Posts

Making it work for me Some thoughts on my “secure communications” course that I had Claude create for me. I haven’t really started the distributed systems course, and may remove it, since two intensive courses may simply be more than I can handle. Great information. The technologies covered are fundamental and important. It was important to me that I learn something that is foundational and (because it’s foundational) slow to change, because I find that such knowledge helps the greatest in understanding higher-level constructs. Too hard?: Some of the challenges are out of reach based on 2-3 hours per day. Obviously, I had to get a lot of help with parsing a raw ClientHello (day 8) for example. I wish I had the time to just slog through it, because I think that yields the deepest understanding, but I still got a lot out of having Claude talk me through it. And since we have the aid of AI, maybe these sorts challenges are actually appropriate. There is always of course the danger of getting it done without learning, so it’s important to keep yourself honest when learning with AI help. Interesting activities. Looking at TCP packets with WireShark is really fun and interesting. Rabbit holes. Following up on my previous musings about whether it was sustainable to go down every rabbit hole, I’ve modified my rule a bit. I’ll take a look down every rabbit hole, and if it’s not too crazy, I’ll consider going down it! My previous rule was somewhat tempered when I ran into deep cryptography stuff. Honestly, I haven’t even attempted yet to understand what “Elliptic Curve” math is all about because a) it just sounds really intimidating, and b) I think that’s somewhat beyond my role anyway, and I’m not sure how immediately useful that knowledge will be. And while we’re at it, the “key schedule” is a tad bit confusing. I decided to be satisfied with knowing that it’s for creating keys from the shared secret. I had to take a deep breath when I learned that the shared secret is not used directly in encryption. This course is kind of a new experience for me. My previous self-teaching was all project-based. I would learn what I needed to learn in order to build what I wanted to build. But now my goal isn’t building something, so the choices I have to make are very different in nature. I really prefer project-based learning when it can be done. But projects are most motivating when it’s a tool that you’ll use every day, like my “course manager” project that I built for using as a teacher. And while I could have started any number of additional “personal tool” projects, it was difficult for me to think of one that targeted the sort of knowledge I wanted to gain – foundational systems knowledge. ...

Slower than expected I got sick, so that put a damper on things. And watch was a little busier than expected. So I’ve fallen behind. Firstly and most importantly, and perhaps unsurprisingly, the quality of the course plan is not very high. Claude has me reading entire RFCs. Ok, I’m not sure that’s the best way to learn something. I find RFCs not to be written for the uninitiated at all. You kind of have to know your way around before diving into one of these. I mean, it’s interesting, but I don’t understand most of it. ...

Updates I’ve been working on learning Actual Budget and I ultimately decided to finalize the switch and canceled my YNAB account. I must say, Actual is quite a nice piece of work. It has a simple, intuitive, and snappy interface, and I think its capabilities actually exceed YNAB’s in some places. Don’t quote me on this, but for example I believe you can only have one spending target per budget category in YNAB, but in Actual, the equivalent concept of templates allows as many as you want. Quite powerful. Templates are an experimental feature but I’ve found them to be fully functional, bug-free and very useful. The connection between Schedules and templates is quite handy as well. ...

Some news I work at Amazon now. And my family and I moved to the Seattle area. But right now I’m in Florida and I’m on military leave for 6 months. Enough about that! Before my move to Florida a couple weeks ago, I had been setting up a “home lab” - in my case, just several containerized servers running on an old laptop. So in my spare time I’ve been working on setting up a remote version of this. You can see the repo here. ...

So I was using ChatGPT in the browser (that’s the only way I use it), and I glanced up and saw a little icon that said “>_ Codex”. Ok, I have a bit of time. Let’s check this out. Within minutes, I had given OpenAI access to my Github, installed the Codex CLI app, and was able to create agents locally. Scary easy. I have been curious (and honestly, a bit horrified as well) about the whole “agentic” thing and I have wanted to try things out, but I hadn’t taken the time because I thought it would require quite a bit of setup and I was also afraid of the agents doing something stupid to my machine (due primarily to mistakes my own permissions or instructions). ...

So I’ve written about my course-manager project on here. I know, it’s a terrible name. I guess I’m not great at marketing. I think I might rename it to “CourseMark” which sounds catchier, right? I already have the domain! Anyway, I digress. This is not an issue with that project per se but rather with my own deployment setup. I’ve been using Digital Ocean Droplets, which is just like Azure VMS or Amazon EC2 instances, they’re cloud virtual machines. ...

I was taking a month-long course where we had to use the school computers instead of our own. I had some down time during breaks and wasn’t able to access my personal laptop, so I decided to try to do things in the cloud. This took me a while because the school’s Microsoft network just happened to block digital ocean and google cloud consoles which I am more familiar with, but fortunately allowed me to access Azure cloud services (weird, right?? Wonder why?). Azure pricing was a bit intimidating but was able to set up the cheapest virtual machine I could find at about $8 a month, more than enough power (1GB RAM) for my purposes. I just wanted to play around in Linux so I could keep learning. I started setting up Vim for Go development and then decided to try out Neovim. Long story short, this led me ultimately to replacing my own setup on my personal machine! I’ve spent the last couple of weeks getting familiar with this IDE and I have to say it’s been a blast. I’m still not as productive as I am with VS Code but maybe I’ll get there soon. ...

Happy Birthday to my Dad, and happy day before birthday to my daughter Yaretzi! We are celebrating her birthday today with a little family party. Music has always been a very big part of my life since I was pretty young. My brother and I took piano lessons (classical music, mainly) from a fantastic lady named Verna Ferrell from the age of 8 to 18, and I also got involved in the school choruses just about every year since middle school. ...

I’ll transferring to Charlotte Engineering Early College for the 2025-2026 school year. I’m excited about the move, but I’m also a little sad because I have really enjoyed working at Phillip O. Berry. The students are awesome, my colleagues are awesome, and the admin is awesome. There is a lot of heart at this school and it will always hold a special place for me. Mr. Johnny Darling and I plan to continue our new podcast, so please watch out for that here Charlotte Tech Talks or on Apple podcasts. ...

Here is the first of what I hope to be many reviews of books, shows and movies. I’m not finished with it yet, but since it’s due back to the library and I can’t renew because of holds, I decided to write my review anyway. This is a serious book about the problems posed by the nascent technologies Artificial Intelligence and Genetic Engineering, and the solutions required. The author lacks nothing in credentials – Mustafa Suleyman is a co-founder of multiple successful Artificial Intelligence companies, including most notably the famous startup DeepMind acquired by Google in 2016. Google’s description suffices here: ...