Lisbon AI

That’s a wrap on LisbonAI.

Two days. Four blocks. Dozens of demos, ideas, and spontaneous hallway debates that might just turn into the next big thing.

From shipping agents to cooking robots, from SLMs to the art of AI - we saw how intelligence is being built, shared, and

And just like that, 2 days flew by

+300 attendees
Beautiful location in Monsanto Park
Sunny day in Lisbon
Good food, great coffee ☕ - it couldn't be a tech conference without great coffee
Tech talks, live demos, networking with amazing speakers and sponsors
Laptops open and

See you in Porto!

“94% of AI projects fail. Why? Our intuition about AI's capabilities is often wrong.” - @cramforce, Vercel

Malte cuts through the hype, showing builders where to focus for real AI success, shipping agents that work.

“Ship and ship sooner than you feel you're ready. You'll learn so much more from your agent being in production than you ever could in dev.” - @julianeagu, Quotient AI

Julia shared why real-world deployment is your best teacher for AI agents.

“Whimsy is the new utility.” - @noahsolomon, Fal

In a world where every product offers similar utility, the differentiator is delight. How do you cut through the noise? It's not just features, it's about creating feeling.

“The web was built for humans to click, not for agents to act.” - @elliotnorrevik, Lovable

But that's rapidly changing.

"You don't have to have a person actually in the physical room wrangle the robot back in. This is what is going to be happening in robotics." - @VilleKuosmanen, Voyage Robotics.

Human-in-the-loop isn't a fallback; it's the strategy for reliable physical AI. LisbonAI builders

The gap between lab and living room for humanoid robots is still vast.

'From a software perspective, I think all of us realize the gap between where we are and where we need to be for humanoid robotics to actually take form in the home.' - @ShadiElaridi, Swish Robotics

How do you make an entire company's issue tracking less painful?

Dana Lawson from Netlify built an AI agent for Linear. It summarizes, finds patterns, and cuts through duplication, making teams faster.

"We've gone into an over-specialization of our tools. We've really compartmentalized our day-to-day as software engineering product builders." - @davidgomes, Cursor.

The tool sprawl is real.

"The non-determinism doesn't really help because it's hard to predict what they're gonna say when we ask models to do some stuff." - @jaimefjorge, Codacy.

Building with AI agents means grappling with unpredictable outputs. How are you managing this challenge in your projects?

“You obviously cannot learn if you have no memory.”

@sarahwooders from Letta cuts to the core of why current LLM agents struggle to evolve beyond workflows. It's a fundamental limitation many builders are grappling with.

"Canvases are great for collaboration with people. Why shouldn't they be great for collaboration with AI as well?" - @steveruizok, Tldraw

Steve showed how to turn the infinite canvas into a platform for mixed AI-human collaboration, despite AI's visual blind spots.

“A really smart model would divide the task into subtasks and maybe guide the model, a smaller model.”

This isn't just theory. Nikita Khomich from Tzafon showed us how to make AI agents truly intelligent by mastering task decomposition.

"Algorithms and data don't guarantee sustainable differentiation. Your unique 'opinion' does." - Pedro Coelho, Zendesk

“If there is an intention, it is art.” - Luis Clara Gomes, Moullinex.

A powerful take on creativity in the age of AI. Does using an LLM make you less of an artist?

Your AI agent is powerful. But is it safe enough for production?

@0xacb from Ethiack showed how to implement robust safety techniques like guardrails and LLM firewalls.

"Evaluations aren't just a checkbox. They're how you know if your V2 is better than V1, if you've introduced regressions, or if you can save costs with a smaller model."

@clarafrmatos from Sword Health on evaluation-driven LLM development.

Block 5 is live - Engineering, Evals, and Security

Watch here: x.com/i/broadcasts/1…
Also live on youtube: youtube.com/live/bpvf7RzgH…

We’re closing Day 2 with a block built for the hands-on builders, those thinking about efficiency, reliability, and safety in the age of AI.