AI Is Coming For Your Software Engineering Job (Or Is It?)

Transcript is auto-generated and may contain errors.

so I tried doing a cold start video it was going pretty good but then I wanted to be safe and when I was putting the camera back I decided I was going to close the screen but I didn't realize that by closing the screen it was going to stop the recording so I don't edit these I'm not going to do it but I'll give you a little rev this is uh the wife's SCV it's an X5 m550i sounds pretty good but that's all we got to get topic today we're going to talk about AI everyone loves talking about AI we're going to talk about Ai and software engineers and if any of us have a future in this career or if we're doomed um I feel like I've talked about this before but I might have written a newsletter article on it and then done

a live stream and not actually done a code commute episode so we'll do that traffic says it's going to be an hour kill me um this is the fourth day at the office this week I was just talking to my skip level and I said I'm not coming tomorrow just half kidding but it's a lot of a lot of travel um I've had a pretty busy week so be nice to not have to commute in so um I don't want to kick this thing off I guess I should remind people if you have topics you want me to go over comment below let me know we'll go over software engineering stuff career stuff and um if you want it to be anonymous find me on social media it's Dev leader and we'll address that I want to see if I can get a a a

gurgle sound of the exhaust cuz oh it's so good one sec one sec one sec can you hear that I don't know I guess we'll find out after um it's probably too windy this thing's got a tiny little they call it a dead cat I don't know why it's a pretty terrible name but maybe it was just wind maybe you couldn't hear the gurgles and burbles but yeah send stuff anonymously or send it tell me you want it to be anonymous just message me on a social media platform if you're choosing or comment below if you don't care if other people can read it okay AI um all our jobs are doomed no such thing as programmers ever again um yeah I'll be straight up I think it's kind of um because I think that it's going to take first of all I think at

some point in the future if we achieve some type of superior AI then sure it'll start replacing all sorts of jobs I don't think it's going to be software Engineers specifically um so I'm not I'm not concerned I'm not discouraging anyone from getting into software engineering I think it's still got lots of potential uh so there's the spoiler you can end the video now um we're done here but to be serious uh I I genuinely don't think that our careers are going to be replaced and stuff um I do think that our careers and what we focus on is going to change 100% I think that that's going to be the case because I think it's like any other uh sort of large advancement in technology I really feel like I've talked about this before anyway so my my belief here is like okay let's

use some of the examples that we're dealing with and kind of extrapolate okay so like when we're working with AI today we have ai like we can ask a question we can get it to Design Systems we can get it to convert code we can get it to do all sorts of stuff which is awesome okay but um the challenge is that it's really hard to just blindly trust the output of it and that means that like if you have ai that's like sort of just doing all the work for you um it's going to be wrong a lot which which right now is okay as long as you're not solely relying on AI doing all your work for you right it's like it's helping at least what in my use cases and people I talk to that feel like they can use it effectively

it's helpful but it's not a replacement okay so what the argument becomes is like yes but it's only going to get better and it can already replace like you know it can already replace Junior devs and can already do really advanced stuff and I just think that the the reality of that going to be that it's going to start replacing cuz it already has it's going to start replacing stuff that we as developers consider to be like mundane now I don't mean simple because it can do complex stuff but when I say mundane I mean like either because it's simple or where it's repetitive or it's like uh it could even be complicated but um your time would be better spent doing something else this thing is a beast man so much fun um it's actually a lot of fun to drive this SUV after

driving that X1 for too long um this is real good so my thought process around the AI stuff is like as a de like as someone who like I guess I should back up I'm not writing software every day at work but I use it I program heavily outside of work every single day so uh I do use AI tools uh I find that they're a huge benefit um so some examples of what I use them for are certainly just like AI code complete right so like using C- pilot and visual studio getting code complete coming up um autocomplete what am I call it code complete I don't know what I'm saying anymore um like the the fact that I can start typing some patterns and co-pilot recognizes like if it knows or is pattern matching to figure out like I'm trying to build like

a factory or I'm trying to build a repository like it it's really good at getting a lot of the patterns that I use regularly um put down I got to this traffic looks horrific it's already backed up we got to move over not a good sign that's for sure um so I like heavily rely on that I've talked about this part for sure before that um when my co-pilot was down briefly for a day or something or a session it felt like I didn't it it almost felt like I didn't know how to program anymore because I was so used to having it um complete really big sections of code for me or at least get them started now when I say that it's doing that I still have to heavily modify what it's doing but it's like it gets the shape of the code

that I want it's a funny thing to say like the shape of the code but truly it's going to get like for a lot of what I'm doing like 70 to 90% And I just got to go touch it up but but I also know what I'm doing because I've been doing it for a long time so I know that when it's autocompleting stuff for me I got to go read it I got to go make sure it's actually what I want but this is one type of thing it's like the autocomplete style of things now I can also and I do I'll go use go to use chat GPT I'll have it um write SQL queries and stuff for me um a lot of the time it's like I'm designing schemas for the like a database and I can I know based on my

design that it's sort of denormalized in a way that I want and like I'm happy with the structure and the organization of things it could it be better obviously all the time but I can go focus on that and then just say like hey like chat GPT I need like a query that's going to pull information from these tables here's their schema and I need a result that looks like this oh this car has hand gestures so I can't speak Italian in here apparently because it picks up on my hand gestures let's not do that um so I'll use it for doing that kind of stuff so the SQL queries here's an example right when I'm getting it to do SQL queries for me like arguably those are complicated that's a complex thing that it's doing because sometimes it's doing stuff where I'm like man

like I'm glad I didn't have to think about that so it's not that it's only for mundane simple things but truly it's like it's mundane to me because I'm like I'm not that's not a creative thing that I want to spend time on okay like I don't want to spend time trying to think about optimizing SQL queries and this kind of stuff like it's not a good use of my time I would rather be you know figuring out other things to be working on depending on what I'm building but almost certainly I could tell you that if given time constraints sitting there optimizing SQL queries is not up there with uh things I want to spend my Creative Energy on so that's one type of thing regular expressions are another really common one uh and then the other thing is like just refactoring code sometimes

uh sometimes I'll have massaged a bunch of stuff over time and I'm like man like this code's pretty messy but like functionally it works and I have tests and stuff on it and I might just say like hey like go refactor it to um and then I'll give it like some type of stylistic things that I'm after and then have it go do that so if you're a net developer an example is like uh I had code where it's using a lot of link and it's kind of messy and it's actually like a a hot path in the code and I will say like hey like this method we actually know the collection sizes like just go initialize these things like you don't need to use Link to go over it uh and other times it's the opposite where I'll say like hey like this

is I know I'm iterating stuff uh you know too many times like go clean this up collapse it down make it more readable blah blah blah like I know how to do it I just don't want to go spend the time doing it I got to move over Lanes here and this is going to be if you watched videos from yesterday it's the same same as yesterday this this spot's terrible on the highway uh and if this spot's bad it means where I'm going is going to be even worse so not going to be fun um so well again you're probably watching this going yeah we know where it's at today but like it's it's obviously everyone keeps saying a few years like we're toast um I still think no I think what we're going to see in a few years is that and and

maybe even before that like the amount that we can trust the output continues to improve come on buddy get in here no I I would love to turn this camera around and just like Show you people trying to zipper merge because people don't understand how it works and it's hilarious and painful this person has two car lengths open in front of me and they're getting impatient and trying to creep up and all they're doing is shortening the amount of room that they have to zipper merge in it's ridiculous and look I just merged instantly easy um okay so I think we're going to see that we can start trusting AI output a lot more over the next little while because the accuracy will go up um and we'll have more like just naturally from using it more and more we as people prompting it will

have better experience like crafting better prompts um but I don't think what's going away is that you're like we're still going to need someone to prompt the AI right um there still needs to be some type of like input to this system and there needs to be a feedback mechanism uh people will say okay well there's AI agents that can orchestrate things together um sure I mean you could say that you put together an AI team and you told It Go build a million dooll business and it did awesome okay does that mean that it replaced people like did anyone just lose their job from that it's kind of weird right it's like not just because it could it did something that a person could do does not necessarily mean that someone's job is now lost it's a kind of a weird thing to say

um so I'm not I'm not thinking hey a job's going to be replaced I'm thinking jobs are going to be enhanced significantly right like as I said I'm already heavily relying on AI for I was calling it code complete but autocomplete it's funny when you're talking to yourself in a car with a camera on you um you say stupid stuff but hey um so I'm not concerned about that I I think it's going to be enhancement because that's what I'm observing um I think where it gets dangerous for people and when I say dangerous I mean in terms of like your career being at risk has nothing to do with AI it has nothing to do with AI and everything to do with these two other ingredients and those ingredients are technological change so that's one ingredient and the other ingredient is being complacent so

if you are complacent in the work you do and we introduce technological change you should be afraid of losing your job and that's it it just so happens that in this case the technological change we're talking about is AI and it seems very much like magic right it's a very fascinating thing it's very powerful but that's technological change and then the people that are at risk of losing their job are the people that are complacent now what does it mean when I say complacent and I mean that people who are number one unwilling to even look at AI as a solution for anything right like oh that's dumb it's a fat or oh it's always wrong or I won't for what make up a reason to not use it right so you're basically just letting this thing that's gaining more and more momentum that's not

going away You're Just refusing to acknowledge its existence or using it to me that's a very dangerous spot to be in the second part to that complacency is about not wanting to change anything that you're doing I want to give you a completely different example um has nothing to do with AI and I don't want to say this to trigger people and have them sort of take this the wrong way but I think it's a fair example okay so um in software engineering there are depending where you work there are roles like QA or tester this kind of thing um now let me add a disclaimer here that I have worked with some amazing testers so what I'm about to say before someone gets gets all Angry is bound to happen is I'm not about to say that I don't think testers are useful okay

I'm not about to say that but what I have seen is that working in teams with testers that refuse to acknowledge or invest time into automation okay now again if you're listening this and this is getting you triggered already I'm not saying that there is no value in a manual test of some sort manual validation a manual test however you would like to Define it there are certainly things that doing manually can add benefit to did I say that right that doing things manually will add a benefit from doing anyway words are hard it's Thursday um so I'm not saying that what I am saying though is that if you ignore the fact that you can automate parts of the manual steps that you are doing I would say you should be afraid of losing your job and I mean that because like if you

say no automation's bad or no I don't want to learn how to automate things I'm only ever going to manually test stuff I would say you are narrowing your F or you're narrowing the space that you can be effective and you will eventually become obsolete because everything around automation is advancing that's not slowing down so if you say I don't want to do that I don't want to learn it or this is how I've always done it you're like I'm not going to move past manual testing well uh yeah I'm never going to automate things then I would say expect that over time like you're going to you're going to be replaced so you should be focusing more and more on Automation and that gap of where it's valuable to manually step through things it will still exist but that Gap will shrink in terms

of like where again more people that can't zipper merge here we go there's still two lanes for like 300 M stop doing this it's just you make things worse for everyone when you do this like seriously zipper like I'm I'm genuinely curious if you're watching this when I say zipper merging if you're like I have no idea what the hell you're talking about like please let me know cuz I feel like sometimes I'm taking crazy pills when I watch people do like just fail to understand zipper merging there's two lanes use them and then when it's one lane you should not be trying to cut people off you should form a zipper one car and then the next car and then the next one it's not rocket surgery I don't understand okay so what does this mean for people that are trying to get into

the industry right did they miss do they miss out like is software development done because by the time they were able to start Landing jobs there's going to be no such thing as programmers again I think the answer is no I think there's still plenty of opportunity see this person figured out the zipper merge they said oh I see that the lane has ended and you are trying to get in I will wait it makes perfect sense so for people that are focusing on becoming software developers again if you're like hey I'm just not going to look at AI um and so you're if you're getting advice from other developers and they're like telling you to not look at AI I would say that's dangerous just because they didn't have ai when they were learning uh I think that that's like a dangerous piece of

advice now when I'm giving advice to software developers up and coming I do suggest they use AI I think it be can be a great tool but I think it's something that you need to think about using properly and I've talked about this recently I can't everything's blurring together so I don't know if it's on code commute or Dev leader or both um but when I talked about this concept um it must have been on code commute I I was suggesting that like one of the things that's dangerous is when you use Ai and you're not thinking about the um about what it's giving you come on let me in here see we can do that in this thing we can't do that in the X1 courtesy vehicle so if you're just getting output from an llm to like to speed up your work and

then not learning from it or not questioning it um I think you're missing out on a huge amount of learning as a junior so I would say like don't not use an llm but I think you want to use it wisely right I think that's important ask it questions use it to unblock yourself we got to get moved over here we got to move over a bunch of lanes I'm behind a stupid truck is wck everything come on buddy move up move up move up Let Me In yeah yeah yeah let me in there's got to be a huge accident somewhere here cuz traffic is backed up on both sides of the road um so I actually I do think that people that are junior that are getting into the industry and learning using AI tools uh I think that they're going to be set

up for a ton of success personally I think it's the exact opposite of what people are saying where it's like oh there's going to be no programmers like there's no such thing as software developers I think there's going to be so many things related to this field still it's just that it's going to look different for example something that could be fascinating is like do programming languages look anything like what they do today right are you writing code in a traditional sense maybe after 5 10 years maybe not for new code but like are you do you seriously think that in 5 years there's going to be no C++ code that anyone has to work with and that just like AI is just going to be dealing with all of this on its own like I just don't think that that's realistic I think that's

a a bit of a pipe dream personally so thank you Google for letting me know there's a stalled vehicle I had no idea based on the flow of traffic I feel like it's a bus that's stalled across all of the lanes God okay let's got to get over one more Lane I am going to I'm not going to make this is going to be a super long drive home so I'll probably end this code commute partway just so that you're not like falling asleep cuz I realize I'm I'm not talking as much as I want to because this traffic is so bad and uh I'm just trying to focus and do the right thing here so um I apologize for that so I want to give you some more concrete thoughts um so people that are up and coming I think by leveraging Ai and

kind of being part of that as the AI space is also growing and booming I I genuinely think that they're going to have number one more experience and familiarity with tools because if you think about developers that have been in the space for 5 10 15 20 plus years the longer you're in the space the more likelihood that you're kind of set in your ways okay now not everyone and it's really good quality if you're willing to always learn right doesn't matter if you've been doing something for 30 years you're like I'm an expert at this I've been doing it for forever if someone's able to come around and say yeah but have you thought about doing things this other way keeping an open mind can be a very powerful thing right because you can continue to learn and improve so I think that for

juniors that are getting introduced while this AI boom is happening they get to ride this wave of like crazy new things that are super powerful super convenient uh super productivity tools right I think there's a huge opportunity there now I want to call out one of the risks a little bit more clearly I think and I've already hinted at it but I want to I want to kind of go down this little thought experiment here so I talked about this in the past one of the risks is that and even earlier in this video but like one of the risks is like when we're using AI instead of in this uh space where we're getting stuck on problems right we're using AI to kind of just bridge the gap we can ask it and say just do this for us right if we if we

do this so much that we don't have to do as much critical thinking and we don't have to understand how things work to me that seems like it's one of the one of the biggest risks of just relying on a gu okay but what I think is very interesting about this and I would be curious your perspective on it I I feel like this isn't going to be a popular opinion I don't even know if I believe this but it's like a thought that I'm entertaining and that's like a lot of us that are software developers now and been in the industry for a while like we learned particular ways right so just to give you a more concrete example I went to University I went for computer engineering so there's a more of a hardware Focus I got to learn about like I had

to have classes with oscilloscopes and learn about measuring uh you know frequency and voltage and and dealing with circuits and then in later classes I got to deal with like how um you know registers work and like like CPU architecture but those classes weren't like Cutting Edge CPU architecture it was like here's like some of the most primitive you know thoughts around computer AR or like CPU architecture just so you could have like an understanding of some of the basics right but these are all like the building blocks that they introduced and then it was like and you're set free so what I find really bizarre about that is like my formal education felt like this really like old school way of looking at things even when I was going through it so I was like okay I know you're trying to teach me the

fundamentals but like man like it would be really cool to go take these courses and feel like I know something that's being actively used okay so I go through this process and then I get into the industry I'm learning by practicing things you keep hearing people say build projects I'm constantly building building stuff um people will say like learn these different types of languages so you know I'm I'm learning in a pretty traditional sense I would say and I feel like people that you know are around you know same level of experience as me probably did a very similar thing um people that were before Me Maybe sort of like Oh look The the previous generation of of Learners before me um I bet you they were spending a lot more time and I don't have stats on this but this is the thought experiment

I bet you they were spending a lot more time doing significantly low level lower level stuff like yes I've worked with pointers in C and C++ and unfortunately Objective C I've done this stuff but like I bet you a lot of people have this philosophy where it's like you should start with a programming language to learn about pointers because this is going to be how you learn things effectively you need to be doing this this is how it's done and I feel like I was already riding the end of that where I'm like ah like I should probably understand it but like I'm not going to waste a ton of time trying to like actively program with pointers because like why why would I based on the programming languages I'm using actively I just don't need to I get the concept I'm not going to

practice it a ton because like why so to extend this thought experiment what I'm genuinely curious about is even the advice I give like hey you should be asking the llm to explain the details right ask it to explain the options make sure you understand the details is it possible that we will see that people don't do that and they continue to not do that and it actually ends up just all owing them to be more effective because they can operate at a higher level so like this next wave of software developers coming through they're like I literally don't understand like how a pointer works and You' be like what this is insanity like how do you not know a pointer and you've been a software developer for four years and you can't explain what a pointer is to me that sounds ridiculous to a

lot of us I'm sure but like is it possible that they just literally will not have to know just like I can tell you pretty confidently that I bet you a lot of people don't really understand how like information is put onto the stack right like and having stack frames and stuff I bet you there's tons of developers that don't understand this thing because I've never had to or how like registers work why would they have to know that yet for a lot of us these were Concepts like we were told you need to know this to progress so I'm genuinely very curious if we'll start to see this more and more oh my God horn didn't work this person just swerved out in front of me I had to slam on the brakes you probably uh observed that I drive around a lot of

really stupid people um the stupid people everywhere unfortunately that's okay we just got to keep our distance so I'm try I guess what I'm trying to say is like as I'm kind of building up even this thought experiment is like I think there's still a ton of room for us to be building with AI supplementing our work with AI letting AI take over what we would consider to be mundane parts of our jobs and we will still have people that we call software developers or software Engineers I just think it will look very different and I think it looks very different based on what we're spending our time doing essentially so conceptually we're still part of building software but I might not have to go write particular things anymore just like I was giving you the SQL query example I'm at a point with AI

or I don't have to spend much time at all writing SQL queries and I use Frameworks like omm Frameworks one called Dapper where I need the SQL query in my code I actually like having the SQL query in my code code I might be weird but I like having it I like being able to look at it and understand it know what it's doing because if I need to tune it and tweak it I would like to be able to do so but I don't have to write them anymore and when I say I don't have to write them you might say well there's Entity framework cor Nick like you never had to write it anyway but my point is like I like having it there and I still don't have to draft it myself which is really neat like it's a neat advancement so

um yeah I think that there's a lot of room for people as software developers still with AI do I predict at some point in the future that software developers will be replaced by AI um sure I think that there's a point like that in the future but I think at that point it's because there's a a massive like societal shift where it's not just software developers it's basically like you could effectively replace most most roles most things that people do with AI and that's kind of like why I'm not really fearful right it's I have like a similar philosophy not to do a huge tangent here I'm very good at huge tangents um so when I was in um still living in Canada and like I lived near Toronto area and there's a bit of a like a housing housing price bubble and people kept

saying like oh housing prices like they're going to crash they're going to crash they're going to crash uh housing is it's not going to be worth anything like this is such a bubble and the the reality is it's like I think that if housing were to come down to a point where you're like oh I lost money on this right like it's never going to be recoverable I would say that will happen when there's way bigger problems and the last thing on your mind is going to be oh housing housing crashed it's everything crashed everything is just like wrecked so if it were to crash I don't think it's going to be something that does not recover in which case just keep holding it don't sell it when it's dropped because if it's actually going to crash and not recover it's because the whole world

is in shambles in which case bigger concerns similar thing with AI here I don't think that we're all about to lose our jobs uh and by the time AI is Advanced enough where it's something that could replace our jobs it's because it's replacing or able to replace almost all jobs I just don't think we're anywhere close to that just because AI can do something that a human does or close to does not mean that jobs get like are all replaced right it means you can start to introduce them more you can start to leverage them more but uh trying to make people fearful that it's all like everyone's jobs are just gone I think we're so far away from that um yeah it's an interesting thought like what happens what does it look like when that happens right when everyone's jobs can be replaced by

AI what do we do are we going to like do we just say okay well food food's free now for everyone you can live for free and we'll just have ai do all the work for us cuz like what you have you don't have to pay the AI right and like you also I don't know like you got to live but you don't have an income because you can't no one can have a job because we've all been replaced by AI like it's just a weird thought I think when we reach that point we got we got bigger concerns or like Society shift in such a a way that I can't really fathom so um I guess to wrap up thoughts on this because I still have another 30 minutes of driving in this traffic this is I thought yesterday was the worst today is

going to be far worse cuz I'm in the fast lane crawling here it's like a parking lot I think if you're interested in becoming a software developer there's still lots of room um I like I feel comfortable saying that I am eager I'm excited to see where AI is going uh I'm excited for it to replace the parts of work I don't want to be doing this bus that's coming up on the left is going to try and zip merge in front of all of us and I'm going to lose my mind because once that bus is in front of us it's going to be going below the speed limit here we go see this thing I hate it so dumb it's one of the only things that could make this ride worse there's a stupid bus in front of us okay folks well I'm

just going to be pissed off for the rest of this drive now so I'm going to wrap it up here um sorry that was probably a little bit more disorganized uh even more so than usual uh but it's good I'll be back in the car for videos going forward again so we'll be back in style which is nice um yeah thanks for listening I'll see you next time take care