José Siles | AI | Data

Soy oficialmente LinkedIn Instructor!

A young employee, in his second year at the company, suddenly started standing out. He made fewer mistakes. He contributed more ideas... When I asked him, “What’s changed lately?” I wasn’t expecting his answer ↓↓

A manager’s job isn’t to “manage people”... It’s to create the soil in which people can grow. And I believe the essence of that soil isn’t praise. It’s the act of genuinely paying attention.

He used to make mistakes. And he never really connected with his manager. From what he later explained, he had been thinking: “What’s the point? Nobody is watching anyway.” You don’t need huge applause or extravagant praise to help someone open up. Just as a plant grows toward the light, an employee grows in the direction of the person who truly sees them.

Three years ago, I bought an iPhone 14 and added AppleCare. $6 a month. Without thinking much about it, I paid for three years: $200 in total. Last month, I asked my friend, whose screen was completely shattered: “Do you have AppleCare?” He laughed and said: “No, but I bought it with my credit card, so they’re repairing it for free.” “What?” Friend: “You seriously don’t know about...↓↓”

③ Extra trick: check your home insurance. Many policies include accidental damage coverage that protects your phone even outside the home—and you may already be paying for it. The worst deal is usually carrier insurance: Expensive, high deductibles and full of fine print.

② Cheapest possible option: screen-only insurance. Starting at around $2/month, designed for people who only fear the classic cracked-screen accident.

🚨THIS IS HOW CHINESE CHILDREN LEARN🚨 China has unveiled the first AI-powered blackboard: * Draw a cube, and it instantly turns it into a 3D model, measures the angles and calculates the surface area. * Write an equation... and it graphs it immediately. And it explains every single step. Live. At the World Artificial Intelligence Conference in Shanghai.

🚨STOP USING OTHER PEOPLE’S LLMs. You can build your OWN LLM right now. This repository teaches you how to build your own model from SCRATCH with PyTorch. No APIs. No hype. No pretrained models. It takes you through every step: → Data download → Tokenization → Preprocessing → Attention and Multi-Head Attention → Transformer blocks → GPU training → Text generation → SFT and RLHF All based on the original architecture from: # Attention Is All You Need It includes diagrams and fully explained code to build: → MLP → Single-Head Attention → Multi-Head Attention → Transformer Blocks → The complete Transformer model And the best part: You can train a 13M-parameter model for FREE on Google Colab using a T4 GPU. At that size, it already starts generating coherent sentences with proper grammar. You can also scale up to models with 2B+ parameters if you have access to an A100 or RTX 4090. Basically, it gives you everything you need to go from zero to training your own LLM. I’ll leave the repository in the comments 👇

🔗 Link to the Repo: github.com/FareedKhan-dev…

An iPhone thief was arrested. He stole your phone and unlocked it. Without knowing your passcode. During the interrogation, they offered him a deal: Tell us how you do it, and you’ll avoid prison. “Fine... I don’t need to know everything. Most people leave 6 settings configured WRONG. And when they do… I’m in within seconds.” Here are the 6 👇🏻

@nahid_pro09 not really! The repo has a nice guide :)

@josesilesdata this looks super usefull for trackin newshow hard is install on local pc?

🚨BREAKING: You can hear about major events before almost anyone else. It’s called Crucix, and it monitors the entire world for you. 26 live sources, checked every 15 minutes: 💥 Conflict zones 🚢 Ship tracking ✈️ Real-time flights 📈 Crypto and markets 📊 Economic indicators 🛰️ Satellite-detected wildfires It sends alerts directly to Telegram and Discord. You can reply to it in chat. Type `/brief` and it gives you a summary of everything happening. A 24/7 on-call analyst in your pocket. No cloud. No subscription. No sending your data to anyone. 100% FREE. 100% OPEN-SOURCE. I’ll leave the repo link in the first comment.