About this talk
This talk discusses the impact of AI on software development and content creation, emphasizing the identity crisis faced by educators and developers. The speaker highlights how reliance on AI has led to a significant decline in engagement and learning opportunities, particularly in communities like Stack Overflow. He addresses the necessity of bringing back human connections and understanding in coding practices, advocating for developers to embrace the learning struggle rather than seeking instant solutions from AI. The session further explores how educators can adapt their methods to better nurture junior developers, suggesting that mentorship and collaboration can create a healthier environment for growth amidst the rise of AI tools. Ultimately, the talk underscores that while AI is a valuable resource, the human element remains irreplaceable in technology and learning.
Full transcript
So, good morning everyone. I hope everybody is having a good day. I mean Thank you for sh- Uh, I hope everyone is having a good day. It's really great to be here. Barcelona is one of my favorite cities, honestly. I I came here last Friday. It was like great. I spent a lot of time with my family here, so I feel like already 10 years younger. And
to start off, as you see, the title is different than the one that we had before. Don't worry, in case you want to see that talk, it is the last talk today. Today, after like starting off with technical content, I also wanted to give you some let's say hopeful talk on what we can do to get better in the stage that we are uh, with this talk
called Who Teaches the Operators When the Educators Stop. And why do I think it's a very important topic? Because if you look at YouTube, you will see the content creators slowly making stopping making videos about uh, software development. You will see that, for example, people like Maximilian Schwarzmüller and like many other people actually lowering down the amount of content that they put out there, the technical content
and everything that they have. There are many reasons to this, but the biggest reason is this is from a friend of mine who is who has been a big part of, for example, Flutter community, has been doing content about it for 3 years. It was his main resource for income. And every year, at the end of the year, he explains what's going on and what he did,
what worked and what not. And this was his 2025 message here. He mentioned that the amount of views that he had has been lowered by 60% almost. And with that, he had, of course, less people visiting, less people buying his courses, and so on. So, basically, pivoted. He pivoted into showcasing how AI tools can be used to build software. Which is fine. Uh so this talk is
not about how AI is bad or anything because AI is part of our lives now. But today it's important to remind ourselves like how developers still matter. Because we come from the different time and period. For example, how many of you consider themselves as junior? Like raise your hand. I see one or two. Like okay, couple of hands nice. How many consider yourselves as medium level people?
Okay, a lot more. And rest, do you think you are senior people? Nice. Good. Good. Good. Good. We have like majority of seniors, but we also have a good balance between medium and junior people as well. And for folks who are new to the business, AI wasn't was not a reality that we had. Like actually we had this magical website called Stack Overflow for a long time
period that was giving us answers. And people were bullying other developers if they don't know the answers by shouting the to their faces like write uh read the documentation, do this, and do that. But actually it was the best place for us. We were still going and copying and pasting code and learning a lot from it. this graph here is very well known because it showcases how
AI agents changed the way that we build drastically. The moment that ChatGPT launched the amount of questions that has been asked on Stack Overflow has been lowered more than half. And in 2026 I've seen the number is almost around 10 to 20,000 on all Stack Overflow sub versions as well, which is very low compared to where we were only a couple of years ago. the problem is
bigger than that, of course. The reason that we are seeing less and less content is the biggest reason is the identity crisis. Many people who are building content, they're thinking, "Why am I wasting my time? Why am I doing this when AI can just do it in 3 seconds?" Which is fine and fair because when you think about it, it is actually a problem that everybody is
facing as well. We don't ask questions, we just go to the AI and ask about it and hope that it will get the answer right or we believe that it will give us the right answer. The other part is impostor syndrome. Especially if you're writing in an in a language as you can see from my accent, English is not my first language. when you speak and read
and write in other languages, already you have to express yourself properly. We did a lot of trainings. Like over the years, I did many trainings, I've been reading books, But even sometimes, I mean, right now I can see an AI text immediately. But right now it feels like AI is still doing better than most people and they believe that as well. That's why they even copy-paste the
content they created through that and put it out there. And the feedback loop is lost. Like people go and ask the chat and whatever you do, it is actually allowing you to move away from the feedback loop that you might have uh from the people. And the feedback is actually Feedback is the bliss. Feedback is our happiness. It keeps us moving. And we don't Once we don't
have it, it will slowly turn the content into the weirdness that we are in. And last thing is AI slop. This is the favorite word of many people, I would say. AI slop means for folks who don't know it, basically the low-quality, repetitive content that AI creates thinking that it is normal. And it takes a lot of attention because AI thinks it cracked the system, the algorithm,
which is right. And you will see very weird things. One example that I can give is in Instagram, many Turkish influencers, they put the bio of a famous uh author because it is the best way to get the SEO and the like algorithm to hook into the system. So, you will see random information about like very sad incidents and underneath there's an bio port from Turkey. It's
because like AI slope is bringing people to this low-quality content because it cracked the algorithm somehow. And it brings a lot of noise as well because we are in the sense that oh, what is using AI? Everyone is doing this craziness and so on. But actually reality is a bit different. By the way, every number that I see and so on, I will share the slides, it's
taken from like research and so on. the perception right now from the Anthropic's research in 2026 showcased that like 25% of the people think they are faster when they use AI tools, which is fine, which is fair. But the reality is, even though they are fast, because of lower comprehension and not being able to debug and test things, people actually get 19% slower than usual. They think
they are faster, but reality is they actually slower than this already brings us back to the first point. We have been optimizing for speed, but accidentally we removed ourselves from understanding what we are doing. And I'm not a big meme person, so I will use one meme and this is the one. And this feels like it at this point. It We feel like everything is fine, but
it is it is not. So, today we will talk about how we can first educate ourselves, educate the others, and how we can actually feel that this AI environment that we have is a tool that everyone can feel safe about. So, first I will start with how can we do the educate ourselves because I think it was Goethe once said, "If everyone sweeps their in front of
their door, the whole of the world will be clean." So, first we need to educate ourselves, and if everyone educates themselves, I think it would be easier to adopt the many things as well. The first suggestion is let's bring back the human element. Let's bring back the human connections and connectivity that we have, because currently what we have is like search results are polluted when we want
to go and search something. Tutorials are outdated by day. Yesterday's model today's garbage, and actually spoiled by the amount of content that is out there. And AI is giving you answers without understanding and thinking it understands it as well, which is crazy to me, because we believe AI more than we believe in human beings. And also many code out there are sh- getting shipped without knowing how
it works. And nobody tells you what you don't know at the end of the day, because you think the answer you get is the right answer. we need to change. First, we need to embrace the struggle. It's okay to have hard times to build new technologies. That is normal. Because in case you want to learn a new language, for example, you don't just download the information, or
you don't think a random information out there is the right information. No. If you want to learn, you struggle. You go out there, you start speaking, and so on. In case I wanted to learn Spanish what would it be the best way? Just going to a course in Germany, or coming here and just trying to speak for 3 months here? Of course, being here would like be
harder for me, but I would learn definitely more. And read the source code. The source code that other people wrote. That is why the importance of software open source software comes in here as well. With AI, that is one of the gates that we keep on, thankfully to the maintainers and so on. And it allows you to have still proper code to learn from. So, not just
focus on the dogs. Dogs are created right now to follow the agents. Read the source And build from scratch with AI. It's okay to build a simple today but once in a while to remember what's going on. Because it's harder to remember sometimes even create in a simple product application without AI. I know like I I know it is for me when I don't do it. I
am guessing it's for you, too. If you say, [clears throat] "No, I'm still perfect." then yeah, you are lying. I don't believe you. But I would definitely say that build something from scratch for sure. And verify what AI tells you. Don't believe whatever it is. I will give you one example. I was teaching at the university last year last semester. And one of my students, she called
me at like 1:00 a.m. in the night from the application of the university. And she was crying. I said like, "What's going on?" She apparently created an instance in AWS that was constantly costing her money. And in 2 hours it was $600, and she was crying. And she was like, "I can't pay this. It's too much money." and so on. And I said, "But I taught you
this. Why didn't you follow along?" She said, "No, I asked AI and AI asked me to do this thing. And now I can't turn it off." If you don't verify, you will get bitten by it at the end. So, always verify, make sure. And find humans. Ask them. Like be vulnerable. It's okay to not know things. I know right now it feels like everybody knows everything, but
I guarantee you people with knowledge is lower than you think of. My take book for you is like first of all, if you have a project, break things on purpose. If AI added an library, delete it. See how it works and see what it breaks. Another thing is again, read other people's code. It's better to learn from the others and ask them. Go and send them a
message. And once you learn something, write it down. Teach it to your others. Teach it to your peers. Teach it to your like your colleagues internally. It will help you a lot to solidify because you will get great questions as well. Also, one more time, make sure you have a habit of building without AI. So, in Anthropic is down, Cloud is removed or something, you know what
to do. Uh you are trained for it as well. And create a group of people, maybe they can help you out how to build as well. It's okay to you know, like have this tiny safe space as well in case you don't feel comfortable in public. And most importantly, you're here, you already made a big step to be with the community. Use the hallway track. Hallway track
is basically when everybody's out there, talk to the people. If you reach If you go and reach out to the people, that shows already you are moving in the right direction. But if you are shy, that's also fine. But if you do not take advantage of the conferences to connect and learn from the others, you will be in your silo and you would think this constant echo
of wrong information is actually the reality. But reality is the thing that you will hear from the So, what can educators do? First of all, what is educator? Educators are many things. For example, I work with AWS as a developer advocate. Our role is working with the service teams and like being with the community, doing talks, creating content, and so on. But for me, educators are different.
Educators are everyone who is knowledgeable and able to share. We have teachers, we have like many people, we have like uh tech leads and so on that is sharing information. So, the way that you share publicly is very important. So, first of all, as educators, we need to understand the way create content we the way that we share knowledge has changed. We We are not just in
the driver's seat anymore. We have always a co-pilot with us, not in the AI co-pilot way, but actually a human being next to us. They are in active participation. They want to see you struggle. They want to see your failures as well because they are failing. I don't know like if anyone has the magic trick to make everything work at once, but it is not going to
work like any tool that you would use is not going to work at once. If they are working, then that means there is a performance issue, there is a security issue, there is something wrong for So, make sure while you are talking to the people, while you are creating content, you are actually showing the failures and like fixing the debugging mentality more than whatever you you are
showing. And more importantly, show real-life fun projects. People are done with the to-do applications. People are done with whatever it is, like dashboards and so For example, there is World Cup going on right now. Sorry for Spain yesterday. I'm from Turkey, it's horrible for me, too. But, but like do something with World Cup. Do something with like other sports, other anything at all, like music festivals and
so on. Whatever interests you, do something about it. For example, today we like my colleagues will have a talk about Dungeons and Dragons because they love playing tabletop games. It's a personal, it is more fun, and people can relate to And what content actually causes is like first of all, people like debugging walkthroughs, like how a problem can be fixed. It is actually very important. And fundamentals
matter than most. Like many people are building applications without knowing what it means behind the scenes. And fundamentals like even talking about HTTP, memory, caching, and so on is actually it matters a lot. Another thing is hands-on labs. People want to build. People want to move forward with proper guidance, and they have highly like 50% of it, if they are coming to your product from let's say
hands-on hands-on labs, they will stick with it because they will be already used to it, and they learn they will learn the basics and more with it. And of course, decision explainers. We are not like once we create content, once we create whatever it is, can be blog, talk, videos, whatever it is, open-source projects, if you just use it to show your product, it's not going to
work. You need to make sure why you made certain decisions. You need to earn their trust now through your decisions because earning trust is your identity now. It is not a under the algorithms, it's not under the views that you have. Everything is synthetic nowadays. But the decision that you make, if they can earn the trust of the audience, then you will be golden. Take a project
from open source for example, tear down, share your knowledge with the others, and see and share what you have seen in this project, right or wrong as well. That helps a lot. It doesn't matter your level. Just go through it and say like as a junior, I don't understand why they did it this way. And of course, code review, code review, code review, it's very important to
understand the reviewing process. We need to understand like AI is a great tool. It helps you to build things, but building things doesn't mean they are going to work. Like according to the MIT study that has been made with like more than 100,000 people this year, the 180% there's a there's an extra 180% code out there that has been written with AI, but only 30% among all
of these chunk of code goes to the production. You know how crazy it is? Like we are not sending the amount of code that we have written to the production because it's just created by AI and we're just prototyping or simply seeing what it is, but it doesn't go to the production. Why? Because it is sometimes easier to build it ourselves because we cannot debug through AI.
We cannot find the solutions. So, we need to understand that with all the content that I have mentioned and we are focusing on debugging, tearing down, and so on because it is where the understanding lives. And that is why we teach it in the content. So, do not let AI assisted coding be a like fancier version of copy-paste. I made up I made up this thing yesterday
evening. I was very proud of it, so that's why I'm reading it. But like do not let it be just a be a placeholder. Understand what's going on. Understand the logic. It will make you stronger. And video still matter because like it's your face and your voice and like our your confusion that makes it real and real content gives empathy. And production quality is just at least
as a consumer for me and I'm okay with the production quality videos and so on. But when I see people like for example Maximilian Schwachman or like other people building live and breaking and like crying it's okay. I feel happy because I know I cried in these as well. you might say that I actually cracked the algorithm but algorithms change. But once you experience something you're there
for good because once you write your opinion that it doesn't matter if it's blog or video and your story is what is making you unique and trust is your new C SEO. You can crack the algorithm like I said but in whatever you have written or did it will be found in 3 years. I still get random upvotes on Stack Overflow for the answer that I have
given maybe eight or nine years ago because it was still relevant to some people and they couldn't find an answer for it on whatever platform they are on. And since we're at the conference let's talk about why conferences matter as The conferences bring a lot of empathy. We need to understand that like humans are emotional creatures whether we like it or not. And once we are happy
together we are happy together. Once we are sad we are sad together as well. So the whole way trick is the place that you see that hey like as a junior you can go and talk to a senior person and say that hey actually like I I was suffering from the same thing. And you can see it doesn't matter some things are very common. And you will
see someone believing in you is the biggest positivity that you that you can see. Many people find jobs on the hallways. Many people find jobs during the networks. Like it helps you a lot. It it takes 1 year to find your next adventure, and that is the most important part in your career. And conferences, they bring this because of the face-to-face approach. And online content is there,
but human it is more needed, and the conferences are the place to get them as well. So, lastly, let's talk about why open source matters. I mean, I don't know if you are a fan of open source or not. I may be open source fan, open source Like, it's not just the products and everything. It's the mentality of like sharing and having a collective understanding of building
things matters. First of all, every PR that you have created or someone has created is a lesson. Every issue is a question. Before we were using Stack Overflow to answer questions and learn. Now, you can go and actually check the issues and you answer the questions about repositories. And every review is mentoring because you will see the maintainers of the products is going to be taking a
look at your answer and get your feedback. And you already have a good feedback and learning opportunity. Reading other people's code, I'm repeating this again, is how you learn what docs can teach because I don't know if you worked in SDKs, products, and so on. Docs hardly is the afterthought. They build the product, then as you move forward, you will see that actually with the customer feedback,
with the developer feedback, they add, "Oh, okay, we're missing this in the doc. AI can do it, blah blah blah." No. Once you read the code, then you get rid of the docs. And now, a a lot of people will tell you to read the docs, and that is still valuable, but you will not get the full picture of the products. And of course, contributing forces to
understand. Many juniors ask me like what can I do to learn? What can I do to make a name for me in the field? And it's open source contribution because first of all, you will get real working experience. Next, you will get actually a a great feedback, and most importantly, you will have a sense of achievement that you yourself achieve I have been saying AI is this,
AI is that. AI is not bad. >> [laughter] >> Let me clarify this, but AI is just a tool. So, once you are working with uh like teaching or like with developers and so on, use AI to generate practices, like problems, like coming up with ideas, or scaffold explanations, or using your source code to create source code snippets to explain some concepts You can create debugging scenario
scenarios and use tools like MCPs and skills and so on to automate it. whatever you do, make sure you keep in mind you are in the driver's seat and you fill everything that needs to be written with your voice. And like afterwards, make sure you don't use AI to replace your voice. I don't want to see your write, I don't want to see streamline, I don't want
to see m dashes. Every time that I see them, every time that when I see a call when I see a talk, a proposal with four paragraphs, I know it's coming from AI. Because I've been doing talks for like it will be 8 years this year, and I have never sat down and written four paragraphs of proposals because if I write four paragraphs of proposal myself, what
is the talk going to look like? It's going to be 120 slides or something. No. Make sure AI is just there to support you, but not shadow your voice. Always keep in mind that developers are drowning in AI-generated content right now that they can't verify. And they need a human being to say this is right place to go, and this is the right place to start off
it. So, make sure you do that. And for all the great seniors in the room, what else I have for you? First of all, mentoring is very, very important. You can start with the outside of the company. You can reach out to the non-senior or like um mid-level developers that you see here. Talk to them. And you have their career changes. You help them to find the
right choices and right uh moves for their career and for their interest. You build trust. And you know that like anything can change. I started my career as a PHP developer. I did Kotlin, Android, Flutter, React, React Native, and now I'm a developer advocate. It's more than 10 years in my career. I did so many things. Every time I learn something new. And at base, I'm a
software engineer, I know. But everything changes. And you need to remind the others that everything can change. It's okay to embrace the change remember, if you are a good mentor, they will remember you for more than a decade. I remember my first mentor 10 years ago. I am still grateful for him to put me to the right direction that I really really enjoy today. You will be
remembered forever this way as well. You might be asking inside, "What should I do?" First of all, leave the why comments and not what comments. Like, do not try to undermine the like juniors or mid-level around you. You are senior. We believe you. I believe you. No joke. I really believe in you. But make sure whatever you are doing is not to show your senior, but actually
show the junior and mid people they are understanding your reasoning. And whatever you do as a change in your pull requests and so on, make sure it's like a small blog post they can read. I'm not saying five paragraphs or something like two or three like mini blog, tiny blog, micro blog, whatever you want to call it. It it works. It You can review review the code
out loud next to the junior in a positive way. And comment your debugging journey so they can see also while reviewing your changes as well. You shouldn't just go and say, "Let me fix it for you." No, this is not the way to go. Because put yourself in their shoes. How would you feel? You would feel like you cannot do anything. But we are there to help
them out. We are there to make them stronger. And make sure future Remember that future is also a junior. There's another junior as well. And they are going to wait for your reports as well. it's okay to say I don't know. This is This was the hardest thing for me when I started in my career. Like normal saying I don't know. This way they will actually have
empathy for you. Share their mistakes. It's okay. Being vulnerable matters. And sometimes explain what the junior did to you. You would actually be surprised how much you will learn from other developers as well because some knowledge we forget to do. Programming is like the languages as well. Like in my day-to-day like I'm originally from Turkey, but I speak English and German more than I speak Turkish. And
sometimes I I forget words in my native language. The reason is I don't use that that much. And it's the same with the code, same with the technologies as well. If you don't use it, they will be gone. And block some learning time for them and for you and like keep for it. Fight for it. And give credit publicly. Once you Once a junior or mid-level person
hears like senior person says positive thing about them, it is the best. I know it from my perspective. Whenever I see it, even right now I'm like as I said more than 10 years in the industry. If someone tells I did a good job, I feel the happiest. So before I wrap up, I want to remind you like at 1:00 a.m. somewhere there's a junior developer is
like stuck with something. And a human voice It doesn't have to be perfect, but an opinionated voice is the thing that they need. And please keep showing up for the others. Please keep whatever you believe in in the technology because right now you need to remind yourself that like the work still matters. You still matter. Your job still matters. AI is just a tool. thank you very
much for listening to me and I hope you have a good day. >> [applause] >> Uh we have a time for one question, so anyone? >> Yeah, one last thing is like here's the the open survey that I created. I do a lot of technical talks. I want to do more soft talks. Please give your feedback it was relevant for you. If you don't like it, that's
also fine. Just I just want to get better at this. Thank you. Anyway, question. >> Hi. Um so I'm a junior developer. and generally I'm I've been seeing what happens in the job market and I've been very lucky that I have a lot of seniors who help me with my job. But uh how do you advocate to have more junior roles and more mid-level jobs to companies?
Because it sounds like sometimes the businesses are not investing in people, but they're investing in tokens more. >> Yeah. I I do it in multiple ways. From my point of view, I I try to work with younger people as much as I can because the voice and the creativity that they bring is different. Also, since we're in tech, tech is used by young people and with the
current average age that we have in tech, we are losing empathy. And we need to have more empathy towards actually the user. So if you have more young people bringing the younger voice and bringing that empathetic side to us, it's actually a win-win. Another side is young people has a lot of energy. They want to build more. They want to learn more and they're very excited as
well. are willing to get the job done more than like sometimes we feel tired, but they don't. And they are like very good at this. And last thing is I really like the fact that they are more open-minded because as you are senior in this software development world, it is hard to find someone who is willing to let's say change their field, change something. But for young
people they're like, "No, I want to learn. I want to like make something of myself." And since they are not putting themselves into a box and they are open for learning, for me to that's another reason. I don't like I don't consider I don't offer juniors to the in my company or any other company as like, "Oh, they are cheaper to hire or something." No. The energy
that they have, the creativity that they have, the empathy that they can bring to the audience that we actually target is very important. If you don't have those people, we are just lying to ourselves. If we act like, "Oh, our target audience is let's say Gen Z and like we our average age is like 36 years old." No, that's not the way to go. There are many
aspects of it of course and most importantly I see it I've seen it with my students. Like I know the job market is hard. I can guarantee it's hard for everyone. So like please don't take it like anyone around you as well as well you can tell them as well. Like job market is hard what I tell my like students or like the people I mentor is
also it's hard for everyone, but make sure you make the right connections, you make you make the right networking and show your real talent because in the software world age is the most deceptive part because you I've met many great creative and knowledgeable people who is way younger than the average engineer that is out there. So it's just a number to me if it matters. Yeah. >>
Okay, we don't have more time. So thank you very much. >> I will be at the AWS booth for [applause] five minutes.
