Episode 63 - Travis Dockter

What if the most useful software in your life isn’t a product, but something you built for yourself in an evening? That’s the spark for this conversation with Travis Dockter, a Rails developer and organizer of Blast Off Rails, where we dig into how AI turns personal ideas into working tools—fast. From a “house health” app that scores chores to a suite of single-user utilities, we break down what’s changed: ideation is quicker, boilerplate is lighter, and the cost of experimentation has never been lower.

We get real about security for personal apps and why network-level access with Tailscale can be the right fit when you’re the only user. It’s a conversation about risk, not dogma—matching controls to stakes and keeping momentum. We also examine the blurry space around AI-assisted pen testing, the difference between “won’t” and “can’t” in model behavior, and how to navigate that responsibly. Then we push forward: what happens when an agent can manage a Markdown knowledge base or a SQLite file directly? If the UI becomes conversation, design becomes orchestration and feedback, not screens.

Docs turn out to be the sleeper blocker. Travis details a pragmatic Obsidian workflow: agents.md files scoped to code areas, linked session notes, and templates that help models find the right context when it counts. We round it out with hard-won lessons on token efficiency, choosing the right model for planning vs building, and experimenting with multi-model “counselors” to balance cost and quality. Finally, we share why a Rails-focused, single-track conference in Albuquerque can actually boost your day-to-day work: tighter content, lower travel costs, and rooms full of people solving the same problems.

If you’ve been itching to ship something small and useful, this is your nudge. Subscribe for more builder-first conversations, share this episode with a friend who loves Rails, and leave a quick review so others can find the show.

Judoscale
Autoscaling that actually works. Take control of your cloud hosting.

Honeybadger
Honeybadger is an application health monitoring tool built by developers for developers.

Judoscale
Autoscaling that actually works. Take control of your cloud hosting.

Disclaimer: This post contains affiliate links. If you make a purchase, I may receive a commission at no extra cost to you.

Transcript

Meet Travis And The Format

SPEAKER_01 0:00

Hello everyone, welcome to another episode of Code and the Coding Coders Who Code It. I'm your host, Drew Bragg, and I'm joined today by Travis Doctor. Travis, anyone who is not familiar with you, would you please do a quick introduction of yourself and how people might know you?

SPEAKER_00 0:14

Yeah, so probably the only way that people would know me thus far is I'm organizing a conference called Blast Off Rails. And kind of started organizing that last year and posting stuff online about it. And other than that, I go to a few conferences now. So you might have seen me there. But I'm a Ruby on Rails developer for the last maybe six years. So excited to be here on Coding Coders and uh see what trouble we can get into.

Personal Software And AI Acceleration

SPEAKER_01 0:48

Yeah, welcome to the show. So for anyone new to the show, the way this is going to work is I'm going to ask Travis three questions. I'm going to ask him what he's working on, what kind of blockers he has, and then ask him to share something cool, new, or interesting that he's recently learned, discovered, built. Doesn't have to be coding related, but it absolutely can be. The name of the show is pretty heavy on the coding aspect. So what are you working on? What do you got going on?

Gamifying House Chores With A Score

Network-Only Access And Tailscale

SPEAKER_00 1:14

Yeah. So this one is going to be interesting because it is kind of enabled by AI, because there's this sort of branch of software that some people are calling personal software, which is kind of all the stuff that you build just for yourself because you can. My version of this was a personal finance app that I actually built originally in the boot camp that I learned to code in. So I built this personal finance app because I had used other personal finance trackers and stuff like that. And none of them really worked quite exactly how I wanted them to. So I was using Excel sheets for the longest time. And then when I learned how to code, I was like, oh, that's the first thing I'm gonna make. I'm gonna turn this Excel sheet into an actual app, make it easier to use. And that was my pre-AI thing. But the only thing that I could really do because I didn't have all the time in the world, was that app. That was the thing I focused on. But now AI has opened up this portal where you can create a lot of stuff. And if you only need something simple, like personal finance tracker or a personal like to-do or calendar project management, I recently made what I'm calling a house health app, and it kind of gamifies cleaning my house because nice otherwise, yeah, otherwise it's like one of those things that I'm like, I have to do that thing. And this like gives you a score, it gives you house a score, and I have like rooms for the house and different tasks for each room, and you have to keep those scores up by doing the tasks, by sweeping and mopping the kitchen, by wiping down the sink and the mirror in the bathroom. And this is stuff that I've always thought about, and I'd be like, ah, that would be cool to make, but never did because I just didn't have the time. But now I can sit down and bang out something like this in an hour, and then you have it. And something also that I kind of want to get your take on, because I haven't heard anybody else talk about this. I'm not a security person. I don't take everything I'm saying with a grain of salt. But my approach to authentication for this suite of personal software that I'm now writing is no auth on the applications, just put it on a server with a tail scale, tail net on it. And I'm the only one that can access it from my devices. I didn't build this for anybody else. Nobody else is ever going to use this. So why go through the pain of you know maintaining authentication and stuff if I can just put it on a private network that only I can access from all my devices and call it good.

Risk, Threat Models, And Practical Security

Pen Testing With AI And Model Safety

SPEAKER_01 4:30

I think the tail scale approach is probably a really good one because there's so many approaches you could take, right? You could pop it on a server and just have the auth sit on the server. You hit the endpoint and like you have to give the password. There's attack vectors for anything, right? Even tail scale, like you're relying on tail scale to be secure. Security, in a lot of ways, you have to do what's right for the situation. You might have two different locks on your house, right? A deadbolt and a door handle. But a properly motivated individual is getting into the door. But if you don't have to worry about a properly motivated individual getting into your house, there's no point in like putting a steel door with multiple reinforced hint, like that doesn't make sense for the level of security you need. You do need a lock on your door. Like it'd be foolish to have a door with no lock. But for most of us, unless you live in like maybe the bad part of town and you need a little something extra, bars on the window, whatever, security is a lot of like risk assessment and stuff. And like it's really easy to be like, oh, I need all of this layers of security. And it's like, but if I'm the only person accessing it, do I really need that? What is the chances of someone giving a shit enough about my personal data to go in there? Like, it's not an account that stores a credit card, it's a game to keep my house clean. You're not even like storing blueprints to your house in it. And even then, like, who wants that? Does it matter? Like, yes, AI is like this super insecure thing because it doesn't know how to do that risk assessment for you. It's a pattern matching tool, so it's pulling patterns off the internet that anybody else can pull off. There is all of this, oh, hey, AI software is insecure. And it's like, well, sure. So is the very first app I ever built in boot camp? It was only even remotely secure because I use device and it did it for me, and I'm trusting that device is doing it right. You know what I mean? But that's app level. Then there's the server itself. Do you have a root password? Is it all SSH? Can only one device get into it? What happens if that device is left at a coffee shop and someone picks it up? It's risk assessment. You have to assess, like, if someone actually got into the server, how fast would I know that they're mining Bitcoin with it and turn it off in wherever the host is or whatever? Or is it a Raspberry Pi that runs in my basement and I can just go unplug the damn thing? That's technically more secure than starting up a Hetner box or a DigitalOcean droplet because I literally can unplug the damn thing. I don't know. But to your point, I am also not a security expert. I've been doing this long enough to have bitten myself in the ass a few times with security-related things, but not a security expert. And AI is a very new level of attack vector.

SPEAKER_00 7:35

Yeah, there is no perfect solution for security. And in my mind, trusting a network-level solution like Tailscale with personal, we'll call it vibe-coded software, makes slightly more sense to me in that what you're saying, like the right tool for the job kind of thing, and the right level of effort and level of security for your situation. Yeah. I had never used it before, actually. I had only heard about Tail Scale, and people, I don't even know what people were using it for when I heard about it. I think it was just people that had like home labs and were doing interesting things with their own Raspberry Pi servers and whatnot. But I only vaguely knew what it was for. And I was like, oh, I wonder if I could just use it for that and not have to put authentication on this application and have to sometimes work around that authentication system, especially when you're working with AI and it's like going down all these rabbit holes of like, oh, we need to add security so that other users can't access your rooms or your tasks. And like, oh, do we need to attach this model to users so that you can access your and I'm like, okay, but nobody else is gonna use this app. Why am I spending tokens on all of this stuff when I really don't need to be? So yeah, I was like, oh, I wonder if I could use it for that. And it's working well so far. I enjoy it.

SPEAKER_01 10:11

Have you at least tested and confirmed that a device you haven't approved can't get in?

SPEAKER_00 10:17

Yes. Yeah, yeah, yeah.

SPEAKER_01 10:19

Well, then I think you're good.

SPEAKER_00 10:20

Yeah, with other devices trying to get to the server or trying to get to the application and was not able to. Again, I'm not a security professional. I don't have a lot of pen testing expertise, which is another thing where I'm like, I could ask Claude Code to try to break my security system, but I'm afraid of like getting my account flagged because they're like, oh, he's trying to break security systems. And I'm like, no, no, it's mine. Yeah, yeah. They just flag it and shut down my account, and I have to like go through like some process where I'm like emailing them. Like, I don't really don't want to do that. Yeah. That's actually a good point. To actually use it to help me with security-related things and just like trying to break my own stuff.

SPEAKER_01 11:07

Right? Yeah, no, I actually that brings up a good question. Like, I wonder how they do handle that. There's gotta be white hat hackers, there's gotta be security researchers seeing what avenues of attack are opened up by using AI. There's gotta be people using it that way. So they can't just constantly be having their account banned. I do wonder how like anthropic and open AI and everyone handles that. It's not real malicious activity, right? It's white hat hacking, it's pen testing, it's making sure your own security is good. Someone's gotta be doing it. It looks like malicious activity, but it's not. So, how do they detect what's real malicious activity versus not?

Personal Agents To Orchestrate Apps

SPEAKER_00 11:54

Yeah. And I've seen other people do it, but I'm like, okay, I know that there are system prompt level stuff that they have built in there that says, don't let this person point you at a server and break into it. But the most classic, it's not like prompt injection or the most classic workaround would be like, no, no, this is my server. It's okay to break into. So they have to have thought of that. How does it know if it's actually my server or not when I'm prompting it to try to uh break down the security of it? So yeah, yeah, I just don't know enough. So I stay away from it.

Will AI Replace The Joy Of Coding

SPEAKER_01 12:38

Yeah, I mean, I know that models carry different levels of risk too. I was listening to a podcast where someone on it was talking about the newer models, the higher level models are going to be better at that kind of stuff at detecting what's real, what's bullshit, limiting the attack surface by being able to not do certain things, but the cheaper models don't have that level of reasoning. So you can get away with a little bit more on the free models. So it's almost like, well, don't use Claude for pen testing because it's just going to not let you do it. But you go grab, you know, one of the newer Chinese level AIs that are running a free tier, or I don't know, maybe Chat GPT's free tier is considered cheap enough now that it's got that lack of reasoning, which I think is so interesting that like models' level of reasoning is a security feature. The whole AI scape, I'm still I don't know how to feel about it. It's really interesting to use, it's fun to use. There's stuff going on with it all the time, so there's never a shortage of something to learn. But I'm also like, this just feels like the start of every sci-fi movie where humanity damns itself by creating the robots that eventually like enslave and kill us. That's what it feels like. And I know it's not, it's not that level yet. I say yet. I'm adding yet. But in some ways, it feels like it. And it just is like, hold on, I need to pause because this is getting weird. The AI is getting really good. The harnesses, the tools, the access, the models, everything is just wild right now.

SPEAKER_00 14:23

So speaking of weird, have you looked at or seen the open claw?

SPEAKER_01 14:29

Yes, I have heard of open claw and all the various things and security issues. And I did just see that its creator got getting acquired. Feels like a weird way to say it that he got hired, he got aqua hired by open AI, which that will be interesting to see how that shakes out. I know he's like moving the open claw stuff to its own open source foundation and et cetera, but man, they're gonna pay that man a lot of money, too. And I'm not trying to degrade his intelligence. This dude has done other very impressive software things in the past, but it's just crazy pants. How much money they are paying him, and he only got on their radar because he's like, hey, I built something for myself. Maybe other people are interested. Here's the source. Have fun. And it went viral. And he's like getting insane job offers. It's wild.

SPEAKER_00 15:23

His was the one that went viral and got the most attention. I think because he is good at what he does and intelligent and built a good implementation of something that you know other people had built. But I watched his interview with Lex Friedman.

SPEAKER_01 15:44

Yeah, that was a beast of a podcast. What was that, like three and a half hour long episode?

Talking To Databases Instead Of Apps

SPEAKER_00 15:48

I'm like, Yeah, it took me like a week and a half to watch it. In the episode, he like alludes that he's like talking with some big company about getting Aqua hired. And the announcement had already come out that OpenAI had hired him like before I was finished with the podcast. So I honestly think part of what made his so successful was that the personality piece of it and like the soul MD and everything that like made it kind of like close to that AGI kind of thing that people are looking for. And his was most interacting with a digital human, and people find that really interesting. So they kind of all glommed onto his project. And it was a you know, it was a very well done uh project, had a lot of interesting pieces to it, and I haven't installed it or anything, but again, that whole personal software thing. I want to build my own very simple implementation of that and hook it up to all these apps that I'm making because that's like the next step is like, okay, I have all these uh personal apps that I've created. Let me hook them all together with this personal agent that can then send me an email this like every morning. Like, why haven't you cleaned the kitchen in three weeks? Kind of give me a little kick in the pants to get going, or you know, just remind me about calendar stuff. That's just like a cool, fun thing that is on the edge. Have you seen that? I think it's an XKCD comic where it's like how much time you spend building a tool versus how much time it saves you and like whether it's worth it or not. None of the stuff that I build makes it on that scale. No.

The Future Of UI And Voice Interfaces

Code Quality Vs AI Readability

SPEAKER_01 17:46

No, no, but but I mean, for anyone who hasn't listened to this podcast, if you have an extra three hours and you're interested in like maybe what the future landscape of AI looks like, it's worth a listen to, or maybe finding a condensed version or something. Like, despite how you may feel about Lex, it is interesting to hear from the guy. I think the biggest takeaway, at least for me, the biggest takeaway from that whole thing was his like quote of like, it's hard to compete with someone having fun, right? Like, he wasn't building a product. He wasn't building even like a tool that will become a SaaS product because I built it to solve my own problems, and now you can pay me to solve yours. Like, he just built a thing because he wanted to, saw it was cool, open sourced it, let other people, and was just like, hey man, I'm having fun. You can't compete with me because I'm not competing with you. I'm just sitting here having fun. I don't give a shit if this is a billion-dollar product or a zero negative on my bank account, you know? Like he doesn't care. He's like, I don't care. Like, I'll use it however I use it. And if you want to use it, go have fun. And I'm like, man, that's the dream, right? Like that you just get to sit here and have fun. High school, the first time I found programming, that was it, man. Like I was having so much fun in the computer lab every day. So much fun learning how to program calculators so I could cheat on math tests, or like, oh, okay, I can add games to my calculator now because I understand this and that. And then I lost it. And when I rediscovered it, when boot camps became a thing, I was just like, oh, right. I forgot how much fun it is to just be like, I can do anything with a computer. Given enough time, I can build it. The knowledge is out there and exists, and I can figure it out over a long enough period of time. And that time with AI has been shortened so much, which is why it's fun. I think there's a lot of room for improvement with it. It's great at spinning up a greenfield app, but the reason why we get paid the money we do is not to spin out greenfield apps, it's to maintain 10, 12, 20, 30-year-old applications that are still somehow running and are behemoths and like database tables with thousands upon thousands of tables. Like AI will get there eventually, but like we will have a place in that for a very long time because we have built monsters of software that AI would struggle to keep in its context window. It'll get there, but yeah, it's yeah, it is fun. It is a lot of fun. It's nice to have that level of like, oh shit, I wish I had a gem to do X. Why am I wishing this? Why not just, okay, open code, pull it up, put it in plan mode. Let's have a 30-minute conversation about it. Cool. Now go build it. I gotta go make dinner. I come back, I spend an hour tweaking its code because it writes shitty code. And I like the way I write code better because I'm a pompous asshole. You know, and now my gem that would have never happened because I'm ADHD and I would have had this brilliant idea and never actually executed on it, now exists. And I had a ton of fun planning it and a ton of fun editing the code, and now I have the gem that I wanted to have. And yeah, personal software is totally going to be a bigger thing. Right now, I think it's still relegated to us. You kind of sort of need to have at least a software engineer's mindset to really be able to empower AI. Like the people who are like, I don't know how to code. I'm like, well, you know, you never learned how to code. You know how to break down problems, you know how to do software engineering, you just don't know how to write code. There's a difference. Your brain does do the thing, and now you're using AI to prompt it into building the thing. You just took a different approach than learning how to write code. Not everybody can do that. Not everybody's going to be able to grab Chat GPT and be like, build me an application, and it works. They're not good at that level of thinking or that type of thinking.

SPEAKER_00 21:49

Yeah. There's more friction, much more friction.

Blocker: Documentation And Memory

SPEAKER_01 21:53

For now, though. Yeah. In the future, I'm sure it'll get to the point where you can say, I need an application to do X, and it'll just be able to pattern match its way through without you needing to iterate very much on it.

SPEAKER_00 22:07

Well, that's the thing that I've people are starting to talk about. Well, no, I think they've been talking about it for a while. I'm just not that plugged in, but I'm just noticing it now. Is like half the personal software that I've written doesn't really need to be an app. Project management to do calendar could just be handled by AI in some simple like markdown files or even just a database. It could just write SQLite queries on a SQLite database, no wrapper around it. Except for the CLI. And I could just talk with it to know what's on my calendar and what's coming up or to add new things or whatever. You don't even need an application to wrap a database now. That's going to be super interesting in the next couple years. People just writing AI wrappers to databases that you can do anything with. You don't even need programming knowledge then, really, because all you're doing is talking with a database.

Obsidian, Session Notes, And Agent Skills

SPEAKER_01 23:20

That's almost the more interesting. I can't wait. I can't wait because I like my job. But like I'm very interested to see how you're just talking with the thing. So why are we even bother typing? Why not just the phone say, hey, I need to know, blah, blah, blah. And it just, I don't need to click a button. I don't need to do a thing. I'm just talking now. It's going to change a lot about UI and UX, right? Like the old ways of doing UI and UX are going to go out the window because we're not interacting with apps the same way anymore. It's interesting. I almost don't want to be interested because I know it's coming at the cost of like this job that I love, right? It takes a piece of it takes a piece of something that I love and was borderline therapeutic for me. Like just being able to headphones on, code because I've figured it out, I've understood the problem, and now I can just vibe code through myself. I can get into the zone and code was great. And someone was paying me for it. And that was awesome. And they still are, right? This isn't a tomorrow or a next year. But in 10 years, my plan used to be retire as a software engineer. Now I'm like, I don't think software engineer is going to be a thing in 10 years. What is like maybe I'm hoping that these are Yeah, yeah. We're all going to be project managers now. No more ICs. They all go away. They become project managers. On the one hand, I'm like, I hate this thing because it's taking something I love away from me. But on the other hand, it's so interesting. And I've got that, like, I'm so curious about it, and I want to know more things about it. And it feels like learning to program all over again. You're just like, wait, I can do that. Wait, how does this work? That felt friction-y. Why can't, like, let's do it this way or that way? Teaching its skills, like growing its ability to do things. It's a tough, weird spot to be in because I want to be like, this is so cool, but fuck this thing.

SPEAKER_00 25:18

Yeah, you're addicted to the drug that's killing you. Right. And the interesting thing is like, I didn't make all of these apps before because I felt like I didn't have the time. But now that AI is a thing, before, outside of work, I would spend probably on average less than an hour, but maybe on weekends a little bit more. Outside of work on a workday, maybe less than an hour on average of coding my own stuff, right? Just because you have life and it's hard. And when I have less than an hour, I'm already tired, like my brain is already tired from work. I'll think of something that I want to do and I'll say, that's gonna take more time than I have. I'll do it later. I just don't have the time or the bandwidth right now. So I wouldn't. And I would average less than an hour per day outside of work, right? With AI, I probably average like two, sometimes three hours a day outside of work just because I don't think about, oh, I don't have enough time to do that. I just start and I'm like, okay, I'm just gonna spend half an hour and doing this thing, and I just start and then it stretches. And if I'd done this before, if I spent as much time coding before as I do now, I would be such a great coder. Yeah.

Context Limits And Structured Prompts

SPEAKER_01 26:50

Yeah. But does that matter? How much does code quality matter anymore? This is something I've been talking about with other people at work, even my boss, my CTO, like, because sometimes I'm like, man, AI really sucks at writing code. It's overly verbose, or like, this is so dumb, or like it doesn't use good idioms or whatever. It like crappy code. And then I stop and I think and I'm like, but I'm also not going to be interacting with this code much anymore. AI is, and if this is what's easier for AI to read, does that matter? It's a weird spot because I do think it matters, because I do think that there's a at least in the near future, like a limit of how much context AI can keep in its window that exists now, but I think that there's a limit of like, all right, we can't get much more without it being like on-prem level of like only enterprise level can have an AI that can understand enterprise big enough app where we're going to have to say, like, you should write good readable code because you will need to interact with it to prompt the AI to interact with it correctly.

Patterns Will Emerge From Tinkering

SPEAKER_00 28:59

Okay. This brings me to the next question of the podcast.

SPEAKER_01 29:03

The blockers. Okay. Yeah, okay. So, what kind of blockers do you have? And if you don't have a current blocker, what's one that you recently had and how did you go about solving it?

SPEAKER_00 29:11

Great. Okay, my blocker, docs, documentation, and project history as part of documentation. We have this constant problem at work where we have a lot of different projects, and we'll like make some decision and we'll say, Oh no, we're gonna go with this tool, or we're gonna go with this implementation or this strategy, and we'll say, Okay, where do we document this? Because maybe it applies to multiple projects, so it's not just one, but it doesn't apply to all projects, and then we'll have the conversation about like where to document it, and then maybe we don't end up documenting it, or we put it in some place that we don't think to look later, and then a couple months later, you come to this same kind of decision and you have the discussion again, and you're like, I know we made this decision in the past, now I don't remember why. And that's been a problem for the entirety of the history of software, right? And now with AI, the friction to writing docs and to documenting things like decisions is very low. You can just ask it, hey, can you look at this code base and write me a README? And it will go and do that. You can have like a coding session with Claude, and after you're done and you've like worked through some problem, you say, Hey, can you take this session and write some like lessons learned from it and put it somewhere in the code base? And it'll do it, right? But now, because the friction is so low, my problem becomes how do you organize this information? And I think a lot of people are working on that right now, but it's still where do I put this? And how do I make it useful in the future to like help yourself and to like give the like whole memory problem with coding agents, make it available to coding agents so that they can say, Oh, remember this decision? We made it for this reason. That's why I'm going to implement this tool in this way, or not use this tool, or whatever it is. Have you guys like had that discussion at your work? Like, are you doing anything for this? Or I'm looking for ideas.

Token Efficiency And Model Selection

Multi‑Model Counselors And Cost Tradeoffs

Conferences That Recharge Developers

SPEAKER_01 31:40

Yeah, yeah, a little bit. We have an agents MD file in basically all of our core Rails directories. So, like in models, we have an agents file in app models, and it only concerns itself with models, right? So the expectation at the high level, the project level agents file, is if you're working with models, load up this file, right? And it only has certain decisions made, like, hey, we have this model, we refer to it as this. Like we have users, we call them creators. We have students, and those are students, but in this next version, we call them members, just so that like when we're talking to the agent, it knows we don't have a great way of AI documentation for certain other decisions yet. We've talked about it a lot. There's been some back and forth. I have a personal strategy for like my personal projects and for my personal work at work is I use obsidian. I used obsidian prior to AI. But what I really think has been cool with AI is I kind of went off, wrote a skill with AI's help on how to use Obsidian's flavored markdown. Obsidian's got a new CLI, which has made things even less friction-y. It can do a lot of like internal linking. So instead of needing to be like reference this file, like it just talks about a decision we made and it names it and it wraps it in these double square brackets, and that tells it to go look for this file, and you can use regular markdown, but it makes these coding session summaries really simple, like one or two paragraphs. Don't need to bog it down with like key lessons learned. It can just say in the session, like we discussed using this pattern or this thing. And when it says we discussed using this pattern, there's a document in my patterns directory that describes the pattern, whether it's hotwire, see hotwire documentation. Or if it's like actually more of an internal pattern, like here's the breakdown. So it works well for me because I was using obsidian before. I'm still getting used to some of the new aspects of it, but it's a bit helpful for AI because it doesn't need to load that up into context all the time. In my personal global agents, I explain how my obsidian directory works so it knows where to look for stuff, but it's not great yet at always saying, Oh, I need to pull up this, or I need to go take a look at that. I do still have to say, like, this project relates to another project where we worked on, pull up that session summary and go. And then it can pull up, like, oh, okay, we're going to use this pattern and this is why we made this decision, or oh, here are the gotchas because of the table naming or whatever issues we ran into last time. That is good, but I still need to kind of prompt it to remember that this has been a thing already. And that's been the biggest challenge for me, at least. Remembering that every time we start an agent session, it knows nothing. It's got access to all the information in the world, but it knows jack shit about you and your project. And that's why like creating skills and commands that are very structured help it so much because you can't assume it remembers the last session, because it doesn't, unless you had it write stuff down. But the memory problem is a real big problem at a certain scale. And I'm sure someone's going to figure out something fancy at some point. But yeah, for individual context windows, I've made huge strides in my personal workflow by using Obsidian and teaching it how to structure docs and using templates and linking stuff together and allowing it to use my second brain as its second brain. But just like me, I have to remember that a note exists about it. It can't remember that. So I need to tell it because I don't need it to remember whatever pattern we're using for X every session. So can't load that up every session. But when it is appropriate, I need it to pull it up. I'm looking forward to the time where I don't need to say, remember about this pattern. I can tell it, remember, we're going to use this pattern. And it knows, okay, that pattern is stored in Obsidian here. But it's still, there's a lot of manual intervention there. And we don't have a good way of doing it at a company level just yet. I think when I get a little bit more comfortable with my current workflow, I'll pitch something similar to the team of like, hey, we usually hold our architectural decision documents here. What if we expand it a little bit and instead of AI, we have a whole bunch of local stuff? What if we just started dumping stuff there and teaching the AI how to search stuff up? Like if we did it in a structured way, it shouldn't be that hard. And then see where it goes and how it evolves. I think AI is so fresh and so new that we're all gonna have to figure these things out, right? Like Rails wasn't built when web applications was built. It came from the frustrations of building enterprise software and all the fucking boilerplate and configuration that went in every single time. And someone finally was like, I've had enough of this. Here's a better way of doing it. And now a ton of people, and not just Rails, you know, Django and sales and all these other Rails-inspired frameworks for other languages exist because it was like, yeah, this is better. I think those patterns will emerge organically as people play around with it. I'm not waiting. I'm just tinkering with my own. I don't have a good answer for you yet, but that's what I've been doing.

SPEAKER_00 37:49

That's kind of the fun part about being a programmer right now is also part of maybe the bad part too.

SPEAKER_01 37:58

But we're helping it get better at our jobs so we can take our jobs.

SPEAKER_00 38:02

It's kind of fun. Teaching it how to make us obsolete is kind of fun. Yeah. I haven't looked at how OpenClaw does this, but apparently they've done it well with being able to recall stuff. There's also the QMD, I think, is the like CLI search for Markdown. I think Shopify is from Shopify or but anyway, yeah, that's cool. I hadn't thought about using Obsidian CLI. I heard that they had come out with CLI, but I hadn't thought about using obsidian features in your markdown files, like the linking and stuff, so that it's not having to draw these connections. The AI isn't having to do that.

Why Blast Off Rails Focuses On Rails

Why Albuquerque: Access, Cost, Venue

SPEAKER_01 38:49

You don't necessarily need to have obsidian. You can just tag files, right? Like it is nice that obsidian offers you a structured way of searching and building dashboards and whatnot, where like I can build a dashboard and it knows to go and look at that dashboard when I give it the good morning command because it looks at all my to-dos from the day prior and whatever. You know, when I tell it to wrap up a session, it knows exactly where to put the session, how it should look, because there's a template. That stuff could also be just in the command itself, like here's how to structure it. But it is nice to have the obsidian way of doing it that's already been iterated on, right? Like obsidian's already solved some of these problems, so it's nice to use it that way and have some structure for the AI. Whether or not it's the best way, I don't know. It just works for me right now, and I'm still tinkering with it. And every, almost every day I tweak something about the setup, or every week, you know, when I run the command, my weekly wrap-up command, I go, oh, you know what'd be better, or you know what I wish it had done? Or, you know, there's still a lot of manual, and it's still like coding, right? Like we still we build the simplest thing and then we iterate on it, we add more to it, and it'll evolve over time. And I'm sure someone will solve it in a way that either gets absorbed into how Claude handles its internal memory at the model level, or and maybe it is more of like the open claw way of doing things, but yeah, we'll see. It's interesting, it's an interesting problem to solve because it yeah, the models have amnesia. They don't know nothing about what's going on other than what we load into it, but they have a small context window. You can't load everything, you can't just be like, here's my app. So a little bit of job security is nice right now because like I know our app better than the AI does, but I don't know for how much longer. We'll see. Maybe it's worth you looking into too, is like, how can Obsidian solve some of these issues that I'm having? And the cool thing is you can just ask the AI, hey, I'm having this problem with your memory, considering a tool like Obsidian. What do you think? And it might give you some interesting suggestions. There's some resources online. Help unblog you a little bit? Or yeah, no, that gives me some things to think about. As much as the memory problem is interesting, and I'm sure someone's gonna solve it. I'm wondering how long it's gonna take to get to the point where token efficiency is so valued that that's what we put on our resume. How token efficient we are, how good we are at getting good results by using less tokens and like what tools we have built to say, oh, I built an agent harness that does X, Y, or Z under the hood before it prompts the LLM and it saves 20% of your tokens or whatever and reduces costs. Like, I know we have a pretty big budget for it at work because we're still in the exploration mode and we know we need to, but there's gotta be a tipping point where people go, okay, I'm not willing to spend this much money on this thing anymore. Right. Let's get efficient.

SPEAKER_00 42:05

You put on your resume, you know, migrated this many users to this new system and like shut down the legacy system or something like that. But in the new age, you have that line on your resume, but it says using only 150,000 tokens or something like that.

Budgeting, Travel, And Early Bird Details

SPEAKER_01 42:22

Yeah, you go to join a new uh new org and they're like, let's see your clawed code stats. Yeah, how much do you how much money are you gonna cost us in addition to your salary? I don't know. Maybe that won't happen. Maybe the models will continue to get more efficient and cheaper and it won't be a problem, but maybe because we're having a huge shortage in RAM and GPUs and all the stuff that it uses. So that stuff's gonna start costing a premium. Maybe it is gonna happen. I've already started playing with the idea of like, okay, which model's better at planning, which model's better at building? Let's try this mode where I use like clawed opus in plan mode, and then I use open AI codecs in build mode and start looking at the differences and stuff. And then I I use open code, where most of my team is using clawed code. There's a few of us using open code, and open code's really cool because it's really easy to switch models. So I've gotten to play with, you know what, this task is like not a big ask. I'll give it to a dumber model, like a not as expensive model. Because I just don't feel like doing this boilerplate BS, but I don't think it needs the deep reasoning of Opus or Codec. So that doesn't matter right now because we don't have any restrictions, but I'm trying to get the muscle memory so that when we do have restrictions, and if we do have restrictions, that the muscle memory is already there to like certain tasks, certain models. Don't sort of like you don't give a senior engineer a junior level task because it's not cost effective. I don't know if that's going to be a thing in the future, but I have a sneaking suspicion it will, so I'm already trying to plan around that.

SPEAKER_00 44:03

Yeah, or just using the right tool for the job, like, or the right programming language for the right application. Like it's gonna be important, I think. There's another tool called Counselors that Aaron Francis made, and it basically is like giving the same prompt or problem to multiple models at once, and then one model will kind of summarize what they all said and kind of bring it all together and see if there is one implementation that is best fit from all of these different uh perspectives. Interesting. I thought that was kind of cool. Probably can get expensive to do. Oh yeah. Yeah. But yeah, interesting to like get the different perspectives and then maybe you just have one implement it based off of the best kind of case that was built from all of them. But I think efficiency will be will be important right now. People are kind of throwing it to the wind because you just have to experiment. There's just so much happening.

SPEAKER_01 45:12

How do you know if you don't experiment? Yep. You have to be free. So the last question, my favorite, is what is something cool, new, or interesting that you've recently learned, discovered, built, whatever, vibe coded, doesn't matter, that you want to share. Doesn't have to be coding related, it absolutely can be.

Speaker Announcements And Where To Find Travis

SPEAKER_00 45:30

I will plug Blast Off Rails here. It is new, so this will be the first time this conference has happened, and I hope that. It will be cool and interesting. I'm trying my very hardest to make it that. But yeah, it's happening this summer, June 11th and 12th. And it is a place for people to gather and kind of get energy. That's the reason that I have gone to and continue to go to conferences is there is for me, there isn't really a substitute for the energy that you get from being in a room with a bunch of people that are interested in the same things as you and being kind of forced to sit through these talks. And I don't mean that in a negative way, like, but for me, I'll see these interesting talks, like YouTube videos. There's recordings of talks, there's people that put stuff on YouTube, there's interesting blog posts that I always like bookmark. I'm like, oh, that sounds really cool. I'm gonna listen to that later. I'm gonna watch that later. I'm gonna read that later. And I don't get around to it. But at a conference, you paid money to be in this room, and people are going to give you the blog posts, give you the YouTube video live, and you set this time aside to do that. And that's really special to me. So there's also a ton of other conferences happening that I'm going to them and I'll just plug them as well. There's RBQ in Austin in March, I believe into March, and then there's uh Blue Ridge Ruby happening in North Carolina.

SPEAKER_01 47:21

Yeah, Asheville, North Carolina. That's Jeremy Smith's baby. That's I've been to that one a few times. That's a great commentary.

SPEAKER_00 47:27

So that one's kind of still newish. Yeah. They did it and then took a year off, and then they're doing it again. So that one's returning. But then there's Blast Off Rails in June, June 11th and 12th. And then there's RubyConf in July. So I'm planning right now on going to all of those. So it's going to be a busy year, but I'm super excited.

SPEAKER_01 47:54

Don't forget there's Rocky Mountain Ruby in October. Yeah, yeah, yeah. I don't think they've announced the dates yet, but that's been going strong. And that's a I love that conference. I love the conference. Spike and Becky do such a great job putting it together. And I love Boulder, the city that it's in. I have pitched to my wife moving there on more than one occasion because I just love it out there. I love seeing the regionals come back, but also I'm like, damn, there's so many conferences I want to go to. I don't have the time to go to all of these. It's like trying to structure my calendar. My we get some time to go to conferences, but not a ton. So I'm like, oh, I'm going to have to use PTO for some of this. And which places does my wife want to go to? Because then I can roll it into a vacation where she won't want to murder me for using all my vacation time going to conferences. So yeah. Actually, the one interesting thing that stands out about Blast off over some of the others is like, it's Blue Ridge Ruby, it's Rocky Mountain Ruby. Yours is Blast Off Rails, not Blast Off Ruby. So you're focusing more on the Rails aspect than just the general Ruby. Was that an intentional decision because everyone else was Ruby and you wanted to be different? Or was it just, well, I use Rails, or you wanted to keep the topic on Rails and be less about the Ruby language? Like, what was the motivator to making it blast off Rails?

SPEAKER_00 49:16

It was partly to sort of differentiate it from the more general Ruby conferences. Like one of the I think it was like a random piece of information from some book I read that was like actually making the thing more niche makes it more interesting to people. And you're more likely to get people to whatever it is, buy your product or come to your conference or use your tool or whatever. If it's more niche, because you can actually focus on satisfying people more because you are doing a more specific thing.

SPEAKER_01 49:59

Yeah, yeah, it makes sense.

SPEAKER_00 50:00

For me, I am more interested in the Rails stuff. I haven't written a lot of other software, to be honest. So I am mostly Rails apps. So that's like always been what I've kind of gravitated towards at conferences. So I was like, you know what? Let's make this conference more specific. And hopefully that will make it more interesting because you can expect, like, oh, I'm a Rails developer, I make Rails apps, I'm gonna go to this conference and it's single track. So if you're not interested in a talk, then you're out of luck. But yeah, if you're interested in Rails and you come to this conference, you're gonna get Rails content and stuff that's applicable to you. So yeah, ended up going with blast off Rails instead of blast off Ruby.

SPEAKER_01 50:53

Okay, cool. And Albuquerque, New Mexico is definitely somewhere I haven't been for a conference before. Is there a big tech scene in Albuquerque? Is it just that's my backyard? So screw it. Everyone come here. It's a pretty cool place, and you're about to find out, or what's Albuquerque?

SPEAKER_00 51:09

I think there's a couple advantages to having a conference in Albuquerque. One is it's kind of centrally located. It's still in this like southwest of uh the US, but for most people, there's at least only one layover. A lot of times there's a lot of direct flights here from from most major airports. And our airport is also really small, so it's enjoyable. So you have the best of uh both worlds. Like you have a lot of direct flights, but there's only like one or two a day from all these cities, and so the airport's still pretty small, and it's not a nightmare to find your way around and stuff like that. There's one terminal, like you're not like taking a train around.

SPEAKER_01 51:55

You just described my least favorite part of some of these conferences.

SPEAKER_00 51:59

Yes. But Albuquerque is a cheap in terms of like cost of living, and it's cheaper to put on a conference here, and it's also your hotel is not gonna be$300,$400 a night to be near the conference. There's like$175 a night hotels that are within walking distance of this venue, so which is the the Albuquerque Museum, which I think is gonna be really awesome. Everybody that comes to the conference also just has free access to the museum, so you can learn about the place that you're coming to. So, yeah, there's a lot of uh cool stuff here. We don't have like uh big tech companies here, but I mean there are like there's Facebook data centers and stuff around, but I think it's just a nice place to visit and nice weather in June. It's just a nice place to have a conference and have it be an overall, generally good experience for every kind of person. So I'm really excited for people that haven't been here to come here and experience it.

SPEAKER_01 53:11

Yeah, very cool, very cool. I like how accessible it'll be from a cost standpoint, too, especially with not every software company out there is sending people to conferences right now. So a junior might have to pay their own way. And if this is going to be the most cost-effective conference they can make in the year, that's a pretty big selling point. It can get expensive going to a lot of conferences.

SPEAKER_00 53:32

So we have like a budget for conferences at work, and it's really easy to blow through that really quickly because you're a couple hundred dollars at least for the flight. If you're gonna stay three nights for a two-day conference, at the very least, you're talking about over a thousand dollars for a hotel or close around there, even if you're staying at like a$300,$250 a night hotel that nowadays is the cheap ones, especially in major cities. You're over a thousand dollars for every conference that you go to if you're on a shoestring. And I think you could do this conference, depending on your flight, you could do this conference for less than a thousand dollars. Because early bird tickets, which there's 30 of them, there will just be that until they're gone, but they're only under$75,$175, and you get two or three nights, or three nights, let's say, at$175 at the best western down the street, which is a nice best western. I actually worked there before I became a programmer. You have one forty dollar Uber from the airport, another$40 to get back. Everything else around the conference is walkable. I think it could be a very uh cost-effective conference.

SPEAKER_01 54:57

And your CFP is closed, so you're gonna have the announcement for who's speaking soon. Do you have a date?

SPEAKER_00 55:03

That'll be coming soon. I don't have a date. Trying to confirm people right now. It'll be probably early March when I make that announcement, just so I can get everybody confirmed and get that all up on the website and everything. So very excited about that, though.

SPEAKER_01 55:21

Sounds like it's gonna be a great one. Good excuse to go to Albuquerque, New Mexico for the first time and blastoffrouse.com. Sounds like early bird tickets are very affordable, so don't sleep on those. Where can people find you on the web other than blastoffruby.com?

SPEAKER_00 55:41

Because of the conference, trying to get the word out there, I'm on everything. I'm on uh blue sky, I'm on X, I'm on LinkedIn. I think if you just search my name, you can find me most of those places. And uh you might have to look at the picture to try to match my face because there aren't many Travis doctors in the world, but there's at least like two others.

SPEAKER_01 56:08

Yeah, and your doctor with it's a D-O-C-K-T-E-R.

SPEAKER_00 56:13

Yeah, which probably the best way to find me is to look at the Blast Off Rails social account. And I think usually I'm like reposting it or something, you can find me pretty easily.

SPEAKER_01 56:23

Yeah, cool. We can also put links in the show notes to make it a little easier for people to find you. So awesome. Well, thanks for coming on the show, man. I really appreciate it. Appreciate your time and looking forward to seeing how uh a Blast Off Ruby comes together.

Closing And Next Steps

SPEAKER_00 56:37

Yeah, thank you for the time and all the good conversation.

SPEAKER_01 56:40

All right, listeners, we'll see you in the next one.