2025 Retrospective

2025 was a year of heavy deep personal challenges that forced me away from the noise of the society and into deep reflection. Time to think about personal life, what’s next, what actually really matters.

Throughout it all, I kept building, refining and learning the same lesson in different ways: systems don’t care about intentions. They care about interfaces, constraints, feedback loops and what happens under pressure.

On AI and hype cycles

AI is everywhere now. I don’t hate it I see its potential, it’s a powerful tool. But I’m concerned about its thoughtless deployment and the familiar pattern playing out again: speculation over substance.

Some years ago cloud, then cybersecurity, now AI. Same playbook. Same “smoke-sellers” riding the hype, same paid-per-ranking “magic quadrants”, same white collars selling bla-bla-bla without knowing or never tried anything, same vendors peddling overpriced garbage that increases complexity instead of solving actual problems.

What’s worse: I see many still stucked in feudal thinking. Everyone protecting their tiny paper kingdom, hoarding knowledge, gatekeeping behind paywalls and certifications. Instead learning, failing, repeating and sharing what works and building on each other’s progress, it’s all about extraction money first, proprietary lock-in second, actual diluted innovation distant third.

It’s exhausting watching this pattern repeat while ignoring what we already have and what we could actually deliver if we stopped treating technology advancement like a zero-sum land grab.

What I built

In 2025, I shipped Net Commander, a network operations module tool focused on removing friction from repeatable tasks. I broke things. I simplified systems that were getting too clever. I kept returning to one principle:

Systems. Tradeoffs. Reality.

I kept building small utilities and workflows because I’m convinced of this: the best automation removes human pain without creating new failure modes.

What I learned

Speed without guardrails scales the blast radius.

The faster you move, the more you need structure, not annoying bureaucracy, smoke-sellers or AI that make coffee, but constraints that prevent runaway failures. This showed up repeatedly: in deployment pipelines, in configuration management, in every system that prioritized speed over resilience and with AI guess? It’s getting worst without structure.

2026

After many years in the industry, I decided it’s time to turn personal notes and long-held ideas into reality. I’m investing time building something meaningful for the open source community that I’ll unveil in 2026, along with some other projects I’m genuinely excited about.

More on that soon.

Thank you

If you read a post, opened an issue, sent a message, or challenged an idean thank you. It matters more than you think.

Happy new year. Tia