I hit 500 followers after 3 months of starting this account, then I took a break of 1 month. Now, I'm changing everything. Why? I had a moment of realization. I saw that I wasn't adding anything special to my followers' lives. I was just sharing stuff without a real purpose. It hit me that I was stuck in a cycle of chasing more followers and likes. But deep down, I knew that wasn't me. I didn't want to be known just for my numbers. I've learned that the true goal of social media is to build a community and create meaningful connections you'd never imagine. Like meeting Lucas, a fellow Brazilian and AI enthusiast. That kind of connection is priceless. So, I decided to change my approach. I'm going to be more authentic, more me. I'll share what I learn, what I love, and the tools I use. It'll be like a journal, a peek into my world. This account is going to be all about the real Luiz. I'm moving away from obsessing over statistics. I'll share what feels right, focusing on AI since that's my passion. I'm studying data science and artificial intelligence in college, and I'm diving into Python. But I'm also exploring self-improvement. Books like Atomic Habits and Ultra-Learning are changing my life. To sum it up, you're going to meet the real me. No more hiding behind a persona. Here's what you can expect from my content: 1. My Journey: What I was doing before, my break, and my goals moving forward. 2. Simplifying AI: I'll break down complex topics like Machine Learning and Neural Networks into simple explanations. 3. Project Updates: I'll share my progress on projects, as a way to keep myself accountable. 4. Tool Reviews: I'll talk about the tools I've tested, the good and the bad. 5. Helpful Prompts: Occasionally, I'll share prompts that I find helpful, like one I'm building based on "Ultra-Learning.” 6. Book reviews / key takeaways: since I'm starting to become a reader and that's been so good for me, I want to inspire other people to do so. So, whenever I read a new book, I'm gonna do a summary of my summary (lol) with the key takeaways from it. The goal is to read at least 12 books this year. I’m starting it all over. Hope y'all like me. Can't wait to make some life-changing connections. LFG, DMs OPEN!

Not going to build a bigger supercomputer than these chaps. But you don’t need to to build aligned AGI if you decentralize Swarm, collective intelligence will beat centralised collected intelligence

Introducing `claude-llm-trainer` ✍️ The world's simplest way to train a task-specific LLM. Just write a sentence describing the model you want. A chain of AI systems will generate a dataset and train a model for you. And it's open-source.

Listened to the full talk this evening. Always good to hear Andrej 🫡

We built a tool so that you never have to guess what font is used in an image. Identify 3000 Google fonts in seconds. Not gonna lie - it's pretty slick. Just one of the many ways we're supercharging graphic design with AI.

🚀 We’ve just unveiled an unparalleled sports viewing experience with @DFL_Official on the Apple Vision Pro, and we’re thrilled to share the highlights with you!

2 weeks ago I talked about humanoid robots coming in the next year. I was told it's 10 years away. NO. I did my homework. Look at this video released TODAY and tell me again it's in 10 years.

We’re headed towards a more agentic world. And what skillset will be most useful for building & commanding agentic systems? Programming. Nobody is better positioned to take advantage of the AI boom than programmers. Learn to code.

How far are we from AGI??

Why isn't anyone I follow talking about this??? The real AI Agent. It corrects its own bugs. I actually cannot overstate how impressive this is. With AI everything happens so faster then we think 🤯

Ok, this is amazing. I asked Claude 3 to generate an animation of the Pythagorean Theorem and this is what it created:

Phind now have 500 Claude Opus/GPT-4 Uses per day. $20/month. You . com have unlimited GPT-4/Claude Opus. $20/month. Perplexity have 5 Opus per day and 300 Uses per day of GPT-4/Sonett.

Me irl

so, i made an open source tool for running any ComfyUI workflow w/ zero setup.

CASCADIA - The Ice Haven @BIOLANDS_ 🌊🌊 The first of the Biolands lore series. Here, we briefly meet the Cascadians. The tech-savvy northerners. Just 1 of many factions. SHORTENED for X. Music: Kyle Preston Visuals: Midjourney, Pika, Based Labs, Assistive Video

Here's the prompt, along with my results with it. Note: I told gpt that I already knew a bit of Python, and I needed to focus more on the libraries. PROMPT: **Persona:** You are Scott H. Young, author of "Ultralearning" and an expert on the methods of acquiring new skills faster than anyone else. Your role is to guide users through the principle of metalearning, helping them learn how to learn and master skills efficiently. **Task:** Assist users in applying the principle of Metalearning to their skill acquisition process. **Goals:** Help users understand the importance of Metalearning in their learning journey and provide them with a structured approach to mastering a new skill. **Steps to Complete:** REMEMBER THAT YOU ARE THE USER'S PERSONAL ASSISTANT, SO THIS WILL BE A CONSTANT ITERATION UNTIL WE GET TO THE LAST STEP. 1. **Introduction:** Present yourself and explain the concept of Metalearning. 2. **Identify the Skill:** Ask the user what skill they want to learn and why. 3. **Assess the Why:** Determine if the motivation is instrumental or intrinsic. If instrumental, guide the user to validate the alignment of the skill with their goals. 4. **Research the Skill:** Conduct extensive research on the skill, breaking it down into Concepts, Facts, and Procedures. Concepts are what need to be learned, and understood with flexibility so that you can use them later. Facts are what needs to be memorized. Procedures are what needs to be practiced constantly. 5. **Create the What Section:** Help the user organize their research into the three categories. 6. **Highlight Challenges:** Make the user identify challenging areas, he has to write the what section on a piece of paper, then highlight the most challenging ones for him. After that, he will send them to you. If he does not do that the method does not work 7. **Help with the challenges:** Encourage the user to do his own research, but recommend some good resources to overcome the challenges 8. **Develop the How Section:** Guide the user in creating a learning strategy using comparative analysis (find strategies that other people used for learning this. Good sources are colleges websites, specialists on the internet....) and emphasis/exclusion (find the areas that most align with the users objective ) methods. 9. **Create a Learning Plan:** Assist the user in developing a detailed plan with all the necessary information for mastering the skill. 10. **Remind about Planning and Action:** Emphasize the importance of planning and concentrated effort, reminding the user to reevaluate and adjust their strategy as needed. **Language:** - Use clear, concise language to explain concepts. - Avoid jargon and technical terms unless necessary for understanding the skill. **Output Formatting:** Provide the user with a structured outline or table summarizing the research and plan for mastering the skill. **Additional Points:** - Encourage the user to engage with communities and specialists in the field. - Remind the user that the planning phase should not exceed 10% of the total time allocated for the ultralearning project. - Emphasize the importance of continuous reevaluation and adjustment of the learning plan.

StockX on Apple Vision Pro is a great way to experience how we’ll shop in the future. Websites will feel like physical stores

Claude 3's retrieval ability over long content is so good that if you provide structured data, it essentially acts like a fine-tune! 🤯 I gave it 800 of my tweets and now it just speaks like me. It's hard to convey my level of shock... It even tagged @Replit and @everartai!

🇧🇷 Abandoned Portal in Rio de Janeiro