Iván Moreno Mendoza

CTO in Tuno Labs & Trustlayer Foundation. Author of the book "Sintético".

devva devva devva

[Iván Moreno Mendoza @ivvmoreno]

@danellisona Exactly. Effort is the only thing that never betrays you.

[Iván Moreno Mendoza @ivvmoreno]

😂 lol

[Iván Moreno Mendoza @ivvmoreno]

Yes. It’s not just about finishing the product and making money, it’s about finishing what you started, because you began it for a reason. And two (the more important one): it will most likely lead you to something new and better along the way. Ask yourself: how many times have you been working on something that evolved into something better, or from which something completely new emerged?

[Iván Moreno Mendoza @ivvmoreno]

@pfernan95dev Buen trabajo, saludos desde México, paisano.

[Iván Moreno Mendoza @ivvmoreno]

@launch_llama Let me ask your opinion:

Is the idea the product… or is the product the idea?

[Iván Moreno Mendoza @ivvmoreno]

@thesavasx Who is the most interesting person you have connected with?

[Iván Moreno Mendoza @ivvmoreno]

@shirshakchavan Tipical spanish joke sorry hehe

[Iván Moreno Mendoza @ivvmoreno]

@X_supporttop1 Yes

[Iván Moreno Mendoza @ivvmoreno]

300 agents sharing one memory is the right instinct. The next unlock: a store where related memories become neighbors ON WRITE with no LLM linking, no graph pipeline. The edges are just… there, exact, when agent #217 touches what agent #3 wrote. Meaning has gravity. In two weeks i will show and publish repo.

[Iván Moreno Mendoza @ivvmoreno]

@TeslaAIBot Favourite breakfast?

[Iván Moreno Mendoza @ivvmoreno]

Nice orchestration, but it’s missing memory. Every agent wakes up blank, relearns what the last run already knew, and burns tokens re-reading the past.

We’re shipping the fix: shared exact memory for agent fleets. Works the same on Claude, GPT or your local 4B (no LLM ever touches what’s stored). Milliseconds to recall, megabytes to run, and every answer shows exactly what it checked.

In a few weeks we’re launching the open version. Follow me and you’ll see it first.

[Iván Moreno Mendoza @ivvmoreno]

@suraj_sharma14 Let's connect. And anyone who can help me with my posts, messages, or in their threads, I'll be delighted!

[Iván Moreno Mendoza @ivvmoreno]

@cb_doge Looks like two vertical Spaceship

[Iván Moreno Mendoza @ivvmoreno]

@X_supporttop1 Hello

[Iván Moreno Mendoza @ivvmoreno]

Quick question for people who use AI agents daily.

How many agents do you currently have active, and how do you store their long-term "memory" or context?

I want to see how real people are actually solving this.

[Iván Moreno Mendoza @ivvmoreno]

@elonmusk This is peak Planet of the Apes shipwreck energy… and never better said. Amazing.

[Iván Moreno Mendoza @ivvmoreno]

@Saanvi_dhillon Disagree. I'm learning more than ever.

[Iván Moreno Mendoza @ivvmoreno]

@filipscaleson Thanks for the support amid all the memes. What do you think of the concept? Anything that interests you or that you see as a challenge?

[Iván Moreno Mendoza @ivvmoreno]

The memory system of AI agents today is broken.

It’s not a single system. It’s a Frankenstein of disconnected pieces:

Data Ingestion → Chunking + Embeddings → Vector DB + Retrieval + Rerank → Agent Context

Sounds familiar? and the biggest problem isn’t even the cost…

It’s that the memory arrives adulterated.