Chorouk Malmoum

AI won't replace you, but someone using agents will. I make sure that's you. Founder AgentX Academy | Trusted by +120K on LinkedIn

[Chorouk Malmoum @choroukmalmoum]

Google's new algorithm shrinks 31GB of vectors into 4GB.

Storing embeddings for RAG eats memory fast.

Ten million documents in float32 take 31GB of RAM.

A new open-source Rust vector index fits the same corpus into 4GB. And searches it faster than FAISS.

It runs on TurboQuant, a quantizer that needs zero training.

Each vector gets normalized, randomly rotated, then bucketed by precomputed math.

The result is a fixed pipeline that works on any data.

16x smaller memory footprint
Beats FAISS by 12-20% on ARM
Matches or wins on x86
Filter by ID at search time
Drop-in for LangChain and LlamaIndex

No data-dependent calibration. No rebuilds as your corpus grows.

No managed service touches your vectors.

Pair it with any open embedding model and you get a fully air-gapped RAG stack with Python bindings.

How much further can data-oblivious compression go?

https://t.co/2laJ3FCwsj

[Chorouk Malmoum @choroukmalmoum]

Sending the whole swarm to fix one bug that one of them wrote.

Nobody knows who wrote it.
Everyone thinks they can fix it.
You're paying for every single one

[Chorouk Malmoum @choroukmalmoum]

@ChrisMPerry Flexible on intent and strict on execution. The verify step is the one most teams skip

[Chorouk Malmoum @choroukmalmoum]

@nrqa__ Out of sight, out of mind is the real reason to do lists fail.

[Chorouk Malmoum @choroukmalmoum]

@AymaneMrini0 Glad you found them useful

[Chorouk Malmoum @choroukmalmoum]

You don't need 100 tutorials to learn AI agents.
You need these 8 videos.
Watch them in this order. Bookmark this so you don't lose it.

1. LLM Introduction
https://t.co/4649Fmy9gk

2.LLMs from Scratch
https://t.co/lAJyVJzOiV

3.Agentic AI Overview (Stanford)
https://t.co/akiK2offlo

4.Building and Evaluating Agents
https://t.co/DheicDxRYa

5.Building Effective Agents
https://t.co/0hXFCanUWu

6.Building Agents with MCP
https://t.co/5BzIZGz6Al

7.Building an Agent from Scratch
https://t.co/pZ4lfobUGV

8.Philo Agents
https://t.co/lqzam96UVr

Start with how LLMs work.
Then learn what makes them agents.
Then build one yourself.
That's the whole path.

♻️ Repost to help someone master AI agents
➕ Follow @choroukmalmoum for practical AI + Business insights

[Chorouk Malmoum @choroukmalmoum]

@ChrisMPerry Exactly. There's so much out there that picking the right ones is half the work. Glad this saves you some time

[Chorouk Malmoum @choroukmalmoum]

@nrqa__ Same here. Each video makes the next one easier to understand. That's why the order matters so much

[Chorouk Malmoum @choroukmalmoum]

@Grow_withAI The consistency across all three shots is the real win here

[Chorouk Malmoum @choroukmalmoum]

AI companies every 3 business days:

“We’re not competing.”

Claude: 🏃‍♂️🪢💨

[Chorouk Malmoum @choroukmalmoum]

Everyone wants to work "in AI."
Few can say which layer.
AI is 5 layers deep.
ChatGPT is just the surface.
Skills live in layers, not in hype.

The map 👇

[Chorouk Malmoum @choroukmalmoum]

@SimslearnAi AI