https://www.youtube.com/watch?v=RNwuFcQ6VmY

无作者

Chances are that when you watch this video, you are a software engineer,maybe a mobile engineer if watching my videos for a longer time already. And I, am just noticing quite heavy trends, quite heavy changes and transitions on, the current job market, on the global industry in regards to software, engineering, in regards to AI integration. And I think it’s just very,very important to talk about that to see what really happens on the market right, now so that you can start adapting on time. Specifically, I see three big, pillars that are changing here. On the one end, of course, AI integration in, companies worldwide. So, what kinds of AI skills companies demand, how that is, changing, so everything around agentic development and all and all that stuff.Then, the second big pillar is about the non-AI tech skills that are required, so, how those are actually changing, which skills are really becoming more in, demand, which roles are less in demand. And lastly, another pillar that I see.

here is just a set of new roles that emerge because AI has in the end opened, up for those. So, those are the three parts that I want to tackle in this, video. Starting off with the AI integration. At this point, of course,it’s no surprise that the increasingly many companies are using and integrating, AI into their workflows, be it for agentic development, be it to automate, some kind of processes or so. AI is a thing, the numbers are rising for for, the adopters, and that’s no surprise because it’s really, really useful. If, you still think AI is just about Slop, then I’m sorry, but the frontier models, nowadays are really, really good. But let’s not just listen to our feeling, here what is happening, but I really also took a look at the data. On the one, hand, I made a survey myself. I’m a big fan of actually having YouTube community, surveys and not just doing them once, but multiple times with a little gap in, between so I can see how trends develop. And I made a survey 7 months ago of when.

I published this video that is all about how much you guys use AI at your, workplace. And the same survey I made a couple of weeks ago again, little, wording changes but all in all the categories are the same. And what this, has shown is that just in these six months in between those surveys, 25%, more people in my community have are working in a company that has fully, adopted AI for agentic development. So, are really writing most of the code just, with agents. And such a change in just half a year, is a crazy pace. Because of course there are lots of medium and bigger companies, where you can’t just snap your finger and make a change, but there is a many, regulations you have to take a look at, many processes you have to accept and, take into account. So, the changes in these bigger companies tend to be very, slow. But this data shows that despite of, these constraints, the changes are made very very fast in regards to AI, adoption. But also also of my own service, the data really shows that uh.

specifically when taking a look at the US job openings and how many of those, openings in the tech industry list some sort of AI skills as a requirement for, the job.

opening

[C] a job that is available 空缺的职位 SYN vacancy
There are several openings in the sales department.
销售部有几个空缺。

Those are 73% of the tech jobs in the US that require, some sort of AI skill as a requirement of all tech jobs. So, this shows how, much in trouble you already are if at this point you still haven’t gotten into, anything regarding agentic development, LLMs, AI because this is not future, stuff anymore, it’s exactly what is happening right now already. And, specifically in regards to agentic development, making agents generate your, code, there are of course various skills that are really needed here, but of, course being able to make agents stick to your conventions, being able to come, up with great skills, great custom agents, on building guardrails around, your agents, verification related skills, integrating agents in a, professional environment, have them create proper pull requests, partial.

review automation, um, orchestrating multiple agents, that all of those are, just very common requirements nowadays in tech jobs listing these AI skills.So, that is the side about AI, and that I think wasn’t a very big surprise that, the industry is changing even more towards that. But, the second pillar, that I’ve mentioned is about those tech skills that aren’t about AI, so our, actual coding skills, Android, iOS, web development, uh, DevOps, system design.What is really happening in that regard? And here it’s very important to say that, we are in a transition phase. So, what I will say in a moment is something that, we already seeing on the market, but it’s also it also means there is another, side of the market that hasn’t yet adapted to that. But, what I’m really, seeing, and not just me, but also many people I work with, is that companies, increasingly demand full-stack roles, or at least this typical T-shaped engineer.So, where you are very good at one specific thing, like native Android, but.

then also have this T-shape on top, where you have very good broad knowledge, in other fields, in related fields possibly. And that’s also what we can,uh, see here on the pragmaticengineer.com,uh, which has made, which is in the end tracked how roles in the tech industry, have changed since 2022. And what this really shows is that roles like a, front-end engineer, an iOS Android engineer, that is getting less and less, in demand if that is your only thing you’re good at, if you are a specialist, in these things. I’m not saying that is worthless nowadays to be a specialist in, Android and iOS, in front-end development, but that alone won’t cut it, anymore.

(not) ˈcut it
(informal) to (not) be as good as is expected or needed (不)如预想的一般好;(不)像所需要的那么好
He won’t cut it as a professional singer.
他的歌艺未达到专业歌手水平。

But, you really need to generalize more. That’s also what my, recent video was about, that I think we really will all have to become, generalists in software development. We have to learn about things like, back-end, about like if you’re into Android, we need to know about iOS,maybe something relating to infrastructure, to DevOps, because AI.

just allows fewer people to do more, and then often times teams simply don’t need, five people anymore, one for native iOS, one for native Android, one for back, end, when one person can actually do all of that. But again, we are in a, transition phase right now. So, on the one hand, I’m seeing this, the roles, that demand these full stack skills are really really increasing.On the other hand, there are a lot of roles where things haven’t changed yet.So, where you still need deep knowledge about Android as a technology, about, compose core routines in that sense, for example. And these interviews still, exist a lot, where you are asked about those things and how deeply you, understand concepts about Android in this case. So, the result of this video, shouldn’t be that you suddenly drop your whole profession and switch into a, completely different field, but just that you realize this is becoming a, thing and you that you start to adapt to this trend ideally right now. Not.

dropping what you’ve always done, that’s still important, that will still be, valuable, but not as a sole thing anymore. These kinds of transitions in, the end are always modelable as a bell curve, where there are some early, adopters on the one hand, then what will probably come in the next 6 to 12, months, where the broad market will also adapt to these, changes, and then there are some latecomers that will adapt very late.And ideally, if you watch this here and you’re still super specialized into one, single role, start learning something else as well. If you’re a mobile, engineer, start learning back end, for example, because you want to be ready, when the general public of the market is also adopting that, when we hit the, middle of the bell curve.

bell curve

a line on a graph that rises to a high round curve in the middle, showing normal distribution 钟形曲线

And while all that maybe sounds a little bit negative,I don’t really think it is. And of course, if you say you are an Android, engineer and an iOS engineer watching this, and this may feel like.

everything you’ve learned is now worthless, because now you’ve spent, possibly multiple years becoming very very good at Android. I spent a a lot of, years becoming very good at Android, you know this.But this is kind of this the sunken cost fallacy. It’s a bias that you think,okay, because I have invested so much time into this, it’s not worth dropping, that, because you already have this big of a time investment. But that is not, true. It’s a fallacy after all. Of course, it doesn’t make sense to keep, going this path if you know it doesn’t make sense to go. But, of course, also, everything you’ve learned or a lot of what you’ve learned is not going to, waste now. Your experience is definitely not going to waste now. The all of the, things that you use for Android like software engineering principles, best, practices, things like dependency injection, architecture, system design,how you test software, all that experience that is just as valuable in.

other domains. So, there is not that much to learn. You Of course, you still, have something to learn when getting into back end when getting into iOS.But, especially a very related field to Android iOS development, what what, really is new there? If you don’t really have to know the programming language in, a way that you have to write the code yourself. In the end, most of what’s new, is just that the operating system is a bit different. There are some other, quirks that we have on Android. These are the kinds of things that you have to, learn. But, system design is the same as on Android. Architecture is very very, similar as on Android. We have the same environment on mobile device. And so,the bar to switch to such a related field is so so much lower than it was a, couple years ago.

Here, bar means the standard or level of difficulty that must be reached to do something.

bar → 门槛
: standard
wants to raise the bar for approving new drugs

And what this is also showing, especially if we take a look at, this graph again, that there are certain roles that are just completely new that.

are created because of AI. And this is interestingly something that history has, always shown that when there was on the one hand the disruption of one area. And, I’m not going to lie, AI is disrupting the software engineering industry. But,whenever that is the case, new jobs, new types of roles have started to emerge.And in this case, that is for example the AI engineer, someone who really, integrates AI into products, into company processes, building rag, systems, rag pipelines for companies, making all the internal company data, available via an interface,

In this context, “rack” might actually be “RAG” — pronounced like “rag,” not “rack.”

RAG = Retrieval-Augmented Generation. It’s an AI technique that lets a model retrieve relevant information from company documents, databases, or other sources before generating an answer.

For example:

  • RAG systems — AI systems that search company knowledge bases to answer questions.
  • RAG pipelines — workflows that retrieve, process, and pass relevant information to an AI model.

via a chat for example, or AI consultants who are, in the end consulting other companies how they can best integrate AI into, their infrastructure. So, those kinds of roles are quite new because AI has, created them. And that has always been like that. And those are very, very, high-paying roles. So, it’s really not the kind of doomsday scenario that many, people want you to believe or the feeling that you maybe get if you’re.

scrolling a lot through LinkedIn and social media what AI is currently, destroying is that all of your experience is going to waste. That’s not, true. It is disrupting the industry. Less people can build more at the same, quality, but human intuition in the end still counts. And what I find is that AI, is still a yes-sayer, a confident yes-sayer. And it won’t help you make, good decisions what you build. AI will help you go in the direction you want to, go faster, but it won’t often at least tell you when the direction is wrong.So, all in all to really draw a summary here, if you are currently a specialist,deeply specialized on native Android, on native iOS, on front-end web,then step one is to drop your identity. Drop the identification with that role.Don’t think anymore I am a native Android developer, I am a native iOS, developer. That is the first thing you have to get, rid of because as long as you see your role, your profession as that role, you.

will kind of want to stick to that, which will be harmful in the long run.And the long run here is really a couple of years at most. But those who do drop, their identity and just start learning more, start adapting to where the market, is changing right now already, those will be the winners of the future. And, those who stick with what they’ve always learned about and think that going even, deeper into that is going to cut it in future, that alone at least, they don’t, have that much time left anymore. But luckily, which path you take is, completely your call, your decision. So, it’s completely your influence. But I, think we can also have a very constructive discussion about that in, the comments. What are you currently seeing on the market? What is changing?How did maybe interviews change for you specifically? What is your profession?So, all these kinds of information I would really be happy to hear about to, see not just what’s the data showing, but also what’s your experience. How do.

you feel about all that? So, let me know that down in the comments. Thanks for, watching. I will see you back in the next video. Have an amazing rest of your, week. Bye-bye. >> [music], [music] [music].