.

Speaker C: Hello, hello. Hello everyone.

Speaker A: Before we jump in into the topic, I want to ask you, the Tech Leaders Hub, what is your number one tip for tech leaders? And since we have um, a bigger crowd than usual, I think we should do it one by one. So maybe we will start with Jan. Tell us what you do, who you are and jump into your number one tip.

Speaker B: So uh, let me introduce myself, uh, Jan Plesinski. I work in SDXNext as a solution architect and uh, that is a technical arm of sales team. So I gather requirements from potential clients and propose initial architecture and team composition. And uh, when you ask me about my top one tip for Tech Leaders Hub, it's like super difficult to pick one. I have two in my mind. So I'll let myself present, uh, both of these tips.

Speaker A: Go for it.

Speaker B: Thanks. Uh, the first is, uh, before uh, you introduce change in a team that you are new at, uh, first gain trust. Trust is the most important factor, uh, that has to be established before introducing change. If you don't do that, even if your change or proposition is brilliant, is great, uh, it will be declined by the team and I tested that in real life.

Speaker A: Did you test that in a positive or the not so positive way where you established trust and it worked or you didn't and it very much did not work?

Speaker B: The latter, it m. Didn't work.

Speaker A: Yeah, that's what I expected.

Speaker B: But because of that I learned that thing. And the second tip is, uh, stay, stay curious, not judgmental. Like every situation that you encounter, it has some reasoning, uh, some history behind it. So first seek to learn it, um, and always be curious why things look like they are and uh, just cease to judge. Don't judge. That doesn't really.

Speaker A: Okay, that is a very good tip everybody. And I don't think it's a good tip only for Tech Leaders. I think it's a good tip for everyone in their life. And if we're talking about staying curious, I am curious if some of you recognize, uh, one of our guests today, uh, because Martin Zabava is actually a person who was already once on the show when Jakub was hosting. Uh, so. Hello, Martin. Can you tell us a bit about what you do now? Because I know your role has changed just a little bit and about your number one tip.

Speaker D: Yeah. So thank you very, uh, much for inviting me, uh, again to Tech Leaders Hub. I'm always happy to.

Speaker A: My pleasure.

Speaker D: Here. Uh, my role didn't change that much. I'm uh, right now the Director of Delivery at STX Next. So, uh, I'm uh, doing basically the same stuff, making sure that our clients are getting great value, uh, from cooperating with us and when it comes to tip for, uh, tech Leaders, well, I feel that uh, at current times a tip has to be connected somehow to AI. Uh, so, uh, the tip would be to uh, rethink your recruitment process, especially the tech assessment part, but not to make it chatgpt proof, but rather chatgpt embracing. Uh, because in the near future I'm pretty sure we will encourage, not prohibit the usage of code generating tools, um, uh, in our daily work. So we should check if the candidates know how to leverage them, but also how to avoid some dangers that uh, kind of tools bring. Um, so that would be my tip for uh, leadership roles in technology right now.

Speaker A: That is very interesting and we will definitely dive deeper into the topic, uh, later on in the session. But I'm very curious what you guys will say about the AI tools that have emerged lately and what was in store for us and for the whole software, uh, development world. And uh, as Kuba said, the first ever guest of Tech Leaders Hub returns. So it's a great return everybody. Uh, and our last guest today is, is Krzysztof Sopewa, who is an AI expert. Krzysztof, what is your number one tip and tell us, what do you do?

Speaker C: Hello everyone. I'm a head of Machine Learning and Data Engineering at stx. So my role was clearly connected with the AI and ML and Natasha before the broadcast say about that you asked about the OneTip. I really uh, have a hard time to figure it out. What is the one tip? Because you know, leading the teams in my opinion is uh, doing many things right. So many small things you have to do right and on time. But when I have to choose one, um, I think setting a vision and a mission for your project or company is um, the most important from my experience because people like to work for a cause, like to work to do, want to do the things that really matter. And um, if you choose the mission and the vision right that is, that is very appealing to the people. So many following mistakes will be um, will be forgotten. So this is my tip. Maybe m, maybe a little bit general than Jan and Marcin, but I really, I really like the feeling that I work for a cause and uh, for something higher than me.

Speaker A: Well yeah of course. I think it's a very true statement that all of us want to come into work and feel like we make a difference and we work for something that is, you know, something that is close to our heart, uh, and something that we care about because otherwise then what's the point? Right? Uh, so that was a good tip in my book. Don't worry. Uh, so um, starting out with the topic of today's conversation, as I mentioned, you are all co creators of Python. The creator, but we all know a co creator might mean a lot of different things. So I would love it if you all said just maybe two or three words about what was your role in the project, um, uh, what did you do? And um, I would like to start with Martin because I know that Martin was actually kind of at the point zero of creating the project. So can you tell us a little bit more about your role Martin?

Speaker D: Yes, uh, my role was to initiate uh, um, uh and bring the idea. But um, to be honest, uh, after the initial idea uh, how um, my friend Lucas Gmez and Paulina, uh, Kaiser took it and brought it to another level. So uh, uh, right now I'm looking at the results of Python Tech Radar and I'm uh blown away uh by uh, the evolution of this concept. Uh, but initially um, uh there were two inspirations for this um, uh report and we as a company as SDXNext, we're mostly famous because of our Python development services. Uh, so we are often picked by companies that uh, love Python and I just noticed that many of them, uh, for many of them it's important to give back to community and some of our clients uh sponsor open source projects. Um um SendCloud, one of our clients is uh a great example of this. Um and I was always kind of envy of this um and that inspired us to think what we can do ourselves uh in this regard. And uh we are already sponsoring some uh local meetups and conferences in Poland. But we wanted to do uh something uh for global Python community and we decided to take advantage of the large pool of developers uh of Python enthusiasts that we have access to um and to generate some useful insights uh for the community using this uh uh group of people who dedicated themselves to Python. Um uh so that was uh the main um inspiration and the second one uh which still remains reflected in the name of the initiative uh Python Tech rather. It's uh obviously technology rather from ThoughtWorks which uh uh I and also many of my friends here at STX read regularly and uh this is something that was always ah an uh example that we wanted to uh do something on a similar level. Um so that was my humble role at the beginning uh of Python tech rather and uh again I'm blown away by uh the extent to which it, it grew uh under uh Ukash and

Speaker A: mhm um we already know that

Speaker D: you

Speaker A: uh wanted to give back to the Python community and uh kind of share those insights that we have since we have so many uh Python developers here. But um, if you could maybe share a little bit more. Who do you think uh can benefit the most uh from uh reading and kind of uh getting the deep dive into the ptr uh and uh, who do you think should actually reach out for this resource um and when it can be most beneficial.

Speaker D: So there are two parts of ptr. First one uh are deep dive articles uh about future of Python and this part will be interesting for every techie person who uh uh like Python and use Python uh regularly and want to continue doing that. Um um what I especially like is that uh those are not like a typical articles that we uh publish on the web, uh about uh um um uh some summaries or tutorials uh or comparisons. It's really a deep dive. So if you're doing Python and you uh uh often think about uh the features of the language, the future of the language, how it will evolve, um uh this will be very interesting for you. And the second part is uh results of survey uh of uh um um I think more than 200 people uh using Python regularly and uh, what do they think about Python about uh the directions in which uh uh it will uh go. This is interesting uh for uh even wider group of people because uh even I and I'm uh more about business

Speaker B: um

Speaker D: was uh able to find um really interesting uh facts about uh the state of Python community. Um uh for example seniority of people. What does it mean today to be a senior uh in um people's own uh judgment. Uh how many years of experience uh do I have to have if I want to call myself a senior and uh uh be recognized by uh my peers, other developers as a senior, how many projects um I have to take part in? Uh so um, it's interesting uh really for everyone who has contact with Python, not only technical people but also uh those uh supporting roles, business people who uh uh decided to use Python to uh realize their projects and their ideas.

Speaker A: Okay, Marcino mentioned the first part that included those uh articles and today we're ever so lucky to be joined by uh Jan Pleszynski who was one of the authors of such articles. So Janik, if you could tell us a little bit more about what you wrote about and how did that work look from your side?

Speaker B: Okay, so maybe I'll start with the latter part of the question. How did that look from my side? Because the person who spread that idea uh in the organization was margin. But then it came uh, to a group of interested developers interested in writing articles. Uh and we started with brainstorming like with ideation of what may be the next thing for the Python and what may be appealing to readers. We came up uh with a lot of ideas out of which most we had to remove because it was uh impossible to write about all of them. Uh but finally we landed with uh three that we wanted to cover. Uh two of the articles were technical. Uh my article which I wrote with uh SimonDermohal, uh it was about uh concurrency in Python, basically about its history, presence and future. Uh yet another article was about uh performance increase introduced by Python 3.11. It's worth to mention that the increase that we have experienced as a community is from 20 to 60%. So that's a massive gain. And if you want to know the details of uh, why that performance gain uh is there, where does it stem from? Then feel free to uh read uh that article from ptr. And the third article that we chose to write is uh maybe less tech centric but it's about maybe more business related especially for startups. So uh, it's about the role of Python uh in writing proof of concepts and MVPs because um, Python is considered uh, a great choice for quickly ramping up your business. It's easy to learn, uh, easy to use with uh, broad ecosystem. So this is what we covered in this third article. That's basically a glimpse of what was

Speaker A: created, um, and how would you say the work looked like in a way that you know, because people came with an idea that hey, we want to create something that will give to the Python community. And how did you end up on those three topics? Why they were chosen over the other ones that you mentioned didn't make it

Speaker B: mainly because of expertise that we have in these topics. Like people uh, will do their best in the areas where they are most knowledgeable. So uh, every expert who chose to write to take part in uh, this endeavor, uh, picked an idea uh, for an article that was like first, Which was like something on the edge, uh, of the Python, and second was uh, knowledgeable in that area.

Speaker A: So uh, Python Tech Radar as we mentioned includes not only those three articles that are in the beginning, but also expert commentary. And one of the experts that did the commentary was Krzysztof Sapila. So I would like to ask you about your work and how that looked from your side. What was your role in the project?

Speaker C: So um, when, when I heard about the PTR and that our company stake came uh, to idea to prepare something like that, I think it will be very useful and um, especially for the tech leaders and the developer to see the direction of the upcoming changes. Not only the Python language itself, but I would say uh, Python ecosystem as a whole. My role was basically to uh, help the team to develop the PTR questions related to the machine learning and also analyzing the answer, trying to drive some kind of conclusions. Um, that's basically my role in the ptr.

Speaker A: Okay. And um, one topic that is already something I want to ask. Um, maybe all of you, maybe only one of you will want to answer. But I'm very curious because um, where we're talking about ptr, it was released what about two months ago and since then a lot of things have happened. Uh, ChatGPT4 was released. I don't even remember when the third one was released and now I'm only all about the 4th and uh, a lot of things have changed and do you think that it stands the test of time, the knowledge that is included in the uh, in Python Tech Radar?

Speaker C: Maybe I will start. So yeah, we are seeing a lot of velocity uh, in the technology, so everything moves really, really fast. But uh, even the last, most of the things that we included in the ptr, most of the questions uh, that have been asked, um, to the people and most of the answer are very related because they showed us uh, the upcoming changes. They show that uh, people want to learn. From my perspective, our developers want to learn machine learning. They want to learn the AI. They see that the AI is the future of the software development in many different areas, uh, as a support, uh, uh, for developers, as new things that you have to include in your product, uh, as a new way of working with the software. So um, from my perspective, and I try to think a little bit, uh, what is going on right now. I think it's stood the test of time and um, even if you read it at the moment, you will see how the upcoming months could look like. What is obsolete, what libraries will be obsolete and what uh, new technologies you have to carefully look at, uh, for a few months, uh, they will emerge or maybe you have to uh, incorporate in some way in your project. So I see, yeah basically I see that um, Python tech radar is a radar, so it shows on the spot of the ecosystem of the Python, um, what um, changes are coming and what should you focus about, uh, in the next few months.

Speaker A: Okay, you mentioned that by reading the radar you can find out what changes are coming and what you should focus on. And this is exactly what I want to ask you guys about today. A little bit more, um, not only about what's in the Python tech radar itself of course, but what is your own take on what is in store for Python and what is in store for uh, tech leaders and tech developers in the future. So uh, one of the big topics of course is that we all know that Python is the language to go when it comes to machine learning and when it comes to AI, and AI is the go to solution when it comes to tech right now. So I would love to hear a little bit more. Where do you think um, that is headed? Where we can kind of expect to see most advances and what actually we should focus on right now when it comes to ah, AI development, I think Krzysztof, you are the most.

Speaker C: Okay, so um, you know, in my opinion, um, you know, AI is a really, really hot topic and I think that it will stay with us for a really long time. I'm in a field from 2010 and uh, see how this uh, emerging technology, how this technology emerged from 2007, uh, I would say, but the beginning of the AI was uh, rather in a R and D and uh, department that was related basically on the R and D. But Nowadays most of the companies try to incorporate the AI parts in their system and in their products. So how the future will look like. So obviously large language models is something that uh, revolutionize the way we interact with the AI. We try to feel um, that AI is some kind of companion that we really spoke about, that we can ask questions and give meaningful and uh, mostly factful uh answers.

Speaker A: And I like how you said mostly

Speaker C: factful answers mostly because you know uh, it's not perfect but for I would say one year we will be uh, we will have those kind of models that are very truthful and factfulness and uh, we develop different kinds of technologies that will help us to um,

Speaker D: to

Speaker C: more understand the uh, answers uh, from the, from the AI. Um, and from my perspective this is, with this will be very very hot topic in the 2023. We have a lot of new large language models that we can commercially use. Not only chatgpt but uh, Llama Lambda Bard. Most of the big tech companies like Matteo, uh, Microsoft, Google Azure, uh show their uh, APIs, uh to the large language models. We can easily incorporate those kinds of functionalities in our products. Uh, there is a new field I would say that emerged from this technology that is prompt engineering. Uh, you don't have to be a machine learning engineer, you don't have to understand how those models uh, uh, work behind the scene but you have to understand how to talk with them, um, how to prepare the prompt, uh, that will be useful, how to assess the prompt, etc. Etc. So the whole new domain of the software expertise uh, was open and I think most of the, not only Python developers but most of the developers should learn how to write a prompt, how to interact with AI, what is capable, um, what are the requirements and what are the restrictions of using it and what are the risk of using large language models.

Speaker B: So um,

Speaker C: for me this year will be the year of the large language models and figure it out, how to work with them, how to incorporate them in your products. Um, and one more thing comes to my mind because uh, up to this point most of the AI related work was done by the R and D departments and there is too little AI managers, too little AI product, product owners, too little product managers, AI products managers, uh, too little people who can um, run the projects and understand how to run the projects with, with the AI AI components. So this is also will be huge I think, uh, and demand for those people uh, will be skyrocket because as we see our clients uh, wants uh, to work with AI and wants to start uh, an AI, new AI, AI project. So people who understand this, who understand how to prepare a backlog, for example, with the AI, different AI tasks. This is different mindset. Uh, this is AI is not, I would say is not. So, um. I forgot the word in English. It's not so. It's not, um, guys, maybe you will help me, uh, AI project. It project is uh, very strict, is a lot of different things that you have to deal with. But most of them, uh, when you set it right, uh, could be done in a fixed, um, amount of time. But when you.

Speaker B: Unpredictable is the word unpredictable.

Speaker C: Unpredictable and deterministic. So um, and AI and AI development is a little bit, uh, undeterministic. So you don't. It's, it's really hard to predict how much time, uh, the task will, uh, will last. But you can change, you can change the way you think about, uh, the task related to the AI. Uh, you don't have to set a task, prepare an algorithm for that. But you can say, okay, let's do an experiment that will last for, I don't know, 6, 8, 12 hours. And uh, after that we just uh, drive some conclusion from this experiment. So we will see the different, we will see the difference in preparing the backlog, in gathering the requirements, in uh, running the project as a whole. I would say.

Speaker A: And I'm very curious because, um, of course you Krzysztof yourself work with AI a lot. Probably the most out of uh, any of us here and more than three of us gathered together, me, Yannick and Martin. But I'm curious how it looks from the other guy's perspective. So, for example, Jan, you are a solutions architect and do you actually have the same feeling? Do you use AI in your work already? How does that look like for you?

Speaker B: Um, so yes, I use it every day, uh, every hour. First of all, uh, that's the most common use case. Just to fix, uh, my conversation with clients and other people, uh, I ask it to make my writing superior and that's.

Speaker A: Is that the prompt you ask it to make it superior?

Speaker B: Sometimes I do I have to try that out. Just that, uh, and that's basically more pleasant and understandable and approachable for other people to read. But on top of that, it also enabled us to do something that was, uh, unthinkable, uh, before, uh, in the sales process. And that is to prepare a proof of concept in a very limited, uh, time. Because, uh, what generative AI is to developer, to developers, and basically to every knowledge, uh, worker, it is productivity, multiplier like if you have um, a person who knows how to use AI, well that person can do the job of 10 people easily.

Speaker A: Is this the famous 10x engineer now? Just an engineer that can use AI.

Speaker B: But right now uh, there will be more 10x engineers and maybe like the baseline will be just different. From now on

Speaker A: when Everybody is a 10x engineer, nobody is, you know. That's a quote from the Incredibles guys.

Speaker B: This is why I mentioned that the baseline will just be different.

Speaker C: Mhm. Uh,

Speaker B: off topic, I heard the comparison that uh developers with uh generative AI are like uh, sports people on steroids.

Speaker A: Oh that actually is a good one. Yeah, I can see the parallel. So I wanted to, since we kind of arrived there early, um, let's do the dive already. Uh, I wanted to discuss kind of between us what we all think is going to happen for developers especially um, because there's been a lot of talk that is there still a future in engineering and is there still a place for new developers to pop up? And what is your take? What do you think will happen?

Speaker D: I have to say I really like this angle um, that uh Janek took that we as IT people figure out the technology that uh help us communicate with uh, outside people. So um, that's not something we should have uh, um uh predict. Uh but um, maybe I will take a moment to uh, support what uh Krzysztof uh said because I feel the same about uh uh AI that uh. It's uh um the change that we're seeing is that AI uh went out of R D department then for some time it was uh, in the uh, uh in the buzzword area uh where everybody who came up with uh a new product said that it's AI powered, that there's the secret sauce inside and it will be great because of this. Um and uh, um obviously in most cases it was like a really small part, not significant for the overall usability and uh, in most cases uh, it was uh, inferior to similar solutions that were not based on AI. And now we are with AI on the, on the verge of adoption. So um, uh we are seeing that there uh, we can already see it. There will be multiple AI providers uh, that uh, we will be able to use in uh, in our products. So it's closer to uh, to what we now think about cloud providers, HWS or Azure. So it will be uh, kind of similar. And um, uh it means that uh, uh even for developers who are not primarily focused on machine learning or even data engineering, uh, it will be a subject because There will be uh um some capabilities uh in the cloud probably that they will be able to leverage in their uh products by integrating with uh with it. So suddenly um AI became a subject for everyone and especially from for people who are um um conceptualizing new products and um um before uh uh most uh mostly uh AI was interesting for uh technical people or for futurists. And today it's interesting for uh

Speaker A: product

Speaker D: m, uh um uh owners, product uh managers, visionaries but the ones who are looking at what we have right now in technology and what cool stuff we can build with this. So that's a big change uh for me on the business and product level. But of course there's uh also um uh a significant impact on every developer because of uh generative AI tools. Um and uh there are different opinions of course. Many people fear uh that uh uh um it doesn't make sense to teach my kids programming anymore. Um but uh I think that for software professionals uh the advent of uh those generative AI tools doesn't mean less work but it does mean less like cookie cutter projects. And I think it's a positive change maybe for the people who ah are uh on the uh entry level in software development. This is not a good news because on those standard uh cookie cutter projects uh it's easiest to learn. It's uh also the easiest way to uh get into uh professional uh software jobs. Uh but for everybody else I think it's a good news because most people want to work on uh some cool innovative stuff and not uh another version of the same. Um and another thought when it comes to the impact of uh uh AI to software development is that I think that most people overestimate the code generation uh aspect thinking uh it will replace the developers. So um uh I will be able to just write um uh maybe even in a natural language what I want uh what the system should do and uh I will get back a code, the whole code base or the maybe a ready application on uh some kind of cloud server. Um and I don't think it will replace developers anytime soon. But at the same time I feel that many people tend to underestimate how um AI in form of smart chatbots integrated into ide uh will boost the productivity um the swift learning during coding and uh well for example if you uh have a new library uh that uh you think you can use those tools will give instant access to documentation uh of libraries or uh examples on how to uh use them. Um developers didn't uh uh have this before and I uh think it will be um uh a factor that will lower the bar uh in the same way um uh Stack Overflow lowered the entry uh level the bar for software developers 15 uh, years ago. Um. Oh yeah, that's uh what I think about uh this uh new AI um um stuff that is coming out.

Speaker C: I have two comments to the things that Martin talked about. Uh one thing is that the cost of the AI, the cost of the uh developing AI has lowered um drastically. So um, the case that Martin mentioned that most of the company uh say that they are doing AI and um under the hood there was only a small piece of the AI component there. Uh but nowadays um I think that the case was uh mainly by the reason that one year later doing an AI was really really expensive because you have to do a lot of experiments, you have to gather a lot of data and those kinds of process uh takes time and takes expertise. Nowadays uh, you can drastically cut the cost of implementing AI with your tools. Of course OpenAI, uh API is a little bit costly but it's far less than doing your own research, preparing your own algorithms. Of course in some kind of domains you have to prepare your own model on your data and the um model accuracy on the model performance will be uh much greater than uh the things you can achieve with the ChatGPT. But for most of the projects I can remember uh nowadays they can use an OpenAI tool with the proper prompting and a little bit experimenting with uh the prompts and they can achieve what they built uh previously for a month. So from my perspective the um decreasing of the cost of implementing the AI is really really huge, huge uh factor. And one more thing that Martin say, I also totally agree that uh we underestimate the capabilities of AI in the context of uh getting familiar with the new libraries. So if you know that this kind of library will help you, you don't have to read, maybe you don't have to read but uh, you don't have to spend too much time in order to find the examples to find uh different edge cases that this uh particular function uh has. The um AI will mostly help you with uh um generating the particular snippet of the code but also will help you ask uh help you answer the questions about the documentation. So this, this is, this will be really really huge to have some kind of body that you, that you can easily ask about different edge cases or different parameters. Uh and we'll show you what is the impact of the fourth or tenth uh argument in the function and to ah so from my perspective this will be also a very very ah big um and

Speaker D: I really look forward to uh the moment when AI will also be able to do this library discovery. Uh because in Python there's a lot of like the community is really numerous so there's tons of libraries for everything. And uh, you obviously struggle in picking the right one, finding, learning about the right one. And maybe this will be also uh.

Speaker C: This will be sooner than you think about because um there is a library called LangChain and there is an idea of an agent. So you can um chain the particular task from one model to another model and you can also incorporate the uh for example uh different systems uh in this chain you can query the Google search or query the I don't know, maybe uh PI index with all of the libraries and ask the question which is the most suitable for I don't know, data processing or data processing on the gpu. For example it will uh query the particular site, uh get some candidates and choose the proper one and say okay use the for example pandas, um

Speaker D: and then someone will ask what's uh the most suitable uh JS framework. And AI will waiting for that day

Speaker A: finally AI and not me.

Speaker B: And um, we'll have some guys poisoning ChatGPT saying that jQuery is the best framework to use. Everyone uses jQuery because ChatGPT told us so.

Speaker D: This is the future of CEO actually.

Speaker A: But while we're at the topic we actually have um a question that I wanted to address uh because Andrew in the comments asked um, based on your comments about the unpredictability of AI, do you believe that it will eventually replace professions where humans human factors are critical such as healthcare, social therapy as well as professions like law where a lot of information is already documented and the rules are pretty much the same for everyone. Since we kind of agreed that the software tech uh industry and all that world jobs won't be cut. The market can grow there. How do we feel about the future of other jobs?

Speaker D: I think uh uh it's fair to recognize that uh. M. Vast majority of predictions that uh were made by smart people uh about which jobs uh are vulnerable and which not uh failed. Uh so uh, everybody said that uh uh uh the safe uh job uh is the one connected with creativity. Uh software developer, artist maybe uh uh, visual artist too uh graphic designer, um and the jobs that are not safe are accountants, um um um uh like legal specialist, lawyer, um and it turned out that it's not the case. We have midjourney that uh generates uh uh great images uh we have uh. We had for more than 10 years uh products that generated music, uh on a level that uh. Was not uh um. Uh. Distinguishable from uh. Human composers. And on the other hand.

Speaker A: Oh yeah, you had that great story you told me once about that experiment with uh. Generated music. It's you who told me that now I remember. Yes.

Speaker D: Classical music. Yeah. And it was back in 20, uh, 10 I think so uh. Quite a long time ago. Uh, but uh. On the other, uh.

Speaker A: Uh.

Speaker D: Part of the spectrum, um.

Speaker A: Uh.

Speaker D: If we now think about removing accountants. It suddenly appeared that uh. Uh. A good accountant, uh. Needs to sometimes um. Navigate. Uh. On the verge of rules and not following, uh. Rules. Uh um. Uh. Very strictly.

Speaker A: Uh.

Speaker D: And with lawyer maybe uh. Law is very logical. Um. Has a very logical aspect. Many rules that can be automated. Uh, we had uh. Those uh. Smart contracts, uh in crypto world. Uh, uh. Even um. Decentralized organizations, uh. That uh. Were predicted to use those smart contracts to uh. To do legal binding things, uh when something happens. But there's also this aspect of a lawyer that is connected with uh. Figuring out what the parties want to achieve and uh. Mediating between them and argument, uh, negotiating. I don't see AI replacing this anytime soon. Um, so I think. Think there are still many surprises before us, uh when it comes to jobs which will be uh. Automated. Um,

Speaker C: my friend who is a lawyer and he says when you ask a question, uh. Uh, to two lawyers, then you have uh. Two different answers. So uh. It's not so predictable. Law is not so predictable. Everyone has its own opinion about the law and uh. Interpretation of the. Of the law. So I also agree that that will um. That those kinds of uh. Those kinds of jobs, when you have to think a little bit, you have to figure it out. What are the contexts, what are the cases, what are the corner cases, uh will last. But uh. The AI could help us, could help the lawyers to find uh. Another cases which um. Which talks about. About the particular. The particular manner.

Speaker D: Um.

Speaker C: From. From my perspective I see that those jobs when you have. When you have to have the contact with the human, like, Like. Like care. Like also the um. The law. When you go to the lawyer and you have to explain what is happening, what is going on. And most of the time the lawyer, um. Will act as some kind of psychotherapist and help you.

Speaker A: Oh my God, your lawyer has to be a poor person, very sad person,

Speaker C: maybe a little bit exaggerate. But uh. When you have a real serious problem and you go to the lawyer, uh. One of These job is to um, say about all of the consequences and all of the things uh, that he can do, uh, uh, for you.

Speaker A: So yeah, I mean I feel like most jobs nowadays have a human factor. Like we say. Oh yeah, there are jobs where you have to interact with people and there are jobs where you don't. But I mean it is true technically but also you always work with somebody in mind. Even when for example you're a software developer and you don't have that contact with a client, you still work with a client. You have to know what they want. You have to kind of understand it. And it takes more than just kind of getting a brief. You know. I feel that yeah.

Speaker C: And I have to, I have to add to this that you can pass this context of the whole situation. It's very hard to pass this context of holding situation to the prompt. Uh, exactly. That's why we have to communicate that we have different uh, tools for communicating. The information flow is much m higher at a much higher bandwidth than normal. So you can describe everything in the prompt and all of the context in the prompt. So I think we are safe. One more thing.

Speaker A: The machines won't take over hot new guys machines.

Speaker C: Maybe contact with the human will be I would say um, like a ah. Premium. Premium service. So at the basic level you will talk with the AI agents and if you want something, uh, you want to accomplish something harder, something more uh, complex then you will have to uh, talk with the humans eventually. So.

Speaker A: Mhm.

Speaker D: This is very grim future. I, I already hate uh calling anyone who has this automated uh, uh system that uh uh um, that tries to answer my problems. Uh, and I, I think you're right. There will be more of it.

Speaker A: Yeah, that is true.

Speaker C: Like when you have a really simple question and really simple and what a really simple answer. Uh, those systems are very, very helpful to the company. So uh, for example, if you have a question about the um, internal policies uh uh, in our company and you can't find it through the search, you can easily chat uh, with the chatbot and you can say okay, this is the answer. So uh, in those kinds of situation it will be really helpful. You don't have to deal with other people and the bother their minds.

Speaker A: Mhm. And Jan, you wanted to say something

Speaker B: but we kind of jumped in. That's true, that happened. But no offense, my take on that comment from Andrew is that in the short term I think that we are all safe and the AI will only improve the quality of, of these services. So for example healthcare people will have more time to speak with people like, as you mentioned, like for this, for this human to human relations rather than interfacing with technology, which currently takes them a lot of time. The same like for the law, maybe the legal cases will not take us, uh, three years to complete, but maybe two weeks or week and we can serve more of them. Like, I heard, uh, a statement that there will always be some computations to compute. Uh, and I stand behind that, be that legal or healthcare computations or whatever. So I think that in the short term we are fine, but we have to embrace that AI because people who will embrace it will be, uh, more capable than those who want. But I think that in the long term, uh, like 10, 20, 30 years in the future, the machines may take over, uh, but that may be not that bad future because we might then as humanity just do pleasant things and talk, uh, about philosophy, uh, solve some math problems and party.

Speaker A: That is a very hot take. Everybody. Machines might take over, but, you know, let's make them philosophers.

Speaker B: No, let's make us philosophers.

Speaker A: Okay, Okay.

Speaker B: I mean, it is not machines. Let's do pleasant things. Like, let's indulge in like, intellectual and physical pleasures.

Speaker A: I mean, the point of creating machines.

Speaker B: Crunch numbers.

Speaker A: Yeah, it is the point of creating machines to make life easier for ourselves. So maybe let's take advantage of that and, you know, work less. If the machines can be programmed to do that for us and just enjoy our time here, uh, that would be a nice future. That's true. Okay, I can get on board with that.

Speaker D: That's also not a new idea. I remember there's this famous poem about all watched over by machines of loving grace. It's from the 60s, I think. Um, uh, you can Google it out. Um, it sticked in my mind. So maybe we are reaching this soon.

Speaker A: So, uh, going back to the tech sector and not the overall world, so. Sector. Um, I wanted to kind of ask you some, Some last questions about, um, taking all of that into consideration. What we mentioned, what we talked about. Um, I want to know what you guys think about. What is the next big step? What is the most important thing or action that people can, um, take to kind of prepare for that future where we, uh, as Krzysztof said, 2023 will be the year of AI. What can tech leaders do? What can software developers do to be prepared for that time and kind of use the most out of it? If you had only, like, I don't know, one. One step to take. And that is the moment when they think of their answers and they cannot think of any. So uh, no, but you know, just think about it. If you only had one thing to prepare for the, let's say next year and new advances, what do you think would be the most beneficial action to take?

Speaker B: I would say that just keep, keep an eye on what's happening, on what's trending because changes are super rapid, the velocity is high. Like every week we, we are like the new technology is getting released. So just keep an eye on it, try to get your hands on everything, uh, just for uh, a small amount of time and have opinion on that.

Speaker A: So stay curious, not judgmental.

Speaker B: Exactly. Stay curious and don't be afraid because I've noticed that some people are afraid of AI and they tend to stay away from it because like, oh, it's scary to take our job. So uh, I don't like it and I will not use it. Don't do that. That's for sure.

Speaker A: Okay, that is uh, I think actually pretty good advice because that is true that um, the emerging tech is overwhelming. It's like every week there's a new release and um, if your Twitter feed looks anything like mine, then at least once a week you see a tweet with oh, these are the AI tools to use for this and these are the tools to use for that. I could totally do like my own movie at this point with all those tools probably if I check them out.

Speaker C: Natasha, I have the similar feeling but every new research paper that is coming up on the Twitter, so I have such a long to do, to do uh, to read list with all of the papers that they have to read because it's something new. And uh, every, every, every new day I have something that was uh, obsolete and something that is better than the previous, previous solution.

Speaker A: And yeah, that is, that is so tiring though I will, I will admit that in my.

Speaker C: Yes, I say you always have a feeling that you are behind. So this is very intimidating. And I would say my recommendation

Speaker A: would

Speaker C: uh, be that you don't have to follow all of the news, you don't have to follow all of the things. Just try to keep the direction and try to play with the AI, Try to learn and understand the AI, Maybe do some kind of curse on the Coursera Udemy or other uh, other platform in order to a little bit understand it and try to learn how the prompting looks look like and how the prompting could help and um, yeah, could help you. So it is I would say quite general, general recommendation for all of the developers and also for other mental workers, I would say.

Speaker A: See you don't have to follow everything. You just need to follow Tech Leaders Hub. Um, simple, very simple, um, to double

Speaker D: down on this, uh, I think a wise thing to do uh, in the current situation is to uh, take a moment to find uh, uh, few of your favorite uh, thought leaders or uh, people who are uh, doing a deep dive and summing up uh, the things that they uh, discovered and dev Leaders cup is one of candidates for that kind uh, of role. So uh, definitely uh, you should look for more uh, of the leaders.

Speaker A: And my final advice to you guys, all the listeners and watchers at home is Also go to pythontechrater.snext.com Download Python Tech Radar, read it and uh, I'm sure you will feel way more prepared for what's ahead. And this brings us basically to the end. Uh, so thank you very much for joining me today. It was a pleasure. I have to admit that uh, those sessions where uh, we get to have more guests and more people on the show are really fun because it's more like a conversation between all of us and I really enjoyed that. So thank you for being um, my guests today. Uh, and thank you to all the people who took the time out of their day to listen to us and to ask us questions. Uh, and this is it for today. Bye.

Speaker B: Thank you. Bye.

Speaker C: Goodbye.