This was the year I built my first homelab. I now host all my own media (movies, shows, audiobooks, and ebooks), my notes, files, tools, custom coded software, websites, local AI, and more! It's been a tremendous learning experience and opened up new possibilities for me.

Learning Docker has been a huge unlock for me in running software. Such a useful setup.

I think folks with ADHD or autism might have an advantage in creating things. Autists are naturally obsessive, and folks with ADHD can naturally only concentrate on the things that interest them most. That obsession can be a superpower for building products and companies.

I hope everyone had a wonderful year, and that the one to come is better than the one before. ❤️

When building out my homelab, definitely hitting the point now where decisions on what *not* to self-host become important. For instance, I don't think I want to selfhost a password manager instead of using iCloud Passwords because if it breaks I'm locked out of everything.

humans are crazy

The weirdest thing about making huge personal progress is it makes you feel a bit like all the time before was wasted not knowing what you do now.

I've made extensive use of ChatGPT, Gemini, and Cursor for the past 3 years, but today I start trying out Claude Code. While it's always set the standard for agentic coding, I've appreciated visual interfaces over command line. The past year of learning has expanded that.

This was a fantastic resource for setting up Forgejo for the first time on my homelab: nickcunningh.am/blog/how-to-se…

I gotta say, OpenClaw🦞 is so much of what I was slowly building on my own but way more and better. I'm excited by the possibilities it's opening up for me after only 24 hours working with it. Can't wait to see what we can do by the end of the first week.

When even the CTO of Dropbox (who was also an early engineer at Facebook) is saying coding won't be done by hand any more, you have to start believing it. We are in a new era of software and building things. Adapting is the best path forward.

AI is still very bad at design. I understand all the reasons why work is being done on coding first, but I'd love to see an actually great AI designer this year.

New art project. Train and inference GPT in 243 lines of pure, dependency-free Python. This is the *full* algorithmic content of what is needed. Everything else is just for efficiency. I cannot simplify this any further. gist.github.com/karpathy/8627f…

I feel this so much

There's an interesting phenomenon happening right now where the *best* coders don't write code any more, while lesser coders staunchly avoid using AI heavily. I think most people expected the opposite to happen, and it's kinda wild to see how much different the reality is.

now this is getting really fun :)

things just keep moving faster and faster!

Looking forward to trying this out as a possible replacement for Opus 4.5 in OpenClaw. If we can get significantly cheaper while retaining high performance, it'll be a big boost to potential adoption. For me it means I'll just get to do 5x as much :)

There's people who want to argue and there's people who want to build, and I vastly prefer the latter. It's so much more enjoyable to keep moving forward rather than constantly try to stop things.

Every technical founder who had stopped coding 10 years ago