Latest / DevQuestions with Tim Corey / 311. Your Software Development Career is in Danger

Transcript

311. Your Software Development Career is in Danger · DevQuestions with Tim Corey

• 0:00
  SPEAKER_00: I've been talking to a lot of developers recently, and a common theme is that they're either using AI to do a development at work, or they're being told that AI must be the one to do all the actual coding.
• 0:12
  SPEAKER_00: This is putting your career in extreme danger.
• 0:15
  SPEAKER_00: So let's talk about the new reality of software development and how you can not only protect your career, but also set yourself up for long-term success.
• 0:36
  SPEAKER_00: Software development is more than just writing code.
• 0:39
  SPEAKER_00: So let's talk about the rest of it.
• 0:41
  SPEAKER_00: Specifically, let's talk about protecting your career.
• 0:46
  SPEAKER_00: Now we're off the bat, I want to say that if you're tired of hearing about AI, I am too.
• 0:51
  SPEAKER_00: That's why my Monday videos continue to be about software development and mostly not about AI at all.
• 0:58
  SPEAKER_00: But I want to be clear, those right now are my least popular videos.
• 1:03
  SPEAKER_00: People are less and less interested in learning about software development because they're more and more reliant on AI to do their work.
• 1:11
  SPEAKER_00: But here's the problem.
• 1:13
  SPEAKER_00: What happens when the AI isn't around?
• 1:16
  SPEAKER_00: And people tell me all the time, that'll never happen.
• 1:19
  SPEAKER_00: AI is here to stay.
• 1:20
  SPEAKER_00: It's never going away.
• 1:22
  SPEAKER_00: Well, Uber just recently said they used up their entire annual IT token budget in about three months.
• 1:29
  SPEAKER_00: GitHub Copilot announced that it's going to be changing its pricing dramatically starting June 1st.
• 1:36
  SPEAKER_00: Their own estimates are showing that developers will pay anywhere from four to a hundred times more money for the same token spent.
• 1:45
  SPEAKER_00: When AI companies are forced to report their earnings and their losses, they routinely show that they're losing billions each year.
• 1:54
  SPEAKER_00: Every time you see an AI company raise billions of dollars, you need to realize the reason they're doing that is because they aren't profitable.
• 2:03
  SPEAKER_00: Eventually, those billions of dollars will need to come from their customers.
• 2:08
  SPEAKER_00: And that means skyrocketing costs.
• 2:11
  SPEAKER_00: Companies are now seeing AI bills that cost more than hiring employees.
• 2:16
  SPEAKER_00: And those costs are only going up.
• 2:19
  SPEAKER_00: Economy of scale isn't going to solve this problem.
• 2:23
  SPEAKER_00: So what happens when your company can't afford to pay for AI?
• 2:29
  SPEAKER_00: What happens when you no longer have access to unlimited tokens?
• 2:33
  SPEAKER_00: Local models aren't good enough or fast enough to replace frontier models.
• 2:38
  SPEAKER_00: Even worse, it's been shown that even frontier models are really bad at long-term development.
• 2:44
  SPEAKER_00: At some point, companies are going to have to start relying on developers to be their local AIs.
• 2:51
  SPEAKER_00: What happens on that day?
• 2:52
  SPEAKER_00: Will you're skilled and ready, or will they be atrophied?
• 2:57
  SPEAKER_00: Now, I spent at one point in my career six years as an IT director.
• 3:02
  SPEAKER_00: And in that time, I was kind of doing a bit of everything.
• 3:07
  SPEAKER_00: I was doing networking, I was doing server maintenance, I was doing email server and SQL Server and all the rest of doing basically everything for a small college.
• 3:15
  SPEAKER_00: And for about a year of that time, I spent six years as an IT director.
• 3:20
  SPEAKER_00: For about a year of that time, I didn't do any software development.
• 3:24
  SPEAKER_00: I just didn't have time.
• 3:25
  SPEAKER_00: I was too busy doing everything else and learning how all this stuff worked.
• 3:30
  SPEAKER_00: The problem was that during that time, my skills really atrophied in software development.
• 3:36
  SPEAKER_00: When I came back to software development and I made an intentional choice, because that's my first love, to come back and start integrating software development into my job, I really had a hard time just getting started.
• 3:48
  SPEAKER_00: The idea of even a file new project was a bit intimidating because I had missed things, things had changed, and I had forgotten how to do the things that just came natural just a year ago.
• 4:01
  SPEAKER_00: But years worth of time was enough to kind of drain some of that muscle memory, atrophy those skills, and I had to work hard to get back up to speed.
• 4:12
  SPEAKER_00: Now I did, and you can too if that happens to you.
• 4:16
  SPEAKER_00: But that's not something you want to go through because it is a very difficult process.
• 4:21
  SPEAKER_00: And I didn't have to change jobs during that time.
• 4:24
  SPEAKER_00: I didn't have to get a new job.
• 4:26
  SPEAKER_00: I didn't have to prove that I was a good software developer during that time.
• 4:29
  SPEAKER_00: I was an IT director.
• 4:31
  SPEAKER_00: So I was able to build back up my skills while still being employed and while not being required to be the best software developer I could be.
• 4:42
  SPEAKER_00: But that might not be true for you.
• 4:44
  SPEAKER_00: There's gonna come a time when your development skills are going to be needed again.
• 4:49
  SPEAKER_00: Quite frankly, they're needed now, but it isn't quite as obvious.
• 4:53
  SPEAKER_00: Now, when it is obvious, when companies say, hey, we need you to rate code without an AI or not all the time AI, are you gonna be ready or are you gonna be out of shape?
• 5:06
  SPEAKER_00: If your company says, hey, you can use the AI to create some stuff, but you can't let it do everything, you've got to do some of that work so to save us money.
• 5:15
  SPEAKER_00: Can you do that or is that gonna just cause a mess?
• 5:19
  SPEAKER_00: So here are some things you can do right now to be ready that will also help you in your job today.
• 5:25
  SPEAKER_00: And to be clear, this you know, the idea that your company might not be fully reliant on AI might not happen tomorrow.
• 5:32
  SPEAKER_00: It might be six months from now.
• 5:35
  SPEAKER_00: But every month you go without being in software development and practicing those skills and building those skills up, it's getting harder.
• 5:43
  SPEAKER_00: But also, when it comes time where companies say, we need a software developer, if you can step up, you'll be a leg up on everybody else.
• 5:52
  SPEAKER_00: So here's some things you can do right now to help your job today, but also help you for that day when your company says, hey, we need a developer.
• 6:01
  SPEAKER_00: Number one, write as much code as you can by hand.
• 6:04
  SPEAKER_00: And I know that sounds archaic.
• 6:06
  SPEAKER_00: You know, software development is all about AI now.
• 6:09
  SPEAKER_00: Write code by hand.
• 6:11
  SPEAKER_00: The things you will learn will definitely improve your skills.
• 6:17
  SPEAKER_00: This is the same reason why when you, you know, when I was in college, we had laptops, we had the ability to just type all our notes.
• 6:26
  SPEAKER_00: But people who are older than me encouraged me, and it was a good advice, take a notepad, an actual notepad with a pen paper.
• 6:34
  SPEAKER_00: Write down your notes that way, and then put in the computer if you want to.
• 6:37
  SPEAKER_00: But the the very act of writing those notes helped change and put those things in my brain in a way that just typing things out didn't.
• 6:46
  SPEAKER_00: So writing code puts things in your brain in a way that just copying and pasting, even, doesn't.
• 6:54
  SPEAKER_00: Now, your employer might not want you to touch code at all.
• 6:58
  SPEAKER_00: That's understandable.
• 6:59
  SPEAKER_00: Well, it's not.
• 7:00
  SPEAKER_00: But that's I understand that happens, but just because it does happen doesn't mean you can't write code by hand.
• 7:07
  SPEAKER_00: It just might mean that you don't write code at work by hand.
• 7:11
  SPEAKER_00: Now, number two, focus on.
• 7:14
  SPEAKER_00: Here's some things focusing on whether you're writing code yourself or having AI write it, focus on security, performance, and edge cases.
• 7:23
  SPEAKER_00: These are three things that AI does not do well that you can take a step up in, learn more, and help the AI out, and at the same time help your skills out and help your company.
• 7:38
  SPEAKER_00: So when you're learning, focus on security, performance, and edge cases.
• 7:43
  SPEAKER_00: When you're reviewing code, focus on security, performance, and edge cases.
• 7:48
  SPEAKER_00: You'll find a lot of value in doing that.
• 7:52
  SPEAKER_00: You will discover the things that the AI has done wrong, that you can kind of be the hero and make things work better by saying, hey, that's a performance problem.
• 8:01
  SPEAKER_00: Hey, that's getting a security hole, because there's been numerous cases where software guards production that just isn't secure, performant and doesn't cover edge cases.
• 8:12
  SPEAKER_00: We've seen this from the very AI companies that are leading the charge on frontier models, where they put things in their code that are unsecure, that are underperformant, that don't take into account edge cases.
• 8:27
  SPEAKER_00: So it's not like this is just a you're holding it wrong situation.
• 8:32
  SPEAKER_00: This is something that happens throughout the AI industry when people rely on AI to write code.
• 8:38
  SPEAKER_00: So learn to be better.
• 8:40
  SPEAKER_00: Learn to be better than the AI in these areas.
• 8:43
  SPEAKER_00: Number three, build something in your free time for fun without AI.
• 8:48
  SPEAKER_00: So too often people think that software development is only about 40 hours a week.
• 8:54
  SPEAKER_00: You go to work, you put your 40 in, and you leave, and you never do anything outside of work.
• 8:59
  SPEAKER_00: I'm sorry if you feel like that's what this field is, but really, if you want to succeed and get ahead, it's gonna be more than that.
• 9:06
  SPEAKER_00: And I know that people don't want to do that.
• 9:09
  SPEAKER_00: They want it to be a 40-hour week job.
• 9:11
  SPEAKER_00: But here's the reality: software developers in the US, the median software development salary means the middle salary, not the average, the middle salary.
• 9:20
  SPEAKER_00: That means that the the people working at Microsoft and Google and all the rest are not throwing off this curve.
• 9:27
  SPEAKER_00: The middle salary for software developers is about double of the median salary for quote-unquote normal jobs or every job in the US.
• 9:39
  SPEAKER_00: We get paid double.
• 9:41
  SPEAKER_00: So with that comes extra requirements.
• 9:46
  SPEAKER_00: We get paid more than nurses, more than some doctors, we get paid more than a lot of different areas, lawyers potentially, sometimes, um, where even they've had to spend years and years at school and racked up hundreds of thousands of dollars in debt from school, we can kind of roll in, be software developers, and get a salary that's that's almost or equal or even a little better sometimes than those fields.
• 10:13
  SPEAKER_00: And yet we expect to only work 40 hours a week when none of those fields work 40 hours a week.
• 10:18
  SPEAKER_00: They work a lot more.
• 10:20
  SPEAKER_00: So, yeah, spend a little bit of your free time.
• 10:22
  SPEAKER_00: The thing is, if if software development is not just a job for you, it's also a passion, it's a lot easier.
• 10:29
  SPEAKER_00: Um, you don't have to, you can do it as a job.
• 10:31
  SPEAKER_00: But if you enjoy it, then do something for fun.
• 10:35
  SPEAKER_00: Pick something fun and build it.
• 10:37
  SPEAKER_00: This is why I did the game development master course, because it's fun to build games.
• 10:43
  SPEAKER_00: It's fun to try things out, and you can do so with the same C sharp skills or similar C sharp skills that you have in your day job, which will mean you'll be practicing C sharp, you'll be practicing your development skills and logic skills and all the rest while building something fun.
• 11:01
  SPEAKER_00: And so you can have a little hobby, you can have a little side project that also builds up your skills and improves what you can do and protects your career long term.
• 11:10
  SPEAKER_00: Number four, keep up to date with modern techniques.
• 11:15
  SPEAKER_00: AIs are not good at keeping up to date with modern techniques.
• 11:20
  SPEAKER_00: Yes, if you push them, kind of, maybe sometimes.
• 11:23
  SPEAKER_00: I did a video where I asked an AI to um use a new async zipping in C sharp from.NET 10.
• 11:33
  SPEAKER_00: And I did the video, I think like a month after.NET 10 came out.
• 11:37
  SPEAKER_00: Now it'd been in pre-release for a year, but um it came out in November.
• 11:42
  SPEAKER_00: I did the video, I believe, in December.
• 11:45
  SPEAKER_00: And I couldn't get to use the.NET 10 libraries, even when I pushed it and pushed it and pushed it.
• 11:50
  SPEAKER_00: And people told me, well, it doesn't know that.NET 10's out yet, which is true.
• 11:55
  SPEAKER_00: It didn't.
• 11:55
  SPEAKER_00: If you asked it, is.NET 10 out, it would say, no, not yet.
• 11:59
  SPEAKER_00: It's in it's in preview, which is not true.
• 12:02
  SPEAKER_00: It was actually in production.
• 12:03
  SPEAKER_00: And you keep pushing it, it's like, oh yeah, yeah, no, it's out.
• 12:06
  SPEAKER_00: But I still wouldn't know how to use that thing.
• 12:08
  SPEAKER_00: It would cite sources on how to use the async zip library that you know were out of date or were incorrect.
• 12:16
  SPEAKER_00: So how can you use the latest technologies and techniques?
• 12:21
  SPEAKER_00: You need to know about them.
• 12:22
  SPEAKER_00: You need to know what's out there, what's changed, and how to push the AI to use the right modern techniques.
• 12:31
  SPEAKER_00: Because the older techniques will often work because of backward compatibility, but they won't be as secure.
• 12:36
  SPEAKER_00: They won't be as performant, they won't be as reliable.
• 12:39
  SPEAKER_00: And remember back to uh point number two, talking about focusing on security, performance, and edge cases.
• 12:46
  SPEAKER_00: Well, this is part of that.
• 12:48
  SPEAKER_00: Keeping up to date with modern techniques allows you to make sure that you are using the best tools.
• 12:56
  SPEAKER_00: When you don't use the latest version, often you expose your company to security holes and vulnerabilities.
• 13:03
  SPEAKER_00: Now, you gotta be careful because also latest version sometimes can be injected.
• 13:07
  SPEAKER_00: We've seen a lot of that, especially the NPM, but a lot of that kind of thing going around the internet today where even the latest version is the wrong version to use.
• 13:15
  SPEAKER_00: You have to keep up to date and know what is the modern thing to do, but also how to protect yourself against these security threats.
• 13:26
  SPEAKER_00: So those are four things you can do today that will help you in your job today, but will also help prepare you for the day when AI might not be available in the same way it is right now.
• 13:39
  SPEAKER_00: You might have to rely more on just being a good software developer.
• 13:43
  SPEAKER_00: Your career matters.
• 13:45
  SPEAKER_00: Don't get lazy and let the use of AI hurt you.
• 13:48
  SPEAKER_00: Instead, use it as an opportunity to set yourself up for long-term success.
• 13:54
  SPEAKER_00: All right, thanks for listening.
• 13:56
  SPEAKER_00: As always, I am Tim Corey.
• 13:58
  SPEAKER_01: Thank you for joining us for this episode of Dev Questions.
• 14:01
  SPEAKER_01: When you're ready to learn to think and code like a professional developer, head over to IamtimCorey.com and enroll in a course.