Videos de Ciencia
Videos sobre avances científicos y descubrimientos.
Ciencia 255 videos
What is an LLM?
An LLM is a model trained on massive amounts of text that learns how words relate to each other. This lesson covers how text becomes tokens, how the model generates a response one token at a time, and why the transformer architecture is what makes modern LLMs so effective. 🖊️ Learning objectives: - What tokens are and why models use them - How auto-regressive next-token prediction works - What the transformer brings to the picture Every token requires a full pass of calculations across billions of parameters. Multiply that by thousands of users sending requests at once and you start to see why inference speed becomes a hard engineering problem. For more resources, you may check out our blog here, where you will find information on: - What is AI inference? Meaning, benefits and how it works - Inference speed or throughput? With RDUs, you don't have to choose #AI #LLM #Tokens #Transformers #SambaNova
Bigger Than Nvidia?? BUY These 3 Embodied AI Stocks Before They EXPLODE -
#aistocks #crypto #cathiewood #growthstocks #investing While everyone is still chasing NVIDIA, the next trillion-dollar opportunity is emerging across humanoid robots, warehouse automation, and AI-powered surgical systems. In this video, we break down 3 of the best AI robotics stocks to buy right now that are positioned to benefit from the explosive growth of embodied artificial intelligence throughout 2026 and beyond. 📈 Stocks Covered ✅ Tesla (TSLA) ✅ Symbotic (SYM) ✅ Intuitive Surgical (ISRG) You'll learn: • Why Embodied AI could become the next multi-trillion-dollar industry • Tesla's Optimus humanoid robot opportunity • Symbotic's massive warehouse automation backlog • Intuitive Surgical's AI-powered robotic surgery leadership • Latest financial results and growth metrics • Competitive advantages and long-term catalysts • Key risks investors should know • Why these companies could outperform during the next AI supercycle As labor shortages increase and AI moves beyond software into the real world, robotics is becoming one of the fastest-growing investment themes of the decade. If you're looking for the best AI stocks, robotics stocks, growth stocks, or long-term investments for 2026, this video is for you. 👍 Like this video if you enjoy AI investing content. 💬 Comment below: Which stock has the biggest upside—Tesla, Symbotic, or Intuitive Surgical? 🔔 Subscribe for more AI stock analysis, growth investing ideas, and market updates every week. AI Stocks, Robotics Stocks, Embodied AI, Tesla Stock, TSLA Stock Analysis, Symbotic Stock, SYM Stock, Intuitive Surgical Stock, ISRG Stock, Best AI Stocks 2026, Best Growth Stocks, Artificial Intelligence Investing, Robot Stocks, Humanoid Robot Stocks, Tesla Optimus, AI Revolution, Future Technology Stocks, Long Term Stocks, AI Infrastructure, AI Investing, Growth Stocks 2026, Cathie Wood AI Stocks, Nvidia Alternatives, Next Nvidia Stock, Warehouse Automation, Medical Robotics, AI Portfolio, Stock Market 2026, Millionaire Stocks, Best Stocks to Buy Now #AIStocks #Robotics #Tesla #Optimus #Symbotic #ISRG #ArtificialIntelligence #StockMarket #GrowthStocks #Investing #HumanoidRobots #AIRevolution #TechStocks #Nvidia #LongTermInvesting #AIStocks #SemiconductorStocks #ArtificialIntelligence #StockMarket #Investing #Micron #Intel #Amkor #GrowthStocks #TechStocks AI stocks, photonics stocks, optical fiber stocks, Nvidia suppliers, AI infrastructure stocks, Corning stock, Amphenol stock, Lumentum stock, Coherent stock, Fabrinet stock, silicon photonics, co-packaged optics, 800G networking, 1.6T networking, AI data centers, hyperscaler stocks, Meta AI, Amazon AI, OpenAI infrastructure, next AI boom stocks, best AI stocks 2026, future technology stocks walmart stock analysis 2026 | costco stock 2026 | starbucks stock turnaround | best retail stocks to buy now | safe dividend stocks may 2026 | consumer stocks portfolio | stocks for recession 2026 stocks under 50, high growth stocks 2026, best stocks to buy 2026, SoFi stock, Rivian stock, UiPath stock, SoundHound AI stock, SOFI, RIVN, PATH, SOUN, AI stocks, EV stocks, fintech stocks, voice AI, robotic process automation, growth stocks under 50, stocks to watch 2026, earnings breakdown, stock market 2026, beginner investing, high upside stocks Cathie Wood stocks 2025, ARK Invest 2025, Cathie Wood latest interview, Cathie Wood Tesla prediction, Cathie Wood Coinbase, Cathie Wood Roku, Tesla stock 2025, Tesla AI stock, Coinbase stock analysis, Roku stock analysis, Best growth stocks 2025, Stocks to buy now October 2025, AI stocks to buy now,ARKK portfolio 2025,Innovation stocks 2025, Stock market analysis,Future tech stocks,High growth stocks 2025, Top stocks 2025,Cathie Wood top 3 stocks,Elon Musk Tesla AI,ARK Invest portfolio update,ARK Invest stocks to buy,Cathie Wood Tesla price target, Coinbase crypto bull run,Streaming stocks 2025,Wall Street news 2025, Investing for beginners,Stock market investing 2025,Cathie Wood 2025 prediction, ARK Invest portfolio, Tesla price target, AI and automation stocks, Coinbase crypto boom, Roku advertising growth, best growth stocks 2025, innovation stocks, stock market analysis, investing strategies for 2025. ⚠️ Disclaimer: The content in this video is for informational and educational purposes only and should not be considered as financial, investment, or legal advice. I am not a licensed financial advisor, and the information shared reflects personal research and opinions. Always do your own due diligence and consult with a qualified financial professional before making any investment decisions. Investing in the stock market involves risk, and past performance is not indicative of future results.
A IA passa um filtro matemático na sua foto. O que ela encontra vai te surpreender #shorts
#aulas #militares #portugues #matematica #cienciassociales "Seu celular te reconhece em 100 milissegundos. Sabe o que acontece por dentro? Aula 8 do curso gratuito de IA do zero — Visão Computacional, parte 3. Professor de Física e Matemática certificado pela Anthropic. Você desbloqueia o celular com o rosto centenas de vezes por semana. Mas sabe o que realmente acontece nesse instante? Não é mágica. São quatro etapas matemáticas: Primeira — detecção: a rede localiza onde está o rosto na imagem, separando do fundo, do cabelo, da roupa. Segunda — alinhamento: ela ajusta rotação, escala e posição pra um padrão fixo. Terceira — extração: mapeia centenas de pontos do seu rosto — distância entre os olhos, largura do nariz, formato da mandíbula. Tudo isso vira um vetor numérico único. A sua assinatura facial. Quarta — comparação: ela compara esse vetor com o que está cadastrado. Se a distância matemática entre os dois for pequena o suficiente — abre. Se não for — não abre. É geometria. É álgebra linear. É Visão Computacional acontecendo em menos tempo do que você pisca os olhos. ━━━━━━━━━━━━━━━━━━━━━━━ 📘 MATERIAL COMPLETO DO CURSO — R$ 17 Resumo das 10 aulas + mapas mentais + glossário com 33 termos + exercícios extras com gabarito + tabela ChatGPT x Claude x Gemini + desafio final. Feito por professor certificado pela Anthropic. Link na bio. Dezessete reais pela base que vai te separar de 95% das pessoas. ━━━━━━━━━━━━━━━━━━━━━━━ 🔔 Segue pra não perder os próximos cortes da Aula 8. #VisaoComputacional #ReconhecimentoFacial #IA #InteligenciaArtificial #CursoGratuito #Anthropic #Professor #FisicaMatematica #FYP
IDL Lect 5A Why Deep Learning Architecture Matters for Explainable AI
In this lecture of INF-8605: Interpretability in Deep Learning, Prof. Dilip K. Prasad explains why deep learning architecture is important for both model performance and model interpretability. Different types of data need different neural network structures. Images, text, time series, graphs, tabular data, and multimodal data all have different patterns. This is why architectures such as MLPs, CNNs, RNNs/LSTMs, Transformers, and Graph Neural Networks are designed in different ways. This lecture explains how architecture affects: what features a model learns how information flows through the model where information is stored which explanation method is suitable how humans can inspect model behavior This lecture builds a foundation for later explainable AI methods such as feature attribution, saliency maps, Grad-CAM, attention analysis, probing, and representation analysis. Key topics covered What is deep learning architecture? Why different architectures exist Architecture and data type Architecture and explanation method Why interpretability depends on architecture MLP, CNN, RNN/LSTM, Transformer, and GNN comparison Architecture-aware interpretability Suggested audience This lecture is suitable for students, researchers, engineers, and AI learners who want to understand deep learning and explainable AI in a structured way. #DeepLearning #Interpretability #ExplainableAI #XAI #NeuralNetworks #MachineLearning #ArtificialIntelligence #CNN #Transformer #GNN
NOTICIAS SEMANALES SOBRE IA. SIN HYPE. NOVEDADES CON IMPACTO ECNONÓMICO, POLÍTICO Y SOCIAL.
Esta semana analizamos las noticias más importantes sobre inteligencia artificial, tecnología y futuro humano: el proyecto de Ley de Sociedades en Argentina, las sociedades automatizadas y las DAOs; el impacto de la IA en los puestos junior; las pérdidas millonarias de OpenAI; el avance de Microsoft Copilot Cowork y Microsoft Scout; el uso de machine learning en el sistema eléctrico de Chile; y el rol de la tecnología en la respuesta ciudadana al terremoto en Venezuela. En este resumen semanal de IA miramos más allá del hype: qué cambia realmente, qué riesgos aparecen y por qué la inteligencia artificial no es solo ChatGPT. Hablamos de automatización, responsabilidad humana, trabajo, regulación, infraestructura, costos económicos y usos sociales de la tecnología. Temas del video: - Argentina: sociedades automatizadas, DAOs e inteligencia artificial - ¿Desaparecen los puestos junior por la IA? - OpenAI, ChatGPT y el costo real de la carrera de la inteligencia artificial - Microsoft Copilot Cowork y Scout: de asistentes a agentes de IA - Machine learning aplicado a la red eléctrica de Chile - Venezuela: tecnología, IA y organización ciudadana ante la emergencia Ayuda a Venezuela: donarseguro.com La IA no es solo innovación. También es contexto, regulación, trabajo, poder económico y responsabilidad humana. Sumate a Instituto Futuro Humano para comprender la inteligencia artificial con criterio, sin hype y con mirada humana. Sumate a nuestra comunidad en skool: https://www.skool.com/ai4good-academy-1487/about Sumate a nuestra comunidad de whatsapp: https://chat.whatsapp.com/B5NzFrEaQvb4aTlWJ6W4ua #InteligenciaArtificial #IA #ChatGPT #OpenAI #MicrosoftCopilot #MachineLearning #Automatización #FuturoDelTrabajo #Tecnología #Argentina #Venezuela #InstitutoFuturoHumano ``` Fuentes que usé para chequear algunos datos actuales: Chequeado sobre la reforma societaria en Argentina, Ed Zitron/FT vía reportes sobre OpenAI, Microsoft sobre Scout y reportes recientes sobre el terremoto en Venezuela.
El Lente Digital: Visión por computadora en la valoración inmobiliaria
Bienvenidos a esta presentación titulada 'El Lente Digital: Visión por computadora en la valoración inmobiliaria'. Durante décadas, la industria de bienes raíces ha dependido de datos planos y bidimensionales para determinar el valor de un activo. Si observamos la parte izquierda de la pantalla, vemos el clásico plano arquitectónico en 2D: líneas, medidas estáticas y distribuciones frías que dejan fuera la esencia real de la propiedad. Sin embargo, estamos presenciando un cambio de paradigma sin precedentes. A la derecha, vemos cómo la Inteligencia Artificial y la visión por computadora transforman ese plano en una realidad tridimensional cuantificable. El 'Lente Digital' no solo ve una imagen; comprende el espacio. Automáticamente segmenta una isla de cocina, calcula su área, identifica que el material es mármol blanco de alta calidad, y detecta un ventanal panorámico midiendo sus dimensiones y orientación. Va incluso más allá, mapeando el flujo de circulación principal dentro del hogar. Hoy exploraremos cómo esta tecnología, impulsada por el análisis de visión artificial, la estimación de valor automatizada y la detección de características espaciales, está eliminando la subjetividad humana. Prepárense para descubrir cómo lo cualitativo se vuelve cuantitativo, y cómo lo invisible a los ojos de un algoritmo tradicional se convierte en el nuevo estándar de precisión para la valoración inmobiliaria del futuro
America Is SHOCKED By China's Army Of Humanoid Bots!
China has unveiled a new generation of advanced humanoid robots, and the world is paying close attention. This shocking development in robotics, artificial intelligence, and automation technology is raising questions about the future of work, warfare, manufacturing, and global competition. From lifelike movements to powerful AI systems, these humanoid bots represent a major step forward in the race for the next era of technology. As China continues to push the boundaries of AI robots and robotics innovation, many are wondering how this could change the balance of power in the future. In this video, we explore China’s humanoid robot army, AI technology, the future of robots, and America’s reaction to this incredible technological breakthrough.
Os Robôs Humanoides Estão Chegando às Nossas Casas Muito Mais Rápido do que o Esperado
Hoje, vamos mergulhar em um mundo onde a robótica humanoide está mudando radicalmente. Será que os robôs podem realmente reproduzir emoções humanas? Quem está liderando essa revolução industrial: a Tesla ou a China? Quais são os desafios éticos e tecnológicos de integrar a inteligência artificial dentro das nossas casas? E, por fim, será que interfaces cérebro-computador, como a Neuralink, vão nos levar em breve a uma nova forma de consciência? Com base nos avanços mais recentes de 2025, descubra como a IA e a robótica estão se tornando protagonistas do nosso dia a dia. ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ 📌 Fontes: https://pastebin.com/raw/GMDdfQxk 📧 Para questões de direitos autorais ou comerciais, envie um e-mail para: atech@clapnetwork.com 📧 Any questions about copyright or commercial, feel free to get in touch via email: atech@clapnetwork.com ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ▫️ ▪️ ✔ Obrigado a você! ©️ ATECH | Todos os direitos reservados
Best AEO Tool: Boost Your ChatGPT & Perplexity SEO (GEO)
https://crowdreply.io?fpr=guylian Start your 7-day free trial of CrowdReply and get your AI Visibility Score today! https://calendly.com/d/cv5z-ctz-x4g/crowdreply-demo Book a 1-on-1 demo call to see how it works for your brand. Most AI visibility tools are either way too expensive or just give you "nice data" without telling you what to do next. In this video, we take you inside the CrowdReply dashboard—the all-in-one tool designed to not only track your AI search visibility but give you the exact, actionable steps you need to move the needle. Whether you want to rank higher in ChatGPT, Perplexity, Gemini, or Google AI Overviews, we'll show you exactly how to track your prompts, reverse-engineer your competitors, and build the right citations to dominate AI search. What We Cover in This Tutorial: LLM Visibility Score: How to instantly see where your brand stands across different Large Language Models (LLMs). Cross-Model Prompt Tracking: Why you need to track user prompts across ChatGPT, Gemini, Perplexity, and more all at once—and how to do it without losing your mind. Competitor Reverse-Engineering: Discover exactly what sources (YouTube, Reddit, Wikipedia, high-authority blogs) the AI is using to recommend your competitors over you. Actionable AEO (Answer Engine Optimization): Stop guessing. Learn how to engage directly with cited Reddit threads and buy strategic backlinks that actually tell the AI to rank your brand. Why Actionable AI SEO Matters: Generative Engine Optimization (GEO) isn't just about knowing your rank; it's about closing the loop. By tracking brand mentions, share of voice, and exact citation sources, you can mimic the strategies of top brands in your niche and position yourself as the #1 alternative. CrowdReply puts everything under one roof so you can stop juggling multiple expensive dashboards. Ready to win in the era of Generative AI? Hit subscribe for more tutorials on AI search strategy, AIO tracking, and the future of digital marketing! Timestamps: 0:00 - The Problem with Most AI SEO Tools 0:42 - CrowdReply Dashboard: Your LLM Visibility Score 1:13 - Tracking Prompts Across Multiple LLMs 1:56 - Competitor Analysis: Stealing Your Rivals' Strategy 2:36 - Actionable Citations & Engaging on Reddit 4:35 - Building the Right Backlinks for AI Search 5:28 - Get Your 7-Day Free Trial! Tags & Keywords: #AISEO #CrowdReply #AISearch #ChatGPTSEO #PerplexitySEO #LLMOptimization #GenerativeEngineOptimization #AnswerEngineOptimization #DigitalMarketingTools
Creé un coach de sentadillas con IA en 5 minutos | Visión computacional explicada
¿Y si tu cámara pudiera ayudarte con los ejercicios? En este video te muestro cómo creé un coach de sentadillas con visión computacional que cuenta tus repeticiones en tiempo real. Sin código complicado — solo la lógica, las tecnologías y cómo funciona detrás de cámaras. 📌 Lo que verás en este video: 00:00 - Introducción: ¿Y si tu cámara pudiera...? 00:15 - Demo en vivo del coach funcionando 00:30 - Cómo la computadora ve tu cuerpo (YOLOv8) 01:38 - La matemática del ángulo de rodilla 02:25 - Los 3 estados (arriba, bajando, abajo) 03:28 - Arquitectura: cámara → cerebro → voz → orquestador 04:19 - Cierre y próximos proyectos 💻 Código y documentación completa: https://github .com/PipetoBlack/coachSentadilla ────────────────────────────── 🔔 Suscríbete para no perderte los próximos laboratorios 📸 Instagram: @pipetolab ────────────────────────────── #VisionComputacional #Python #YOLO #OpenCV #InteligenciaArtificial #MachineLearning #DeepLearning #ComputerVision #PipetoLab #AprendizajeAutomático #IA #TutorialPython #PoseEstimation #Automacion
Muon Explained: Adam's First Real Challenger
AdamW has been the default optimizer for training large neural networks for nearly a decade. But a new optimizer called Muon may be its first serious challenger. In this video, we visually explain how Muon optimizer works, why it is different from Adam and AdamW, and why researchers are paying attention to it for large-scale LLM training. Instead of treating every weight independently, Muon looks at weight matrices geometrically. It orthogonalizes the momentum update, reshapes the singular value spectrum, and pushes training updates across more useful directions. We’ll break down the core idea behind momentum orthogonalization, Newton-Schulz iteration, polar factors, and why Muon can be more compute-efficient than AdamW. We’ll also explain the catch: why vanilla Muon can destabilize attention layers at frontier scale, and how QK-Clip turns it into MuonClip, making it more stable for large language model training. Topics covered: Why AdamW became the default optimizer Adam’s blind spot with matrix weights Singular values, SVD, and matrix geometry How Muon orthogonalizes momentum updates Newton-Schulz iteration explained visually Why Muon can reduce training compute Why attention logits can explode with Muon QK-Clip and MuonClip explained Why Muon matters for future LLM training If you’re interested in LLM training, optimizers, transformer architecture, AdamW, Muon, Newton-Schulz, scaling laws, and frontier AI training, this video gives you a visual explanation of one of the most interesting optimizer ideas in modern deep learning. Muon optimizer Muon explained Muon optimizer explained AdamW vs Muon Adam vs Muon Adam optimizer AdamW optimizer LLM optimizer LLM training optimizer deep learning optimizer neural network optimizer AI optimizer transformer optimizer training large language models LLM training large language model training how LLMs are trained optimizer explained AdamW explained MuonClip MuonClip explained QK-Clip QK Clip explained attention logits attention instability frontier LLM training Kimi K2 Muon Moonlight Muon Moonshot AI Muon Newton Schulz iteration Newton-Schulz iteration momentum orthogonalization orthogonalized momentum polar factor singular value decomposition SVD explained singular values matrix geometry matrix weights fast LLM training compute efficient training training compute scaling laws LLM scaling laws Pareto frontier AI Megatron Core Muon NVIDIA Megatron Core transformer training deep learning explained machine learning explained AI research explained large language models explained modern AI training future of LLMs Adam’s blind spot Muon vs AdamW optimizer for transformers training at scale LLM training stability loss spikes query key clipping QK norm multi head latent attention MLA attention frontier AI models AI training explained
The Math Behind Deepfakes (GANs Explained)
Discover the math behind deepfakes and how Generative Adversarial Networks (GANs) changed AI forever. Before 2014, teaching a machine to create highly realistic data from scratch involved incredibly slow and complex probabilistic calculations. That all changed when Ian Goodfellow and his team introduced a completely new paradigm. By pitting two neural networks against each other in a continuous Minimax game, they bypassed the heavy calculus and unlocked the modern era of generative AI. In this video, we break down the exact mechanics of GANs. You will learn how the Generator acts as a counterfeiter trying to create perfect fakes, while the Discriminator acts as the police trying to catch them. We also dive into the training loop, the significance of the Jensen-Shannon divergence, and the mathematical proof that guarantees these models can perfectly mimic reality. 00:00 - The dense math of early generative models 00:58 - Re-framing generation as an adversarial game 02:27 - The Counterfeiter and the Police analogy 03:11 - The Minimax game and training loop steps 04:24 - Visualizing the push and pull of data distribution 05:58 - Proving the perfect fake mathematically 06:40 - The Jensen-Shannon divergence explained 07:31 - Why GANs matter for modern AI and deepfakes 🔗 Stay Connected 👉 Subscribe on YouTube: https://www.youtube.com/@insightforge_9 👉 Read the Blog (AI, Chatbots & Automation): https://insightforge-ai.blogspot.com/ 👉 Connect on LinkedIn: https://www.linkedin.com/in/mohit-rathod-7991241b5/ 👉 Join the Newsletter: https://www.linkedin.com/newsletters/7330620395449937920/ 👉 Follow on Instagram: https://www.instagram.com/insightforge.ai/ #GenerativeAI #MachineLearning #Deepfakes