Blog on mzll

The failure of a public pledge

lg@mzll.it (Luigi Mozzillo) — Sun, 17 May 2026 01:22:22 +0100

I haven’t lost any weight—I’ll say that right off the bat. Or rather, I’ve lost just under 1 kg (~2.20 lbs). This public commitment I made to myself didn’t go as I’d hoped. I kept up with my swimming sessions, the occasional brisk walk through the traffic of the Neapolitan suburbs, and some home strength training, but nothing that actually helped me lose weight. I think I need to consult a nutritionist and start a structured, serious program: I just need to find the time.

In short: experiment failed. Making my intentions public didn’t help me stick to it. Even though I had your eyes on me, I’m still the same as I was two months ago. I’m the same person. The same idiot who doesn’t take the idea of sticking to a diet seriously.

But sooner or later I’ll make it. Maybe.

Thank you to everyone who has written to me over the past few weeks to encourage me and give me a boost: I’ll have to repay you with some chocolates.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Stochastic falls

lg@mzll.it (Luigi Mozzillo) — Mon, 27 Apr 2026 01:22:22 +0100

I left the company where I had been working for the past (almost) year. There are various reasons for this, but they all boil down to two main issues.

The first is related to my previous experience. Since I am still involved with my former company, I am working – together with the chairman of the board of directors, who is just as committed as I am –to get a portion of the business back on its feet so that we can try to get back on track and resolve some old issues with the tax authorities and creditors.

The second reason is that the company’s ownership has completely thrown itself into the arms of AI, viewing it as a panacea for all ills and failing to consider the boomerang effect that decisions made today – in haste, without professional advice, without preparation, without in-depth analysis, without understanding the facts or the context, without planning, without shame, without vision, and without trusting those who have been telling them for months: What the hell are you doing? – will have.

AI is certainly a challenge for workers. But I know for a fact that it will also be a huge problem for many entrepreneurs who embrace it without conducting serious, thoughtful assessments of their business strategies. Without knowing the path ahead or paying attention to their steps, the likelihood of stumbling and getting seriously hurt becomes far stochastic.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

When I travel for work

lg@mzll.it (Luigi Mozzillo) — Wed, 15 Apr 2026 01:22:22 +0100

I’m back to traveling for work. My old project and my new job will require me to travel more often, much like I did in recent years, when four out of every 14 days I spent 800 km away from home.

Over the past few weeks, I’ve taken trains and planes several times. I’ve set foot in Italian train stations and airports again, with goals to achieve and no museums to visit. I like trains and planes, but I don’t like absolutely everything about them.

I don’t like it when the train is late. Damn delays. And not to mention the fact that sometimes you have a lot of free time, and you can’t just spend it watching TV shows. And I don’t even like the morning train from Naples to Rome, because the car is packed to the brim. And as for the plane and the airport, I don’t like the prices. Damn, what prices! And I don’t like it when you undress, wait for your suitcase, and then get dressed again – maybe, if they don’t throw anything away. And I don’t like waiting—all that damn standing around. And I really don’t like airport parking lots.

Instead, what I like about the train is how comfortable it is, and having time to watch a TV series in the evening on the way home, sprawled out. I like the stretch between Rome and Naples on the return train, when there are usually only a few of us. I like it when he’s punctual. At the airport, on the other hand, I love that even at 4 a.m., at the airport there’s always a festive atmosphere, full of joy and hope. Dreams come true, at the airport. And I love the world inside the airport, and the languages, and that liminal space between life and travel, between everything and nothing – like Schrödinger’s.

But what I like best about my business trips – and what brings me joy, whether I’m traveling by train or plane – is the hug from my daughter when I get home.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Dotfiles and nerdy stuff

lg@mzll.it (Luigi Mozzillo) — Sun, 12 Apr 2026 01:22:22 +0100

I need to replace my work Mac within a few weeks. ¹ My work laptop is synced with my personal one for everything that involves iCloud. I have pretty much the same apps, and I use Homebrew for everything I can. So, with the switch coming up, I’ve been thinking about rebuilding my dotfiles for macOS.

The idea is to clone all the main configuration files (git, neovim, claude, micro, binary scripts, ghostty, aliases, zsh, ssh, …) and keep them synchronized in a GitHub repository. In addition, I’ve included all the applications (both command-line and GUI-based) installed or updated via Homebrew in a script, so I can install or update them quickly.

I used dotfiles years ago, back when I only had Linux devices. Then I got a bit too comfortable, and when I switched to the magical world of Apple, I lost that very sane habit. With my increased use of the terminal – both because of and thanks to LLMs – I figured rebuilding my collection of common configurations would be a good idea. And it certainly was.

While improving my setup, I discovered stow, a command-line utility that allows me to classify, organize, and update the symlinks for configurations to be replicated more quickly and efficiently. It’s extremely useful for updating the symbolic links to all files (using a loop, I update all configurations in the same installation script mentioned above).

In short, if you use the terminal, it’s a process I highly recommend. It’ll be a lifesaver if you unexpectedly switch devices or use multiple devices at the same time.

I really missed all this nerdiness.

Yes, I’ve decided to leave the company where I currently work as CIO to take on a new professional challenge. We’ll talk about it. ↩︎

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Slop and translations

lg@mzll.it (Luigi Mozzillo) — Tue, 31 Mar 2026 01:22:22 +0100

AI has recently been used on this blog just to improve the graphic design: I’ve asked for help on several occasions to fine-tune aesthetic details that would otherwise have taken me forever. But never when it comes to the content. And that will never happen because, since this is a personal blog, writing using soggy stuff generated by an unconsciously system would be really stupid.

But my English sucks. And writing in English wastes my time. And I don’t think in English – if only! So I write my content in Italian first and then translate it.

I’ve noticed that, as the months go by, some of the systems I use to translate this idle chatter are increasingly relying on algorithms that, thinking of themselves as generative, add a touch of creativity to my texts. And my work of reconstructing the original meaning is sometimes turning into an epic undertaking.

These days I use DeepL and Google Translate: I use the first one to translate and the second one to correct. ¹ But more and more often, I find myself having to rework sentences that weren’t mine to begin with and that have been mistranslated.

And so, here’s my dilemma: does my content come across as truly my own, or—even though the words are my own—am I ultimately writing a blog as if an AI were writing it?

If you have any other products to recommend, please feel free to do so. ↩︎

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

New needs and new possibilities

lg@mzll.it (Luigi Mozzillo) — Sat, 28 Mar 2026 01:22:22 +0100

I had a PDF in English, so I asked Claude Code to translate the text into Italian and convert it to EPUB. Done – it works and reads well. It took me less than 5 minutes. Things that, I don’t know, might not even have occurred to me to do before.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

The Gratitude Subscriptions, March 2026

lg@mzll.it (Luigi Mozzillo) — Thu, 26 Mar 2026 01:22:22 +0100

I won’t go into the reasons why you should use a password manager. I’m assuming that if you’re reading this blog, you’re nerdy enough to already have one. Right?

Anyway, I’ve used plenty of password managers over the years. I’ve used Enpass, LastPass, Apple Password, Bitwarden, 1Password, and Dashlane, and there are many others I’ve at least tried out to see how they worked. Switching from one solution to another has never been complicated, and when I did switch, it was either because I needed a better, more structured way to organize my data or, sometimes, for simple cost reasons.

For a few months now, I’ve been using VaultWarden – though I’m not sure if it’s permanent.¹

VaultWarden is a server implementation compatible with Bitwarden. Essentially, it’s a “behind-the-scenes” alternative to the Bitwarden server—lighter and fully suitable with all Bitwarden applications (desktop and mobile). It’s self-hostable, and it’s one of the solutions I host on my personal VPS.

This month, I’d like to extend my gratitude to the VaultWarden lead developer, Daniel García: my March “Gratitude Subscription” goes to him.

All my gratitude to you, Daniel.

As with any software, I might switch to something else sooner or later. I must be suffering from some kind of sickness. ↩︎

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Put full post in your feeds

lg@mzll.it (Luigi Mozzillo) — Wed, 25 Mar 2026 01:22:22 +0100

If you have a blog and you’re not making any money off it, for crying out loud, just put your full post in the RSS feed – let us read it without having to use some gizmo tool to view it in its entirety. We’ll check out your website anyway, we probably already know it, but – for heaven’s sake! – forget about the stats and just let us read your content. If we read, it’s because we like what you write. Leave that to us.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

We need the illusion that we matter to someone

lg@mzll.it (Luigi Mozzillo) — Sat, 21 Mar 2026 01:22:22 +0100

It’s impressive how writing a blog and feeling responsible for your readers – no matter how little you know them – can make our brains stick to promises we’d normally break without a second thought. Kev Quirk wrote about this recently, discussing his weight and his desire to lose weight.

I found last time that writing about it really helped to motivate me, not to mention hold me to account. It’s easy to eat that chocolate bar if no-one knows I’m going to gain an extra half kilo, but if I’m writing about it publicly, I’m less likely to do so. I dunno, my brain is weird. Anyway, expect monthly updates, just like before.

They are brain flaws, Manuel wrote a few days later.

Anyway, to get back on track, the issue I wanted to discuss is related to disappointment. Specifically, my issue with the concept of disappointing others. […] but in the context of this discussion, the thing that matters is that as long as I’m blogging and I’m sharing my experience, the irrational pressure of disappointing someone keeps me on track. At a rational level, I know that no one gives a fuck if I fail at these silly experiments, and yet, for some reason, that extra pressure is what keeps me in check.

Perhaps the public exposure and the desire not to disappoint those who have set aside a few minutes of their day to read our posts makes us irrationally more motivated. It makes us feel, in a way, “watched” when we stray from the path. Keeping the blog updated on this, therefore, becomes a tool we can use to our advantage. You who read—we who read—even though you’ll view our failure with utter indifference: you’ll turn your presence into a spur.

We need the illusion that we matter to someone in order to stay true to our own transformation.

Back in 2008, I weighed 132 kg (~291 lbs). Too much.

One spring day, a dear cousin of mine stopped by my house, invited me over for coffee, and – somewhat unexpectedly and without my realizing it – took me to see a nutritionist. Surprised, offended, but perhaps also happy to have mentally taken the first step on a new journey, I had an appointment and she put together a diet plan for me. She gave me advice, guidelines, and goals. Within ten months, I lost 40 kg (~88 lbs). Through healthy eating, exercise, and motivation, I managed to reach a goal that had seemed completely unrealistic to me less than a year earlier. I was absolutely thrilled.

When I reached 92 kg (~203 lbs), the nutritionist told me she had nothing more to say, that I could continue on my own, that I had adopted a new lifestyle, and that I no longer needed to come to her office. There was no need. I had reached my goal. She had given me the all-clear.

A few months later, my weight settled in at around 96 kg (~212 lbs), and it stayed that way for a long time.

Over the past couple of years, due to high work-related stress caused by my circumstances rather than my job itself, I’ve gained nearly 10 kg (~22 lbs). That’s still too much. And I can’t seem to lose it.

This post is a promise: I want to give it another try in public, just like Kev. I’ll try to be more active and post regular updates here, so that my fear of letting you down will help me reach my goal.

If I win, we all win. If I lose, I’m the only one who loses.

God damn it, we’re such a mess.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Du-ddu-ddum

lg@mzll.it (Luigi Mozzillo) — Tue, 17 Mar 2026 01:22:22 +0100

Maybe, for the first time in my life, after trying dozens of them and reselling them or sending them back because they hurt or bothered me behind the ears due to the temples of my glasses, maybe I’ve found the over-ear headphones that, in addition to having an exceptional sound for my expectations, don’t hurt behind the ears: the Major IV by Marshall. What a treat. And what a hell of a basses.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Digital DIY

lg@mzll.it (Luigi Mozzillo) — Thu, 12 Mar 2026 01:22:22 +0100

Even the most foolish entrepreneur today thinks he can replace consultants, technicians, lawyers, and accountants with a credit card and a subscription costing just twenty dollars a month.

I wrote a few lines on my Italian blog about why corporate digital bricolage needs to be governed. Otherwise, that idiot will only become even more of a fool.

The problem isn’t technology, but a lack of methodology. In too many companies, IT is still managed “by gut feeling,” but in an era where IT is the nervous system of every process, ignoring governance is silent suicide.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

The Gratitude Subscriptions, February 2026

lg@mzll.it (Luigi Mozzillo) — Thu, 05 Mar 2026 01:22:22 +0100

In February, my gratitude subscription was expressed with a donation ($10) to Johan Sanneblad, lead developer of Notebook Navigator.

Notebook Navigator is an excellent plugin from Obsidian that made me fall in love with Obsidian all over again.

I wrote that Bear was definitive, and in fact he was for a long time. This was primarily for two reasons: because it’s beautiful to use, and because its nested tag system, rather than folders, has always made managing my notes faster, quicker, and more personalized.

Notebook Navigator, among many other things, allows you to manage notes with nested tags, add icons to them, change their colors, add a calendar and customize many other aspects of the daily use of the application and its two-panel interface that, not like Bear but almost, I would even lick Obsidian, right now.

I’ve replaced many plugins with this one, reducing the complexity of my vault. And it’s fast. Both the plugin itself is fast, and Johan’s development is incredibly fast. It has many features, and releases are consistent and rapid. It was one of the most downloaded plugins of 2025, and it deserves my full compliments. It’s also worth a try if you use Obsidian. But even if you don’t.

In short, thanks to Notebook Navigator, I’ve rediscovered the pleasure of using Obsidian again (though I no longer sync it to iCloud) and have significantly reduced its complexity.

All my gratitude to you, Johan.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

7HE INB0X

lg@mzll.it (Luigi Mozzillo) — Sun, 01 Mar 2026 01:22:22 +0100

In my professional experience, I’ve often dealt with corporate communication. Many users don’t know what an email is, don’t know how to write, don’t know how to reply, or – even worse – sometimes don’t even realize they need to reply. The inability to communicate within an organization, both internally and, above all, externally, is a hurdle that absolutely must be overcome.

Over time, I’ve gathered a lot of tips and rules that I’ve jotted down in documents and training presentations. I frequently find myself having to go back over them because, unfortunately, there’s always someone who has no clue what the skill of “managing email properly” actually means.

So, over the last few weeks, I’ve decided to organize everything I’ve gathered and, as I rewrite it, I’m posting it on a new blog.

It’s called The Inbox, it’s spelled 7HE INB0X, and it’s at https://inb0x.it.

It’s in Italian, and I haven’t planned an English version yet. I’m still thinking about it.

These are short, blunt, direct posts, with no frills or flourishes – a bit harsh, a bit philosophical, and not very practical. A minimalist blog: minimal in concept, aesthetics, and content. The goal is to manage to write at least a hundred of them and gather a whole series of topics, tips, and common practices for using email (in general), along with small insights on personal productivity, to eventually produce a small publication.

The nerdiest thing about this whole story is that it’s built with Astro, a static site framework that uses the same logic as Hugo but offers way more possibilities and customization. My blogs are built with the latter, so this was the perfect chance to try it out – I’d been stalking it for quite a while – and learn something new. And I’m glad I did.

That’s all, I hope you like it.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

G pleasure

lg@mzll.it (Luigi Mozzillo) — Sat, 28 Feb 2026 01:22:23 +0100

One thing about AI is that I use it. I use it for work, but also for personal stuff.¹ And there’s nothing wrong with that. I also wrote about it here (in Italian): I don’t remember how to do some things that I now do with a calculator, and sooner or later I’ll probably forget how to do things that I now do with an LLM. All in all, it’s not a big deal. The serious thing is that I had distanced myself from Google.

By paying for a more beefed-up version of Google One to also get Gemini, I went back to using the services included in the subscription and, as a result, fell head over heels back into the arms of Big G. Something I didn’t expect to happen.

My privacy is probably suffering as a result, and I’m disappointed about that. But it’s fucking awesome to use web services that work so well.

I apologize. I promise I’ll think about it again.

I use Google Gemini APIs to summarize or translate content from feeds, for example, but also to build AI-powered automations, or to help me improve the aesthetics of this or other blogs. ↩︎

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

The Gratitude Subscriptions, Jenuary 2026

lg@mzll.it (Luigi Mozzillo) — Fri, 27 Feb 2026 01:22:22 +0100

Over the past two years, I’ve started this my virtual subscription to projects for which I express deep gratitude. Projects related to applications I use, or blogs I follow, or podcasts I listen to, and so on. I wrote about them for 2024 and 2025.

This year too I decided to continue with this habit, but unlike other years I decided to write about it (more or less) monthly, so as to be able to dedicate more space to individual projects.

Let’s start in January, albeit a bit late. And in January, I donated my monthly subscription (€10) to Alexandre Alapetite, aka Alkarex on GitHub. Alexandre is the lead developer of freshRSS, which I’ve already written about extensively and which I’m hosting on my self-hosted server.

For those who don’t know it yet and don’t want to click around the blog, FreshRSS is a simple, lightweight, open-source RSS feed aggregator designed to give you full control over your news sources. It offers a fast, customizable interface. I use it to keep my feeds in a private, secure location.

I use it on a client and not directly on the web. Today on SmartRSS, before that on Unread and before that on Reeder Classic. I use it and it always works. Always.

I use it because it makes me feel less tracked, and I feel like I’m protecting one of the most important things I have: the carefully curated, decades-long selection of my information sources.

All my gratitude to you, Alexandre.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

We're always talking about blogs, anyway

lg@mzll.it (Luigi Mozzillo) — Wed, 25 Feb 2026 01:22:22 +0100

With a toot¹ on Mastodon in which he appreciated this post, my dear friend Nicola made me think about how little I put into promoting my professional blog on these shores.

Lately – due to a work context that, at this stage of my life, I still consider unstable – I’ve been updating it more frequently than in the past. The blog’s content is mostly related to the governance of a company’s IT, managing a team, data and processes, the use of AI in the office, software development, and so on. It’s an Italian-only blog whose very marketing-oriented content, in some ways, was previously on this one².

In short, this is an opportunity to revisit it and, equally, to evaluate whether to share – and translate – that content on this blog, or just keep it there. For now, using an online translator may be sufficient.

If you like them, I’ll publish them. Let me know.

A term that, after a few years and despite the fact that I generally appreciate the platform, now gives me a bit of hives. ↩︎
Before I differentiated “personal blogging” from “professional blogging” ↩︎

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Which are worth a lifetime

lg@mzll.it (Luigi Mozzillo) — Tue, 24 Feb 2026 01:22:22 +0100

When my daughter gets a haircut, she comes home sad and angry because she doesn’t like it, that haircut, and then her mom washes it again and fixes it, and she looks at herself in the mirror and smiles, and then, in bed, under the warm covers, shortly after, happy, she tells me, Dad, I love my hair.

What are the things or moments that, for you, are worth a lifetime?

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Vibe self-hosting

lg@mzll.it (Luigi Mozzillo) — Fri, 06 Feb 2026 01:22:22 +0100

In a recent blog post, Jordan Fulghum noted how, through the use of AI agents via the terminal, the management of self-hosted solutions has radically changed. Today, it’s easier, faster, more convenient, and even more fun to install and configure utility, monitoring, or productivity applications.

The approach Jordan used was to directly run Claude Code on a local machine and have it perform actions to install, configure, and manage open-source software.

I too have implemented a few self-hosted solutions in the last few months, and as I already wrote here, I used LLMs to get support. Calmly, I was saying.

Frankly, I preferred not to grant direct access to the production machines¹ and managed the various tasks manually. It was very easy anyway, and it gave me the ability to manage a lot of software in a clean, fast, and economical way—and also fun, I must admit!—that would have taken me months to do before.

Unlike him, I’ve taken a different approach. Some solutions are on a local machine (a Raspberry Pi 4), others on a remote VPS². Eventually, I’ll try to list the solutions I’ve decided to host on these pages, in addition to FreshRSS, which I’ve already mentioned.

I don’t know if 2026 can realistically be considered the year of self-hosting³. For now, I can say that, if you’re willing to put in a little effort, you can integrate dozens of services, abandon some well-known ones, and bring a bit of complexity in-house. But do you want to consider the satisfaction?

On my MacBooks, however, yes. ↩︎
Hosted by Contabo, a hosting provider based in Munich, Germany, where I have been sourcing professional solutions for over ten years. ↩︎
Some say that, instead, vibe coding could be a brake on open-source. ↩︎

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

Calmness

lg@mzll.it (Luigi Mozzillo) — Thu, 05 Feb 2026 01:22:22 +0100

One thing I wanted to say, regarding my last post. I am not a pessimist, and I do not want to imagine any kind of block on artificial intelligence tools for supporting creative work or development. On the contrary, aware of my past mistakes, having experienced defeat, mortificazione and frustration in my youth, I believe these tools should be used to produce faster and, perhaps, to sell the fruits of one’s labor better. Clearly – needless to say – cleary the results of these tools must be verified, corrected, analyzed, and adjusted before being put into a stable context. Already said. I am optimistic about this, but with calmness.

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.

The quality shop

lg@mzll.it (Luigi Mozzillo) — Sun, 01 Feb 2026 01:22:22 +0100

As a boy I was a freelancer for many years. I built and sold websites. I did it like a craftsman: I designed the layout with Macromedia Fireworks, refined it with lines, shadows, and reflections, and then manually converted it to HTML¹. Then I used my own CMS, built from scratch with my own PHP framework, to make it dynamic.

My work was reasonably priced, and it took me a long time to complete. My customers were satisfied. The quality of the product, even though we were talking about the early 2010s, was evident.

Over time, I met a small web agency that, more or less, sold the same product as me. They built it with Microsoft FrontPage starting from existing templates. They had no technical skills whatsoever. The work didn’t go through any design, coding, or PHP engine development process (sometimes I did it for them). The sites were mostly static; they updated them manually by uploading new pages via FTP. It was already “slop” for all intents and purposes.

To build a website this way, they took at least a fifth of the time it took me. The price was the same. The quality was extremely poor, but no client had ever noticed.

How did it end? I started collaborating closely with the agency because I worked more, producing more, on the low-end product they produced instead of creating my own from scratch.

I don’t want to be melodramatic, but it was clear: quality pays, but only when there’s no other option. If quality can be eroded to move faster and produce more economically, those willing to spend will—almost always—choose the latter option.

Today, over twenty years later, there are LLMs. Software can be produced much faster. The quality of the code? Well, it’s often mediocre without any human intervention. But the seller takes a tenth of the time to produce it. And the customer doesn’t notice the difference, especially if they pay less.

I see myself drawing a gray vector line and then a black one, trying to simulate a shadow under a menu bar. And I imagine my former colleague pressing a button and getting a similar but not identical effect, on one of the worst software programs ever. Who will notice anyway?

Commercially, he had won. Commercially, today, those who can get AI to support them in producing their own processes will win. It’s the law of the market. And unfortunately, the market rarely rewards quality.

The good products will remain on display, priced dearly, in a shop in a place kissed by the sea and frequented only in the summer.

Even for that shit Internet Explorer 6. ↩︎

✱ Thanks for enjoying the RSS feed and reading! You can leave your comments by emailing me.