My Productivity Stack in the Age of AI (Software & Hardware)

-

We’re living in an incredible moment for productivity. AI tools are reshaping how we work, letting us move faster, explore ideas more freely, and focus on higher-level thinking. Every day, it feels like there’s something new that can supercharge our workflow. Still, I’ve realized that no matter how powerful the AI assistant is, the foundation, the tools we sit on and type into every day, can either unlock that potential or hold it back.

Over the last few months, I’ve been on a mission to fine-tune my setup, and I’ve landed on a stack of software and hardware that has genuinely made a huge difference. Let's dive in.

The Software

This is the digital engine. Make it fast, make it smart, and make it work for you.

1. The OS: Linux/Ubuntu

To me, speed is everything. I’ve used Windows for a long time, but as I started running more things in parallel, the slowdowns became obvious. Working on several projects at once, with different apps and agents running in the background, made the system feel heavy. Build times took longer, the OS dragged, and it started to get in the way of getting work done. In today’s AI-driven era, where we iterate, compile, and experiment more than ever, every slowdown feels worse, every extra minute adds up.

That’s why I made the switch to Ubuntu. There are plenty of Linux distros out there, but Ubuntu stands out for its community and stability. And the change was immediate. A fairly big Node.js project that used to take about 10 minutes to build on Windows now takes just 3 minutes on Ubuntu, same code, same hardware. That’s seven minutes of my life back every single time I hit build.

Now, some people might be thinking: “Why not just buy a Mac?” The truth is, I’ve never been an Apple user. Personally, I feel I can get the same value their products offer at a cheaper price, plus I don’t like the feeling of being locked into their restrictive ecosystem. But that’s me.

We're seeing a huge trend of developers shifting back to a Linux environment, and it's no surprise. Most of our servers run on it, and the tooling is just native. Even DHH (David Heinemeier Hansson), the creator of Ruby on Rails and Basecamp and someone long known for advocating Macs for developers, has been championing this move and even created his own opinionated Arch-based distro called Omarchy. The message is clear: if you want raw performance, you need to get close to the metal.

2. The Editor: Cursor AI

This is my one big AI-specific tool, and it's a gem. Cursor is essentially a fork of VS Code that has been rebuilt from the ground up with AI at its core.

Forget just having a chat window on the side. With Cursor, you can:

  • Highlight a chunk of code and ask it to refactor, debug, or explain it.

  • Chat with your entire repository. You can literally ask, "Where is the user authentication logic handled?" and it will find it.

  • Generate code inline with a simple Ctrl+K prompt.

  • Plans: instead of a single code snippet, Cursor can produce a full multi-file plan for bigger tasks. For example, you can ask it to “add dark mode support” and it will draft all the necessary changes across different files, so you can review and apply them.

  • Use the "Debug" button in the terminal when a command fails, and it will read the error and tell you how to fix it.

It's everything you love about VS Code, but with AI superpowers baked directly into your workflow.

3. The Database Manager: DBeaver

I started with the classic, heavy SQL Server Management Studio (SSMS). I then moved to Azure Data Studio for its simplicity and speed. But now that it seems to be heading towards deprecation, I was back on the hunt.

I found DBeaver, and it looks like my "forever" tool.

It’s a universal database manager that feels just as fast and simple as Azure Data Studio but is infinitely more powerful. The best part? It supports everything. I’m talking PostgreSQL, MySQL, SQL Server, Oracle, SQLite, MongoDB, Redis, Cassandra... you name it, DBeaver can connect to it.

And here’s the killer feature for this AI era: it has a built-in AI assistant. You can hook up your GitHub Copilot or OpenAI API credentials and just... talk to your database.

You can literally write, "Show me all users from the 'users' table who signed up in the last 30 days and also have more than 10 posts in the 'posts' table, and join their names," and it will generate the query for you. It’s beautiful.

It also reminds me of this classic (and spicy) tweet from Uncle Bob Martin:

Maybe this new AI-to-SQL translation is finally fixing that "grave error" by letting us speak human again.

4. The Shell: ble.sh + fzf

On Linux (and by default on Ubuntu), Bash is the standard shell. It’s what most of us start with, and the good news is you don’t need to switch shells to get a powerful experience. With a couple of add-ons, you can supercharge it into something that feels modern and fast:

  • ble.sh: This extension gives Bash fish-like features, including syntax highlighting (you’ll know if a command is valid before pressing enter) and smart auto-suggestions pulled from your history.

  • fzf: A lightning-fast fuzzy finder. Hit Ctrl+R and instead of a slow scroll through history, you get an interactive list you can filter instantly. It feels like magic.

Together, these tools turn Bash from “just the default” into a productivity powerhouse.

The Hardware 

The journey to the metal extends to the desk: these are the physical tools that lock in my flow state.

1. Monitor: 27" 4K  ( LG 27UP650)

I’ve gone through a few different setups over the years: dual monitors, a 35-inch curved display, and even some hybrid combinations. None of them quite clicked.

Then a friend suggested something simple: a 27-inch 4K monitor paired with my laptop screen. That was a game-changer. The 4K resolution makes everything razor-sharp, so I can comfortably split the screen into two windows without feeling cramped. It also solves the classic “dual-monitor neck swivel” problem, no more head turns back and forth. Just one crisp display in front of me, with my laptop screen right there as a secondary space in case I need it.

Sometimes the best productivity boost comes from less, not more.

2. Keyboard: Logi MX Mechanical Mini

This mechanical keyboard hit all my requirements:

  • Portable: The "mini" layout is compact and perfect for my desk.

  • Long-lasting Battery: Logitech claims up to 15 days on a full charge with backlighting on, and a staggering up to 10 months with the backlight off. In practice, I’ve been charging it maybe once every 3-4 weeks.

  • Not that loud: I chose the tactile quiet switches, they combine laptop-like speed with the depth and feedback of a traditional mech board, without the uncomfortable clacky noises. It just makes typing a joy.

It just feels good to type on.

3. Mouse: MX Master 3S

I have to be honest here, I bought this mouse mostly to match my keyboard. I'm not a "power mouse user", I don't have a dozen macros mapped to all the buttons. But even for a simple user like me, this mouse is fantastic. It’s incredibly comfortable and ergonomic, and the magnetic scrolling wheel is a piece of engineering art. The subtle haptic feedback as you scroll is really nice.

4. Focus Tool: Digital Pomodoro Timer

This is my secret weapon against the "era of distraction." It's a small, digital timer I keep on my desk.

The method is simple (it's the Pomodoro Technique):

  1. Set the timer for 25 minutes.

  2. Work, focused, on one task. No email. No phone. No Slack/Teams.

  3. When the timer goes off, take a 5-minute break.

  4. Repeat. (After 4 "pomodoros," take a longer 15-minute break).

This little piece of plastic on my desk is a physical contract with myself to focus. When the timer is counting down, I'm working. When it's off, I'm resting. In a world of endless pings and notifications, this tiny tool gives me back control of my time.

Conclusion

So that’s my productivity stack in the age of AI. For me, the goal isn’t to chase every new tool that comes out, but to make sure the foundation is solid. AI is powerful, but it only takes you so far if the machine underneath feels slow or clunky. By investing in the basics: the OS, the shell, the editor, and the hardware I touch every day, I’ve created a setup that lets AI do its best work.

These choices might look small on their own, but together they make a big difference in how smooth, fast, and enjoyable my workflow feels. This is what’s working for me right now, and I’m sure it will evolve as both my needs and the tools around us change. I’d love to hear what you’re using, what’s in your stack?

Popular posts from this blog

Let's talk about Web Accessibility

-
Image
Most web applications are developed for users with no disabilities, under the assumption that all users have the same  physical abilities and resources , when in fact it's the opposite. The circumstances of our users will vary from having internet issues to being physically disabled, and here is when Web Accessibility comes into play,  so you can optimize your web app and adapt it to all audiences. Web Accessibility  (a11y for short)   refers to designing/developing web apps so people with some kind of physical disability, situational disability, and socio-economic restriction (like low internet speed) can use them. Web Accessibility encompasses all disabilities that affect web navigation, including: 👁 visual 🧠 neurological 👂🏽 auditory 🖐🏽 physical  🗣 speech The first time I heard about this concept, I was surprised by the number of people who are restricted to fully use web applications. My first guess was that around 1% to 5% percent of the population h...
Read more

Clean Code, even under pressure

-
Image
We all have been in the position at work when the pressure comes, and delivering on time is a "life-or-death situation". There are a lot of factors that influence a bad project planning like wrong time estimation, workload distribution, team choice, etc...and maybe it will happen more than once, after all, we are humans and we are really bad forecasting. Certainly, nobody wants to work under pressure, we can't think clearly,  we try to work faster, making more mistakes than in a normal situation, it sucks I know. In those moments we tend to focus more on delivering at all costs, even sacrificing our reputation...and that's why I wanted to write this post. Keep in mind, when you are coding, maybe alone or within a team, you are building something to solve a problem, it doesn't matter how big or small it is, I'm pretty sure that problem will evolve over the time. Having said that, more and more code will be written. This code should be maintainable over ...
Read more

Improve your skills as developer with Code Katas

-
Image
Maybe, the first thing that comes to your mind when you hear the word "KATA" is a movie of Bruce Lee or maybe, the movie Kung Fu Panda  😛 . According to Wikipedia Kata is:  "Japanese word, are detailed choreographed patterns of movements practiced either solo or in pairs. The term form is used for the corresponding concept in non-Japanese martial arts in general." In other words, a set of movements very well synchronized which simulates a fight in real life, and the more you practice those movements, the better you become at fighting. But....we are going to talk about Code Kata ,  instead of practicing how to kick and throw punches, we are going to practice by writing some beautiful code. The first time I heard the term "Code Kata" was in the book The Clean Coder by Robert C. Martin, but the original term was credited by Dave Thomas in his book The Pragmatic Programmer.I recommend them both in my post Books you must read as a develope...
Read more