Burned Out But Still Coding? Here's Why That Matters

Transcript is auto-generated and may contain errors.

Hey folks, I was just reading one of the threads for experienced devs. Figured I'd talk a little bit about that and then jump over to just some brief AI updates from over the weekend. The thread on Reddit was someone asking about how people end up finding time uh or like maybe they kind of frame it like how they find uh mental energy in the evening to be able to continue to work on things that are that are technical or complex um because they're saying by the time the evening comes around their brain feels fried which is kind of interesting. So, the post itself is pretty short, so there's not a ton of context into maybe what they're what they're after.

But I think just figured I'd talk about this cuz uh it's kind of funny because I actually like I have both sides of this where there are days where uh I feel fried and then other days where even though I'm doing what I think are similar things or working as hard or working on complex things um that I still find uh that I get very creative in the evening. So I I should start by saying I'm definitely not like not a doctor. I don't understand all the details about how your brain works. Um certainly not this is just like anecdote I guess but my understanding for some of this at least is like for my own I'm going to say like I'm going to use the word research but I don't mean like conducting studies. I mean from for me like surfing the internet and looking for information on it.

Uh I've been over the past couple years trying to shift my my sort of sleep schedule and wake schedule for uh being a morning person. And traditionally I'm absolutely not. I've never been my entire life and I don't really have a desire to be except just how it I think needs to work with the schedule that I have with my wife for we go to CrossFit in the morning and uh otherwise like this our our days don't overlap. So um I've been kind of doing this shift over the past couple years and one of the most tricky thing you're probably like what are you talking about Nick? Um one of the tricky things though is that I am like a night owl. That's I don't know why, but usually between like 10:00 and 11 at night historically, that's where like all of a sudden something turns on for me and I feel creative.

I feel um I feel like I need to do. So like I go into like a creative build mode and it's just kind of always been like that. And what ends up happening is that I have this burst of I don't know this like creativity or productivity for uh for like mental type things, right? So like problem solving um that kind of stuff and then it doesn't keep me up like it's like I can't fall asleep. Usually what happens is like I get to a point where I'm tired and that would historically be around like 2:00 in the morning and then uh I would lay down in bed and I I put it's going to sound kind of weird but like I would put myself to sleep by by continuing to think about hard problems. And uh that's just how it's been for a really long time.

Like I don't know as long as I can kind of remember. I don't know what it was like before doing that. Maybe I just never paid attention to my sleep uh before that. Anyway, point being that in terms of having like energy or like mental energy like this person's asking for, I I genuinely wonder if for some people it's kind of uh you know, just how they are. And just to briefly tie it back to when I said I was kind of doing a little bit of homework or trying to look this stuff up. Um, sure there's a lot of people that are saying, of course, like, you know, circadian rhythm or like uh, you know, waking or rising with the waking up with the sunrise kind of thing. And there's supposed to be some evidence backing this. There's also people, uh, that say there's evidence to say the opposite for some individuals, right?

Like um, you you just may not be wired this way. And I don't know the the threshold or the balance between figuring out if you've given it a good of enough shot cuz for me it's been almost 2 years and I I feel like now I can wake up at you know 500 5:30 in the morning and it's okay and that's only been I would say the past couple of months otherwise it's been a really long time. So, I don't know if it's a matter of like you didn't try hard enough or just kind of really not how you are, but I've heard some people saying it only takes like I don't know like less than two weeks of like trying to do it consistently and I'm like, dude, no, it's not like that. But creativity, productivity for me is an evening time thing. Uh I I don't feel that way when I wake up at all.

It's a very different mindset. I don't wake up and I feel productive. I feel energized. Um it's the same thing like after I go to CrossFit. I don't go to like some people are like, "Oh, I work out in the morning because it gets me energized for the day." I'm like, "No, I go to CrossFit. I'm going to be getting there. They shifted our classes, which is kind of nice." So, I get there starts at 6:30 a.m. Used to be 6:00 in the morning on the dot. Get there, work out for an hour. Um, for that like almost 2-year period I was just telling you about, most days I would go home and sleep after like I don't I don't feel energized after dying at CrossFit. I don't feel energized for the day. No, not for me, right? So, I think people are just different and that's okay.

So, I don't feel creative and productive in the morning. I do feel that way at night. And for me, I've just been slowly trying to shift to being a morning person. Now, where it gets kind of different, cuz I said I get a little bit of both sides of this, is that I find um I find like peopling, cuz I'm very introverted. I find people all day is uh very taxing. So, for example, um my Thursdays are a lot of one-on- ones that I have with remote employees and uh I try to organize my my calendar such that I don't have uh cuz well, I have a bunch of direct reports. I don't have my 101 spread out across every single day because uh because it is mentally and I don't know like emotion maybe emotional is not maybe the right word but I'm going to just stick to mentally taxing for me as an introvert.

So I I do schedule those on particular days and do them kind of in bulk. I got screwed here. What's going on? We got stuck behind a really big truck that was going way below the speed limit and then we couldn't merge. Um, so on days where I have a lot of peopleing, so whether that's meetings, one-on- ones, and that kind of stuff, I feel exhausted at the end of the day. Um, but it's a it's a different kind of thing, right? It's like uh I think those days are where I either do nap or I feel like I need to nap. I haven't really needed a nap in a few months now. Um but those are the days I really feel it after work where I feel like the mental energy is not there. But almost always it's just that I need a quick nap to like reset my brain and then I'm fine.

So, I personally don't really have a uh like brain fry from thinking about hard problems. I'm not saying that never happens. Sorry, but that's not like my my regular uh type of daily thing or weekly thing. I think it's the types of problems that I'm thinking about. And this kind of ties back to some of the the other conversations from over the past week or so. in the videos about engagement and stuff like that. I think if I'm trying to solve difficult problems that I'm not super interested in, that's where I start to feel more of that brain fry. And uh I think to me that's kind of interesting because it's like it's almost not your brain is fried from thinking about hard problems. It's like you're for me it's like my brain is fried from thinking about stuff that I'm not trying to put energy into things that I'm truly just not engaged in.

And uh again, not that's not consistent. I'm not saying like I oh I never get tired of, you know, working on my own things. Like that's not the case. I I certainly do. But for the most part, I find that it's the types of problems like whether or not I'm engaged in them uh and less about the complexity because there are some things that are for me complex problems and they're actually exciting to go solve or they're exciting to spend time thinking about. Um, there are times where I'm working through things that are hard problems for me to solve and it might be like couldn't get it for over the weekend or something and just making this up as an example and then Monday morning like as soon as like my alarm goes off or whatever, I'm excited cuz I want to go run to my computer to go try something, right?

Like it's an excitement thing. And that almost like I find for me is generally um at odds with this like mentally burnt out from working on technical things. Um so I I don't know if this person was asking this question because they were interested in like you know how are people moonlighting or working on side projects and stuff like that. They don't really offer a ton of context. Maybe they responded to people in the comments and I didn't read it. But um if that's the case for them, I just wonder if like the stuff they're doing during the day is is genuinely not engaging and they're spending a ton of mental effort on it and as a result like it's just like they they feel fried and less about the specific technical challenges. I don't really know. Um I also don't know for them if they have something that is outside of work where it's an interesting engaging problem space for them to go to go work on.

Uh and the final thing I'll kind of mention about that is like given that I don't know the context they have. Uh you know it's it's not to say that you need to go have this kind of stuff outside of work, right? I'm I'm not I'm certainly not sitting here saying you have to go, you know, have your your moonlighting side hustle kind of thing. Um, and like you better go do that, bro. But b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b based on how they were asking, that was kind of my interpretation is like how do people have mental capacity for this kind of thing into the evening. Um so for me it's just genuinely a lot more about projects I find exciting.

uh not every day, but projects that ideally get that to me most of the time. Anyway, just a light one for today. I was kind of curious when I read that title. Uh figured just do the last little bit about uh sort of an AI update. Speaking of doing things after hours, uh this weekend was a bit hit and miss for feeling productive. Um I I've been kind of sticking to uh the similar pattern I've been doing where I have a few things going on at the same time. I if I step back and think about or look at the progress that's made, I think there is actually a lot of um productivity, but I think that this weekend I felt more frustrated by um AI doing kind of like dumb things more than I wanted it to.

for just to give you an example um because I was frustrated last weekend with some some skill usage from AI uh where it was apparently using skills and then I would like get the result and I'm like wait a second like you clearly clearly didn't do this step and it would be like oh yeah I decided not to. Um, I wrote a or I had it write a skill that's basically like explains another skill like a workflow. So, what I've been experimenting with is before I get into a session where I have um a big like orchestration kind of skill to go do a bunch of work, I'll ask the model run this other skill which explains the orchestration skill I want you to use. and like given the context of the problem we're trying to solve. And so what it will do is like write out a workflow that's almost like a waterfall kind of thing.

And then that tells me or I don't know I I don't know if it's a guarantee or not. Probably not that uh it's aligning like the skill flow to the problem at hand. And now that's also in the context. And my theory has been that if I do that, number one, I can catch if it's going to do something stupid early on because it explains to me what it's about to do. And then number two, if that's in the context and we've agreed on it, then it will stick to it. And that's given me a little bit more confidence for some of this. Um, but I still had some cases like uh last night I was thinking or I'm thinking about last night where uh I was asking the sonnet model uh to go you know do this kind of thing and it gave me the results and it was like basically 100% fail and I'm like I know that I just know that statistic is not true for the work that it was doing.

And so I questioned it. I said like, you know, what did you use for it to be 100% fail? And it's like, well, I used this regular expression. I said, okay, like did you get that regular expression from the skills we discussed? And it said, no, it said I and then it wrote out, I know that we talked about using these skills. I see two of these skills say exactly how to go do this matching. They there are two skills that literally have the exact same thing for how to match. And then and then I decided to go write my own regular expression instead. And like so what what do I do with that? It's like I gave you instructions and you just decided to go do something else. Um so it's kind of tricky because you know I'm looking for the feedback from it to be like okay well tell me logically what what got you to that point.

and it wasn't really able to say, which is kind of the frustrating part. And sometimes when you ask it that it like tries to defend itself, like, "Oh, you're right. Like, I'm so stupid. I should have." Um, and I'm like, I have to tell it, "Don't defend yourself. Just explain like the logical progression for like how you ended up selecting this." And, uh, it just didn't really have a good answer. So, I've I've just been noticing at least this weekend I had a what felt like a few instances like this where just kind of doing stupid stuff which is annoying because I go run these big workflows um and it's you know churning through a lot of stuff seems to be making progress and then I second guess myself like can I trust any of that?

Um, it's easy when it's uh verifiable in front of your face, but otherwise I'm not uh totally sure like if it's a records in a database like do I have to go read through all these database records like that kind of stuff. Bit of a pain in the butt. Um, I did have a good win uh with another scenario where um I was getting frustrated with a development loop with C-Pilot uh where it would build things for me and say, "Oh, it's built and tested and now it's like and so like ready for you to go try it out." And uh this is for my side project where I'm revi coding this like role- playinging game engine that I've had for like 20 years and I'm just kind I'm just doing it on the side because it's kind of fun to see it come together.

And so started integrating Unity and what was happening was that Copilot would say like okay the features done you go try it in Unity and then I would switch over to Unity and like the code doesn't even compile. So having to keep pushing back on this LLM to be like what do we need to do here so that this doesn't happen again right um did you write tests for it? Okay, it's saying it's writing tests, but like the code's still not compiling. And then a lot of the time it's like, oh, we didn't like copy over the new binaries. So, honestly, probably like 20 different things to to try and close this gap. And it's mostly there, but dramatic improvements. And one of the big ones was that uh doing based on the types of changes like so like file paths and stuff that are touched doing strategic pre-commit uh test runs.

So what would happen is that when it's ready to hand off work to me it would say like hey I've committed this you know we have the test the features done go try it and then it wouldn't work. So now what it's doing is at least running the test because it can't finish because the commit hook is literally forcing it to run. And that has I don't have like stats to to prove it because I I just wasn't keeping track. But like anecdotally, this went from happening like probably like two to four times per feature being done where it would just get stuck like, "Okay, I'm done. Oh, now I fixed it. I'm so sorry." Um where I keep going back to Unity to reload things.

And so like just to verify a quick feature, it's now like 20 minutes of work back and forth, which is super dumb to like it's uh it's more often completing the feature tested and like the feature I'm not saying the feature is necessarily done the right way, but like I can see the change and it's working, which is exactly what I would want. So or more in the direction of what I want. So things like we were doing movement and I wanted to be able to use WD and click to move and like it did it and I was able to use WD and it was fluid and click to move did something but it was not uh registering the input always correctly. But at least like it was hooked up and I could use it. So then I could give it the feedback and tell it to go hey like this doesn't work.

You need to go try it in this scenario. blah blah blah. Oh man, no parking spots. Okay, that's dumb. But yeah, closing that loop was just a real pain in the ass because like no matter what I was telling it with like types of tests and stuff, it was just like just missing stuff. And I'm not using CI/CD because I'm just doing it on my my local machine. So, we needed some type of uh feedback loop to work with here. Oh, there's tons of spots right here. I drove past all of them. So, that was a big thing. But, yeah, kind of agents doing dumb stuff. And then, uh, getting pre-commit hooks was a big one. And then last night, switched back over to Claude because I was hitting co-pilot rate limits for the first time uh, that I've ever seen. So, time to go back to Claude because not going to sit there being rate limited when I got to do.

So, thanks for watching. I will see you in the next video.