Bruno Andrade

Founder @ SenseLab - Cognition for AI agents

[Bruno Andrade @balovespizza]

Models are eating tokens at all costs these days

Me: "Hey"

Claude: "I planned the full Hey app, both backend and frontend. I'm building it and will open a PR in the end"

[Bruno Andrade @balovespizza]

Having a fine-tuned model served through a continual learning layer that improves both your model and your agents, all while reducing model costs by ~50%?

Yes, please!

This is how we unlock teams to finally deliver a self-improving agent infrastructure

@SenseLabAI

[Bruno Andrade @balovespizza]

@vakada_rohit @daimacedoa Thanks! Please give it a try! :)

[Bruno Andrade @balovespizza]

Ask your agent something about a project. Ask a teammate's agent the same question. You'll get 𝘁𝘄𝗼 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗮𝗻𝘀𝘄𝗲𝗿𝘀, and both agents will sound completely sure of themselves. 🤔

That's fine with two people. Once everyone on the team has an agent, some on ChatGPT, some on Claude, someone in Cursor, it gets messy fast. 𝗡𝗼𝗻𝗲 𝗼𝗳 𝘁𝗵𝗲𝗺 𝗵𝗮𝘃𝗲 𝗲𝘃𝗲𝗿 𝘁𝗮𝗹𝗸𝗲𝗱 𝘁𝗼 𝗲𝗮𝗰𝗵 𝗼𝘁𝗵𝗲𝗿.

At @SenseLabAI, we built 𝗥𝗼𝗼𝗺𝘀 for this.

Your agent writes down what actually happened, and everyone else's agent reads it and keeps adding to it.

Agents are actually learning together, the way we always wanted 𝗰𝗼𝗻𝘁𝗶𝗻𝘂𝗼𝘂𝘀 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 to work.

[Dai Macedo @daimacedoa]

Big milestone for Bruno @balovespizza and me.
SenseLab is backed by a16z @speedrun

This is our third company together. We've spent years building infrastructure, and we started SenseLab because we found a problem we couldn't stop thinking about.

AI agents generate valuable knowledge every day, but almost none of it becomes shared across agents, teams, or workflows. We started SenseLab to change that.

Huge thanks to the entire a16z @speedrun team for believing in us. Excited for what's ahead.

If you're building AI agent solutions and you've run into challenges around shared knowledge between agents, I'd love to hear about it. DM me.

[Bruno Andrade @balovespizza]

@dimakhanarin @daimacedoa @speedrun Like Marcus say, maybe a bit passed the point of “seasoned” 🤣

[Bruno Andrade @balovespizza]

@hthieblot Yeah, building distributed memory for self-improving agents across models at https://t.co/f22Lmqn5i9

[Bruno Andrade @balovespizza]

It doesn't matter if the team is using @claudeai or @ChatGPT anymore, because @SenseLabAI distributes the knowledge across

It's awesome to be able to leverage whatever model I want for my agents without losing the learnings

[Bruno Andrade @balovespizza]

Aaaaaaand @github is having trouble again

[Bruno Andrade @balovespizza]

@jacobkieser @victorsunderlan @speedrun @ycombinator @EmbrasureAI Congrats! 💪

[Bruno Andrade @balovespizza]

Everyone's building agent memory. Nobody's building agent coordination.
Memory is what one agent remembers. Coordination is what happens when five agents touch the same system, and none of them know what the others just did, and that's the problem that actually breaks production.

[Bruno Andrade @balovespizza]

@azed_ai @veo_tutorials Another option you can have is to use @plixel_studio , especially if you are trying to keep character and location consistency across shorts and your ending credits