Week 12, 2024 - Homelabbing, Leadership and AI Music

Week 12, 2024 - Homelabbing, Leadership and AI Music
Oldschool tinkering - Fortepan / Kádas Tibor (1975)

🦸 What I learned this week

Oh boy, did I learn! I haven’t spent this much time in the terminal for ages. I received the replacement SSD on the weekend for my 2011 Mac Mini and spent most of the week playing with it and learning about virtualization, networking, and other technologies. As I write this on Friday, I have Proxmox running 4 LXCs and 2 VMs, the latter hosting a combined 15 Docker containers, with the media stack I used to have on MacOS, Pi-hole providing DNS for all the devices in the family, and some more. It was a rollercoaster that included a long support session with my ISP after I accidentally turned off IPv6 on their router, but I had so much fun that I almost felt guilty about it. Tinkering with new toys is always addictive (and the victim is my sleeping schedule).

More importantly, after last week's article on the opinions that could suggest challenges in an Engineering Management career, this week I wrote about characteristics and behaviors that could help someone become a great Engineering Manager. The list could be useful for self-reflection for Individual Contributors considering a leadership career path and for managers looking to start conversations with potential leaders in their team.

Working on these two slightly tongue-in-cheek lists made me think about what really makes a good manager. I realize it’s very different for everyone, so I’m not looking for some universal truth. But the way I work, it boils down to these six critical aspects:

  • Authenticity. Everything relies on a solid, authentic self-image. Authenticity requires transparency and builds trust. It takes second-guessing and doubt out of the equation, making everything smoother and more reliable.
  • Empathy. Being able to see the world (and ourselves) from the other person’s perspective is a superpower that can make a leader’s work much easier in every situation, whether they are figuring out how to support someone, navigating a challenging negotiation, or discussing uncomfortable feedback with their managers.
  • Communication. Everything is made easier or harder depending on how well someone communicates, from setting objectives to resolving conflicts, team coordination to managerial feedback, project planning and interviewing, change management and mentorship. I couldn’t find who said it first but they were right: “Leadership is communication.”
  • Positivity. A leader needs to be able to build a vision and motivate people to follow it, and that’s impossible to do without being positive. But even beyond that, genuine everyday positivity can be infectious and help the team overcome morale lows.
  • Decisiveness. Being comfortable with ambiguity, and having to make decisions with limited information. Utilizing frameworks, and understanding the various inherent biases. These are crucial skills of an efficient manager.
  • Patience. Patience with people on the team and persistent support until the results show. Patience with peers and managers when they have a bad day or need more time convincing. Patience with customers and vendors, because a sustainable business is built on long-term relationships. And patience with ourselves, accepting our imperfections, and focusing on the lessons we can learn from our failures.

🤓 What I want to try next week

I have a few half-written articles and rough ideas, but I can’t commit to finishing anything yet. And who am I kidding: I’ll probably spend a lot of time experimenting with self-hosting, maybe moving part of the stack of this site to home too. If that happens, I’ll need a solid offsite backup and maybe some redundancy in the infrastructure too. What a rabbit hole!

🤔 Articles that made me think

This is a new chapter, so let me introduce what I have in mind. I was never a fan of “drive-by sharing” links on company Slacks and newsletters, just a URL without added context. Tell me why I should read it! Was there a particular aspect that personally resonated with you? Do you think it ties to an ongoing project or issue at the company? What new ideas did you get from it? I try to lead by example and will use this part of the weekly newsletter accordingly: instead of summarizing articles I recommend, I’ll add what I got from them, hopefully creating more value for the readers than some links with summaries would.

With that intro aside, here’s one article that made me think this week (triggering a little nostalgia at the end):

Paul Graham: The Reddits

Probably a deliberately timed nostalgic article to boost the Reddit IPO, but for me it was a good reminder about the importance of change and focus. One of my weaknesses I’m working on is around this area: I understand why constant change is important, but I can get anxious seeing the rug pulled under me in different directions, and it makes it increasingly difficult to keep my focus.

YCombinator, the VC company financing Reddit in the beginning needed to change their policy of “investing into ideas and not founders” when they had a gut feeling about these guys. And these guys needed to accept to scrap their idea and work on something entirely different too. They agreed, and in 3 weeks launched their first version! This shows incredible focus — but more importantly, how critical it is to deliver something fast, to be able to start receiving feedback, and change according to user needs as soon as possible.

I also liked that the initial idea of Reddit came from an unintended use of a different product. People started to utilize the “trending” page of del.icio.us, an online bookmark manager, as a discovery tool for new content! It shows how important it is to focus on the way users are actually interacting with a product and be ready to change quickly to react to that.

I had a similar experience in the mid-2000s, right around the time Reddit was founded. A few friends and I were writing a popular Hungarian blog (inspired by the likes of Boingboing) and published an article that was a simple quiz where people had to recognize old computer games from a screenshot. Probably because the game was insanely hard (it only accepted exact name matches!), and didn’t require you to understand Hungarian, it got so popular that our server died under the traffic. We immediately looked for an alternate host, made a whole new site in English for similar games, and had quiz-em-all.com up and running in a few days. The popularity lasted for a few months, and we made a little money from advertising after our hosting costs, but eventually, the whole thing died off, like most startups. I still think of this as a good example of embracing change, taking advantage of an unintended use of your product, and pivoting to something else fast.

🕺 Something cool: AI-generated Music

I see how this section is turning into “something AI”, but bear with me, this one is really cool.

Suno.ai allows you to create music based on your prompt, and the result can sound better than this statement. Unlike previous attempts at freely available AI music generators, Suno can also handle singing surprisingly well – in non-English languages too.

So, to stay on the topic of this newsletter, as a quick experiment, I turned the Agile Manifesto into a pop song. I know, it’s far from great, but you have to admit, it’s catchy! (There are better examples on their Explore page.)

I don’t know what this means to music as we know it today, but it’s fun to guess. There is AI already being used in playlist generation, but this field will explode when the functionality is built into assistants. Just imagine, someone can wake up and turn on the “radio”, and get matching background music for the day based on things like weather, news, commute traffic, upcoming calendar events, shopping lists and reminders, browsing history and geolocation, the active window's content and the person’s current heart rate… really, anything. Are you going to a big meeting? You probably need some motivation. Is it Friday? Let’s party. Did you have a long evening yesterday? Better keep it quiet and instrumental. Or, you’re a medical university student and need to memorize the Latin names of all the leg muscles in order. How about a catchy song to help you do that? Or, how about a simple hardware device that you can just put in your café corner, connect to speakers, and let it generate music based on what it sees on the street with its camera? “Getting late, evening's glow / Girl on bike and a guy with dog. / Whispers pause, eyes will meet, / Brief encounter on an empty street.” Everything is ready for this, you just have to wire the tools together. 🤯

For (human) musicians though, the future might not be this much fun.

First, they’ll need to adapt, just like they had to with every technology change: moving from live music to radio, then to the stores on recorded media, and now online streaming. I guess that they’ll need to get great at building and maintaining a community of fans around them more than ever. (Just like journalists, I’m afraid.) The tools are there to use, from Substack to Patreon, but this takes time and effort that could’ve been spent on, well, making music.

Second, in the coming years, we’ll need to figure out what this whole AI-generated content thing means to copyright protection. How can you claim intellectual property on something that was trained on others’ work, without attribution, let alone compensation? But on the other hand, isn’t learning to play music the same? Why should studying art to learn from it be different if the student is not flesh and bones but software? In the end, even Suno ends up advising you to get a good lawyer.

Well, this got pretty long, thanks for bearing with me. Have a headbanging weekend,


Subscribe to my newsletter

I write about engineering leadership topics.
Sign up to receive new articles.