Recent content on Luca Lanziani
AI Engineering is not just prompting or model selection. It is the design of workflows, platforms, quality systems, evaluations, governance, and delivery practices that make AI-assisted development useful in real organizations. 
 My angle is practical: AI makes coding faster, but organizations still need better specs, platforms, review systems, automation, governance, and delivery ownership to…
Shipping faster should not be your end goal!
Hermes + GPT-5.5, will it work?
AI as an accelerant, shifting left on intent, and the rise of T-shaped roles.
Time to test Hermes.
For the past few months I’ve been hacking on Jarvis in the evenings: a small Go agent that gives an LLM a handful of local tools (read files, run shell, spawn sub-agents, that kind of thing). I’m not trying to ship the next OpenClaw clone. I just wanted to see how the pieces fit together when you build them yourself.
Goodbye GitHub Copilot, Hello Cursor.
JS vs TS while testing the Cursor SDK.
Another Friday night project! For years, I’ve used pass on a remote server, but I never had a simple way to use those passwords in the browser without the usual copy-paste dance.
My main issue with AI is that it is still too dependent on me.
TIL: People are saying it will be harder to learn Software Engineering if you delegate the implementation to Agents.
This is tomorrow and you are still in time to join us.
Two months have passed since my post about “Taste,” and @Catherine Wu just said this in a recent interview:
NVIDIA Jetson Thor 🤔
I’ve canceled my GitHub Copilot subscription and I’m missing it already.
Another 3 days and I’m canceling my annual GitHub Copilot subscription and asking for a refund. 🤷
TIL: I’m taking the “Building with Claude API” course on Anthropic’s Skilljar and just learned a nice trick: using assistant messages and stop_sequences to extract structured data.
Remember when everyone was worried AI would take our jobs? The shift in the last few months toward ‘AI as an amplifier’ is spot on.
Project Jarvis: Why I Decided to Build My Own AI Assistant
Another evening, another feature for Jarvis. Let’s see if we can talk to it over IM.
What better way to spend a couple of hours on Easter Monday than testing the new Gemma-4 with Claude Code via LM Studio?
You can now run Gemma 4 locally on your phone!
My Friday evening was spent hitting a key architectural milestone for “Jarvis,” the agent I’ve been developing on top of my langchain-go (link in the comments) project.
I’m not really an “orchestrate multiple AI agents in parallel” kind of person.
This is about YOUR data and AI.
Hey @Cursor team, I believe there is a bug in the editor regarding model selection.
CAREFUL: You’ll see this and you’ll start testing the new model.
Plan mode is NOT enough! Let me explain 👇
Creating it has never been so easy.
So, you finally caved to the AI hype. Welcome to the messy part.
Watch a swarm of agents in action in less than 10 minutes!
I wrote about this here: https://lnkd.in/dWBbes8G
Please Stop Force-Feeding AI FOMO to Others!
Please Stop Force-Feeding AI FOMO to Others! 
 Not too long ago, I wrote a post about the AI Hype Burnout and how it was harmful to the whole industry. 
 I thought that was the end of it, but it seems like the FOMO is starting to kick in. Many people are seeing that AI is finally delivering real value and are “inviting” everyone to quickly jump on the bandwagon.
This is what I mean when I say we are actively using AI at @Nearform. Our engineers use AI every day!
LLM (Large Lanziani Models)
“Less is more” holds even truer in the AI-assisted coding era.
Everyone can now produce tens of thousands of lines of code.
We’re just 7 hours away! ⏳
What I took away from the 1st Forum on Artificial Intelligence for Industry.
Continuing my series of Friday night experiments, this Friday I was on a train from Milan, returning from the First Italian Forum on Artificial Intelligence for Industry.
Attending the First Italian Forum on Artificial Intelligence for Industry.
Continuing my series of Friday night experiments, this Friday I was on a train from Milan, returning from the First Italian Forum on Artificial Intelligence for Industry. 
 Since the models weren’t working well on the train Wi-Fi, I decided to exercise my coding muscles instead.
“SWE will be replaced by AI in 2-3 years” - Dario Amodei. 
 That’s not what Dario means, read on…
Start developing your “AI taste”.
Yesterday, Kelsey Hightower shared a post from David Crawshaw.
The biggest bottleneck in the AI-Native SDLC isn’t just writing the code. It’s verifying that it actually works.

 I haven’t had this much fun in 10 years! 
 
 This is what a colleague just shared about their experience with SDD. 
 This tech isn’t just transformative in terms of speed; it’s making the work itself incredibly exciting again. 
 There’s no going back! 🚀
Today, we are all talking about AI and the magical things it can do.
We are giving developers Ferrari engines (AI Agents), but is the rest of your SDLC still driving a go-kart? 🏎️💨
Another example of how AI is rewriting the productivity playbook.