---
title: Artificial intelligence videos
seo_title: Artificial intelligence video library
description: A library of short, technically rigorous 5sigmas videos about artificial intelligence, reasoning, multimodality, agents, energy and infrastructure.
keywords: artificial intelligence videos, AI videos, LLM, Transformer, reasoning models, multimodality, AI agents
hide:
- toc
---


<div class="s5-video-library" data-s5-video-library>
  <section class="s5-video-library__intro">
    <div class="s5-eyebrow">Video library</div>
    <h1>One technical idea per video. Full evidence one click away.</h1>
    <p>Explore 46 native-English explanations published on 5sigmas. Every watch page connects the short explanation to its original chapter and related material.</p>
    <div class="s5-video-library__actions">
      <a class="s5-video-library__primary" href="https://5sigmas.com/en/visuales/">Open the visual experience</a>
      <a href="https://5sigmas.com/en/series/">Explore the series</a>
    </div>
  </section>

  <section class="s5-video-library__controls" aria-label="Filter the video library">
    <label><span>Search</span><input type="search" data-s5-video-search placeholder="Topic, concept or series" autocomplete="off"></label>
    <div class="s5-video-library__filters" role="toolbar" aria-label="Video topics"><button type="button" data-s5-video-filter="all" aria-pressed="true">All <span>46</span></button><button type="button" data-s5-video-filter="agents" aria-pressed="false">AI agents <span>6</span></button><button type="button" data-s5-video-filter="security" aria-pressed="false">AI security <span>6</span></button><button type="button" data-s5-video-filter="impact" aria-pressed="false">Economics, energy and well-being <span>5</span></button><button type="button" data-s5-video-filter="foundations" aria-pressed="false">Foundations <span>5</span></button><button type="button" data-s5-video-filter="history" aria-pressed="false">History of AI <span>6</span></button><button type="button" data-s5-video-filter="infrastructure" aria-pressed="false">Infrastructure <span>5</span></button><button type="button" data-s5-video-filter="multimodality" aria-pressed="false">Multimodality <span>6</span></button><button type="button" data-s5-video-filter="reasoning" aria-pressed="false">Reasoning <span>6</span></button><button type="button" data-s5-video-filter="engineering" aria-pressed="false">Systems engineering <span>1</span></button></div>
    <p data-s5-video-status aria-live="polite">46 videos available</p>
  </section>

  <section class="s5-video-library__grid" data-s5-video-grid>
<article class="s5-video-card" data-s5-video-card data-topic="security" data-search="Production Controls Which controls limit damage when an AI system reads external content, uses tools and one defense fails. AI security AI security production LLM security least privilege dual LLM guardrails MCP security tool poisoning agent observability Separate document reading from actions Give every tool only the permissions it needs MCP does not eliminate the problem: it standardizes a new trust boundary">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/seguridad-ia/05-controles-produccion/" aria-label="Watch Production Controls">
    <img src="https://5sigmas.com/en/series/seguridad-ia/05-controles-produccion.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:48</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI security · 0:48</p>
    <h2><a href="https://5sigmas.com/en/videos/series/seguridad-ia/05-controles-produccion/">Production Controls</a></h2>
    <p>Which controls limit damage when an AI system reads external content, uses tools and one defense fails.</p><small>Separate document reading from actions · Give every tool only the permissions it needs</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/seguridad-ia/05-controles-produccion/">Watch video</a><a href="https://5sigmas.com/en/series/seguridad-ia/05-controles-produccion/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="security" data-search="Red Teaming How to test from the document that enters the system to the action it can execute, before an incident occurs. AI security AI security AI red teaming security evaluation human uplift attack budget AI agents LLM benchmarks The threat model comes before the benchmark Separate what the model can do from what the system executes Testing the model and testing the product are different experiments">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/seguridad-ia/04-red-teaming/" aria-label="Watch Red Teaming">
    <img src="https://5sigmas.com/en/series/seguridad-ia/04-red-teaming.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:48</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI security · 0:48</p>
    <h2><a href="https://5sigmas.com/en/videos/series/seguridad-ia/04-red-teaming/">Red Teaming</a></h2>
    <p>How to test from the document that enters the system to the action it can execute, before an incident occurs.</p><small>The threat model comes before the benchmark · Separate what the model can do from what the system executes</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/seguridad-ia/04-red-teaming/">Watch video</a><a href="https://5sigmas.com/en/series/seguridad-ia/04-red-teaming/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="security" data-search="Poisoning What happens when a document or memory preserves a dangerous instruction and the system uses it again later. AI security AI security RAG poisoning agent memory sleeper agents memory poisoning LLM backdoors unlearning Storing a datum does not make it true Persistent memory is already a measurable attack surface Dangerous behavior can also remain hidden in the model">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/seguridad-ia/03-envenenamiento/" aria-label="Watch Poisoning">
    <img src="https://5sigmas.com/en/series/seguridad-ia/03-envenenamiento.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:48</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI security · 0:48</p>
    <h2><a href="https://5sigmas.com/en/videos/series/seguridad-ia/03-envenenamiento/">Poisoning</a></h2>
    <p>What happens when a document or memory preserves a dangerous instruction and the system uses it again later.</p><small>Storing a datum does not make it true · Persistent memory is already a measurable attack surface</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/seguridad-ia/03-envenenamiento/">Watch video</a><a href="https://5sigmas.com/en/series/seguridad-ia/03-envenenamiento/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="security" data-search="Jailbreaks How an attacker tries many ways of asking for the same thing and which controls remain necessary after a refusal. AI security AI security jailbreaks LLM security GCG adversarial suffixes Best-of-N attack budget Refusing a request does not create a perfect boundary Trying many variants changes the cost of the attack The threat model needs an attack budget">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/seguridad-ia/02-jailbreaks/" aria-label="Watch Jailbreaks">
    <img src="https://5sigmas.com/en/series/seguridad-ia/02-jailbreaks.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:48</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI security · 0:48</p>
    <h2><a href="https://5sigmas.com/en/videos/series/seguridad-ia/02-jailbreaks/">Jailbreaks</a></h2>
    <p>How an attacker tries many ways of asking for the same thing and which controls remain necessary after a refusal.</p><small>Refusing a request does not create a perfect boundary · Trying many variants changes the cost of the attack</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/seguridad-ia/02-jailbreaks/">Watch video</a><a href="https://5sigmas.com/en/series/seguridad-ia/02-jailbreaks/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="security" data-search="AI Security A series on how an instruction hidden in a document can influence an AI system, how that risk can persist and which controls constrain actions. AI security AI security AI security prompt injection LLM jailbreak AI agents RAG security OWASP LLM AI red teaming Contents 1. An instruction hidden in a document can change what the system does 2. Asking the model to ignore its limits">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/seguridad-ia/00_presentacion_serie/" aria-label="Watch AI Security">
    <img src="https://5sigmas.com/en/series/seguridad-ia/00_presentacion_serie.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:48</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI security · 0:48</p>
    <h2><a href="https://5sigmas.com/en/videos/series/seguridad-ia/00_presentacion_serie/">AI Security</a></h2>
    <p>A series on how an instruction hidden in a document can influence an AI system, how that risk can persist and which controls constrain actions.</p><small>Contents · 1. An instruction hidden in a document can change what the system does</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/seguridad-ia/00_presentacion_serie/">Watch video</a><a href="https://5sigmas.com/en/series/seguridad-ia/00_presentacion_serie/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="agents" data-search="From demo to production What it takes to run an AI agent in production: budgets, retries, idempotency, observability, asynchronous work, fallbacks, and criteria for not using an agent. AI agents AI agents production AI agents agent observability retries idempotency background tasks agent cost fallbacks conversational runtime Budgets before promises Retries, idempotency, and terminal failures Asynchronous work and honest completion">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/agentes-ia/05-de-la-demo-a-produccion/" aria-label="Watch From demo to production">
    <img src="https://5sigmas.com/en/series/agentes-ia/05-de-la-demo-a-produccion.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:32</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI agents · 0:32</p>
    <h2><a href="https://5sigmas.com/en/videos/series/agentes-ia/05-de-la-demo-a-produccion/">From demo to production</a></h2>
    <p>What it takes to run an AI agent in production: budgets, retries, idempotency, observability, asynchronous work, fallbacks, and criteria for not using an agent.</p><small>Budgets before promises · Retries, idempotency, and terminal failures</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/agentes-ia/05-de-la-demo-a-produccion/">Watch video</a><a href="https://5sigmas.com/en/series/agentes-ia/05-de-la-demo-a-produccion/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="agents" data-search="Agent security Why an agent that reads external data can be manipulated into acting, and which controls reduce risk: instruction separation, least privilege, approval, and auditability. AI agents AI agents AI agent security prompt injection indirect prompt injection agent hijacking agent identity authorization least privilege Direct and indirect prompt injection Authorization must live outside the prompt Least privilege and separation of planes">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/agentes-ia/04-seguridad-agentes/" aria-label="Watch Agent security">
    <img src="https://5sigmas.com/en/series/agentes-ia/04-seguridad-agentes.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:32</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI agents · 0:32</p>
    <h2><a href="https://5sigmas.com/en/videos/series/agentes-ia/04-seguridad-agentes/">Agent security</a></h2>
    <p>Why an agent that reads external data can be manipulated into acting, and which controls reduce risk: instruction separation, least privilege, approval, and auditability.</p><small>Direct and indirect prompt injection · Authorization must live outside the prompt</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/agentes-ia/04-seguridad-agentes/">Watch video</a><a href="https://5sigmas.com/en/series/agentes-ia/04-seguridad-agentes/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="agents" data-search="How to evaluate an AI agent Evaluating an agent requires measuring the whole task, traces, tools, cost, and recovery from failure. A convincing final answer is not enough. AI agents AI agents AI agent evaluation agent evals agent benchmarks traces LLM as judge task success observability The unit of evaluation is a task Four dimensions worth measuring 1. Outcome">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/agentes-ia/03-como-evaluar-un-agente/" aria-label="Watch How to evaluate an AI agent">
    <img src="https://5sigmas.com/en/series/agentes-ia/03-como-evaluar-un-agente.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:32</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI agents · 0:32</p>
    <h2><a href="https://5sigmas.com/en/videos/series/agentes-ia/03-como-evaluar-un-agente/">How to evaluate an AI agent</a></h2>
    <p>Evaluating an agent requires measuring the whole task, traces, tools, cost, and recovery from failure. A convincing final answer is not enough.</p><small>The unit of evaluation is a task · Four dimensions worth measuring</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/agentes-ia/03-como-evaluar-un-agente/">Watch video</a><a href="https://5sigmas.com/en/series/agentes-ia/03-como-evaluar-un-agente/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="agents" data-search="The anatomy of an agent How an AI agent works internally: context, planning, tools, memory, state, and runtime. A tool call is a contract, not magic. AI agents AI agents AI agent anatomy tools tool calling agent memory runtime state MCP ReAct The minimal loop Tool calling: from text to a contract Context, memory, and state are not the same thing">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/agentes-ia/02-anatomia-de-un-agente/" aria-label="Watch The anatomy of an agent">
    <img src="https://5sigmas.com/en/series/agentes-ia/02-anatomia-de-un-agente.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:32</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI agents · 0:32</p>
    <h2><a href="https://5sigmas.com/en/videos/series/agentes-ia/02-anatomia-de-un-agente/">The anatomy of an agent</a></h2>
    <p>How an AI agent works internally: context, planning, tools, memory, state, and runtime. A tool call is a contract, not magic.</p><small>The minimal loop · Tool calling: from text to a contract</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/agentes-ia/02-anatomia-de-un-agente/">Watch video</a><a href="https://5sigmas.com/en/series/agentes-ia/02-anatomia-de-un-agente/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="agents" data-search="What an AI agent is—and is not The difference between a chatbot, workflow, copilot, and agent. An agent is not just an LLM with tools: it is a system that decides actions within explicit boundaries. AI agents AI agents what is an AI agent agentic AI chatbot vs agent workflow tool calling ReAct A scale of autonomy 1. Direct response 2. Deterministic workflow">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/agentes-ia/01-que-es-un-agente/" aria-label="Watch What an AI agent is—and is not">
    <img src="https://5sigmas.com/en/series/agentes-ia/01-que-es-un-agente.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:32</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI agents · 0:32</p>
    <h2><a href="https://5sigmas.com/en/videos/series/agentes-ia/01-que-es-un-agente/">What an AI agent is—and is not</a></h2>
    <p>The difference between a chatbot, workflow, copilot, and agent. An agent is not just an LLM with tools: it is a system that decides actions within explicit boundaries.</p><small>A scale of autonomy · 1. Direct response</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/agentes-ia/01-que-es-un-agente/">Watch video</a><a href="https://5sigmas.com/en/series/agentes-ia/01-que-es-un-agente/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="agents" data-search="AI Agents Five chapters to understand what an AI agent is, how it uses tools, how to evaluate it, which risks it introduces, and what it takes to operate one in production. AI agents AI agents AI agents agentic AI tool calling workflows agent evaluation prompt injection agent identity production AI Contents 1. What an agent is—and is not 2. The anatomy of an agent">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/agentes-ia/00_presentacion_serie/" aria-label="Watch AI Agents">
    <img src="https://5sigmas.com/en/series/agentes-ia/00_presentacion_serie.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:32</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI agents · 0:32</p>
    <h2><a href="https://5sigmas.com/en/videos/series/agentes-ia/00_presentacion_serie/">AI Agents</a></h2>
    <p>Five chapters to understand what an AI agent is, how it uses tools, how to evaluate it, which risks it introduces, and what it takes to operate one in production.</p><small>Contents · 1. What an agent is—and is not</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/agentes-ia/00_presentacion_serie/">Watch video</a><a href="https://5sigmas.com/en/series/agentes-ia/00_presentacion_serie/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="infrastructure" data-search="The real footprint of a data center Water, energy, minerals and data-center lifecycle: what AI consumes, why impact varies by location, and how to measure its footprint. Infrastructure Infrastructure data center water consumption AI environmental footprint AI critical minerals data center PUE data center WUE cobalt AI mining GPU e-waste golf vs data center water local water conflict 1. The comparison that calibrates the conversation 2. Water: withdrawal, consumption and the technology that determines it 3. Energy: the AI premium and what it implies">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/datacenters-espacio/04-huella-real-datacenter/" aria-label="Watch The real footprint of a data center">
    <img src="https://5sigmas.com/en/series/datacenters-espacio/04-huella-real-datacenter.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:53</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Infrastructure · 0:53</p>
    <h2><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/04-huella-real-datacenter/">The real footprint of a data center</a></h2>
    <p>Water, energy, minerals and data-center lifecycle: what AI consumes, why impact varies by location, and how to measure its footprint.</p><small>1. The comparison that calibrates the conversation · 2. Water: withdrawal, consumption and the technology that determines it</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/04-huella-real-datacenter/">Watch video</a><a href="https://5sigmas.com/en/series/datacenters-espacio/04-huella-real-datacenter/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="infrastructure" data-search="What a data center in space actually is What processing data in orbit really means, which use cases make sense today, resilient storage and megaprojects with decades-long horizons. Infrastructure Infrastructure orbital data center satellite computing in-orbit edge processing space-computing maturity Starcloud H100 Axiom Space ODC D-Orbit AIX Three-Body constellation SpaceX orbital data center digital flag state 1. What it means to process data in orbit Processing observation data General-purpose computing in orbit">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/datacenters-espacio/03-que-es-datacenter-espacio/" aria-label="Watch What a data center in space actually is">
    <img src="https://5sigmas.com/en/series/datacenters-espacio/03-que-es-datacenter-espacio.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:53</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Infrastructure · 0:53</p>
    <h2><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/03-que-es-datacenter-espacio/">What a data center in space actually is</a></h2>
    <p>What processing data in orbit really means, which use cases make sense today, resilient storage and megaprojects with decades-long horizons.</p><small>1. What it means to process data in orbit · Processing observation data</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/03-que-es-datacenter-espacio/">Watch video</a><a href="https://5sigmas.com/en/series/datacenters-espacio/03-que-es-datacenter-espacio/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="infrastructure" data-search="Energy, heat and connectivity Why the cold of space does not mean free cooling, where the real advantage of orbital energy lies, and what limits the connection to Earth imposes. Infrastructure Infrastructure orbital data center cooling Stefan-Boltzmann space LEO orbital solar power vacuum space radiators LEO GEO satellite latency GPU radiation degradation satellite downlink bandwidth 1. Why the cold of space does not mean free cooling The real scale of radiators Why in-orbit validation is still limited">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/datacenters-espacio/02-energia-calor-conectividad/" aria-label="Watch Energy, heat and connectivity">
    <img src="https://5sigmas.com/en/series/datacenters-espacio/02-energia-calor-conectividad.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:53</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Infrastructure · 0:53</p>
    <h2><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/02-energia-calor-conectividad/">Energy, heat and connectivity</a></h2>
    <p>Why the cold of space does not mean free cooling, where the real advantage of orbital energy lies, and what limits the connection to Earth imposes.</p><small>1. Why the cold of space does not mean free cooling · The real scale of radiators</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/02-energia-calor-conectividad/">Watch video</a><a href="https://5sigmas.com/en/series/datacenters-espacio/02-energia-calor-conectividad/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="infrastructure" data-search="Why now Why orbital computing is being discussed now. The pressure AI compute demand places on terrestrial infrastructure and the limits that make space relevant. Infrastructure Infrastructure orbital data center space computing AI infrastructure rocket launch cost data center bottlenecks Starship cost per kilogram SpaceX FCC AI compute demand data center water stress 1. The explosion in compute demand 2. Terrestrial bottlenecks Power grid">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/datacenters-espacio/01-por-que-ahora/" aria-label="Watch Why now">
    <img src="https://5sigmas.com/en/series/datacenters-espacio/01-por-que-ahora.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:53</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Infrastructure · 0:53</p>
    <h2><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/01-por-que-ahora/">Why now</a></h2>
    <p>Why orbital computing is being discussed now. The pressure AI compute demand places on terrestrial infrastructure and the limits that make space relevant.</p><small>1. The explosion in compute demand · 2. Terrestrial bottlenecks</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/01-por-que-ahora/">Watch video</a><a href="https://5sigmas.com/en/series/datacenters-espacio/01-por-que-ahora/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="infrastructure" data-search="Data Centers in Space Technical analysis of orbital data centers: heat dissipation, latency, launch cost and real viability compared with terrestrial infrastructure. Infrastructure Infrastructure data centers in space orbital computing AI infrastructure compute energy satellite computing Contents 1. Why now 2. Energy, heat and connectivity">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/datacenters-espacio/00_presentacion_serie/" aria-label="Watch Data Centers in Space">
    <img src="https://5sigmas.com/en/series/datacenters-espacio/00_presentacion_serie.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:53</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Infrastructure · 0:53</p>
    <h2><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/00_presentacion_serie/">Data Centers in Space</a></h2>
    <p>Technical analysis of orbital data centers: heat dissipation, latency, launch cost and real viability compared with terrestrial infrastructure.</p><small>Contents · 1. Why now</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/datacenters-espacio/00_presentacion_serie/">Watch video</a><a href="https://5sigmas.com/en/series/datacenters-espacio/00_presentacion_serie/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="security" data-search="Prompt Injection How an instruction hidden in a document can enter an AI system and which controls separate reading from action. AI security AI security prompt injection LLM security indirect prompt injection RAG security AI agent security dual LLM pattern 1. The problem starts when a document reaches the model 2. The instruction can appear in a document search 3. Filtering words does not separate data from instructions">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/seguridad-ia/01-prompt-injection/" aria-label="Watch Prompt Injection">
    <img src="https://5sigmas.com/en/series/seguridad-ia/01-prompt-injection.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:48</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">AI security · 0:48</p>
    <h2><a href="https://5sigmas.com/en/videos/series/seguridad-ia/01-prompt-injection/">Prompt Injection</a></h2>
    <p>How an instruction hidden in a document can enter an AI system and which controls separate reading from action.</p><small>1. The problem starts when a document reaches the model · 2. The instruction can appear in a document search</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/seguridad-ia/01-prompt-injection/">Watch video</a><a href="https://5sigmas.com/en/series/seguridad-ia/01-prompt-injection/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="engineering" data-search="Proactive and reactive agents and tool calls A technical walkthrough of Reactive / Proactive Agent and the conversational contract it encapsulates: immediate response, asynchronous work, and deferred completion. Systems engineering Systems engineering reactive agent proactive agent tool calls async tool calls conversational runtime background tasks pending updates The problem is not the tool, but time The first turn accepts work; it does not promise results Asynchronous work needs policy, not only background execution">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/articulos-tecnicos/reactive-proactive-agent-header-demo/" aria-label="Watch Proactive and reactive agents and tool calls">
    <img src="https://5sigmas.com/en/articulos-tecnicos/reactive-proactive-agent-header-demo.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:20</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Systems engineering · 0:20</p>
    <h2><a href="https://5sigmas.com/en/videos/articulos-tecnicos/reactive-proactive-agent-header-demo/">Proactive and reactive agents and tool calls</a></h2>
    <p>A technical walkthrough of Reactive / Proactive Agent and the conversational contract it encapsulates: immediate response, asynchronous work, and deferred completion.</p><small>The problem is not the tool, but time · The first turn accepts work; it does not promise results</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/articulos-tecnicos/reactive-proactive-agent-header-demo/">Watch video</a><a href="https://5sigmas.com/en/articulos-tecnicos/proactive-reactive-agent-and-tool-calls/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="reasoning" data-search="Reasoning Risks and Production Controls Overthinking, prompt injection and agent hijacking in reasoning models with tools. Design criteria for bounding risk in production. Reasoning Reasoning AI overthinking reasoning model risks prompt injection TabooRAG agent specification gaming AI alignment responsible AI systems many-shot jailbreaking CoT illegibility agent hijacking 1. Overthinking: when more reasoning degrades the answer 2. Quality vs cost vs latency in a real product The debt profile">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/modelos-razonadores/05-riesgos/" aria-label="Watch Reasoning Risks and Production Controls">
    <img src="https://5sigmas.com/en/series/modelos-razonadores/05-riesgos.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:00</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Reasoning · 1:00</p>
    <h2><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/05-riesgos/">Reasoning Risks and Production Controls</a></h2>
    <p>Overthinking, prompt injection and agent hijacking in reasoning models with tools. Design criteria for bounding risk in production.</p><small>1. Overthinking: when more reasoning degrades the answer · 2. Quality vs cost vs latency in a real product</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/05-riesgos/">Watch video</a><a href="https://5sigmas.com/en/series/modelos-razonadores/05-riesgos/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="reasoning" data-search="Latency, Streaming and Product Design TTFT, streaming and perceived-latency thresholds in reasoning models. RouteLLM, design patterns and production session-cost management. Reasoning Reasoning AI latency TTFT LLM streaming RouteLLM test-time compute product perceived latency AI response time AI SLO speculative decoding 1. Perceived-latency thresholds Dynamic routing: RouteLLM 2. Streaming and the perception of latency">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/modelos-razonadores/04-latencia-streaming/" aria-label="Watch Latency, Streaming and Product Design">
    <img src="https://5sigmas.com/en/series/modelos-razonadores/04-latencia-streaming.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:59</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Reasoning · 0:59</p>
    <h2><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/04-latencia-streaming/">Latency, Streaming and Product Design</a></h2>
    <p>TTFT, streaming and perceived-latency thresholds in reasoning models. RouteLLM, design patterns and production session-cost management.</p><small>1. Perceived-latency thresholds · Dynamic routing: RouteLLM</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/04-latencia-streaming/">Watch video</a><a href="https://5sigmas.com/en/series/modelos-razonadores/04-latencia-streaming/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="reasoning" data-search="Test-Time Compute Test-time compute as a second scaling axis. The three levers—more steps, more candidates and more structure—and their quality, cost and latency tradeoffs in reasoning models. Reasoning Reasoning test-time compute LLM reasoning best-of-N chain of thought PRM ORM scaling laws extended reasoning budget forcing DeepSeek R1 MCTS self-consistency Tree of Thoughts 1. What test-time compute is and why it matters 2. The three levers Lever 1: More internal steps">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/modelos-razonadores/03-test-time-compute/" aria-label="Watch Test-Time Compute">
    <img src="https://5sigmas.com/en/series/modelos-razonadores/03-test-time-compute.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:00</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Reasoning · 1:00</p>
    <h2><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/03-test-time-compute/">Test-Time Compute</a></h2>
    <p>Test-time compute as a second scaling axis. The three levers—more steps, more candidates and more structure—and their quality, cost and latency tradeoffs in reasoning models.</p><small>1. What test-time compute is and why it matters · 2. The three levers</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/03-test-time-compute/">Watch video</a><a href="https://5sigmas.com/en/series/modelos-razonadores/03-test-time-compute/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="impact" data-search="AI and GDP Today Why AI&#x27;s macroeconomic impact takes time to appear in GDP, where it appears earlier, and which signals are most indicative of what is happening. Economics, energy and well-being Economics, energy and well-being AI GDP impact AI productivity J curve generative AI productivity economic growth AI AI adoption signals Brynjolfsson productivity Goldman Sachs AI Acemoglu AI macroeconomics 1. Why macro impact takes time to arrive The four mechanisms behind the lag 2. Where it appears before GDP">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/04-ia-pib-hoy/" aria-label="Watch AI and GDP Today">
    <img src="https://5sigmas.com/en/series/ia-pib-bienestar-energia/04-ia-pib-hoy.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Economics, energy and well-being · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/04-ia-pib-hoy/">AI and GDP Today</a></h2>
    <p>Why AI&#x27;s macroeconomic impact takes time to appear in GDP, where it appears earlier, and which signals are most indicative of what is happening.</p><small>1. Why macro impact takes time to arrive · The four mechanisms behind the lag</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/04-ia-pib-hoy/">Watch video</a><a href="https://5sigmas.com/en/series/ia-pib-bienestar-energia/04-ia-pib-hoy/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="reasoning" data-search="How Reasoning Fails Sycophancy, shortcut learning, specification gaming and cascading failures: the failure modes of reasoning models, how to detect them and how to mitigate them. Reasoning Reasoning reasoning model failures AI sycophancy shortcut learning specification gaming LLM hallucinations reasoning chain systematic LLM errors CoT unfaithfulness 1. Failure types 1.1 Shortcuts — shortcut learning 1.2 Systematic errors">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/modelos-razonadores/02-fallos/" aria-label="Watch How Reasoning Fails">
    <img src="https://5sigmas.com/en/series/modelos-razonadores/02-fallos.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:59</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Reasoning · 0:59</p>
    <h2><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/02-fallos/">How Reasoning Fails</a></h2>
    <p>Sycophancy, shortcut learning, specification gaming and cascading failures: the failure modes of reasoning models, how to detect them and how to mitigate them.</p><small>1. Failure types · 1.1 Shortcuts — shortcut learning</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/02-fallos/">Watch video</a><a href="https://5sigmas.com/en/series/modelos-razonadores/02-fallos/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="impact" data-search="Measurement: GDP vs Well-being Why GDP does not capture real well-being, which dimensions matter most, and when subjective well-being diverges from material well-being. Economics, energy and well-being Economics, energy and well-being GDP well-being beyond GDP human development index HDI Easterlin paradox genuine progress indicator GPI subjective well-being measurement Stiglitz Sen Fitoussi unpaid work economy 1. What GDP measures and what it leaves out What GDP leaves out 2. What well-being measures that GDP does not capture">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/03-pib-vs-bienestar/" aria-label="Watch Measurement: GDP vs Well-being">
    <img src="https://5sigmas.com/en/series/ia-pib-bienestar-energia/03-pib-vs-bienestar.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Economics, energy and well-being · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/03-pib-vs-bienestar/">Measurement: GDP vs Well-being</a></h2>
    <p>Why GDP does not capture real well-being, which dimensions matter most, and when subjective well-being diverges from material well-being.</p><small>1. What GDP measures and what it leaves out · What GDP leaves out</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/03-pib-vs-bienestar/">Watch video</a><a href="https://5sigmas.com/en/series/ia-pib-bienestar-energia/03-pib-vs-bienestar/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="reasoning" data-search="What It Means for an LLM to Reason What reasoning means for a language model, what o1 and DeepSeek R1 added, and why evaluating reasoning requires looking at steps, cost and failure modes. Reasoning Reasoning reasoning models LLM reasoning chain of thought CoT OpenAI o1 test time compute AI reasoning illusion of thinking Apple reasoning paper chains of thought 1. What &quot;reasoning&quot; means for a human 2. What LLMs can do that looks like reasoning 3. The birth of reasoning models: OpenAI o1">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/modelos-razonadores/01-que-es-razonar/" aria-label="Watch What It Means for an LLM to Reason">
    <img src="https://5sigmas.com/en/series/modelos-razonadores/01-que-es-razonar.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:58</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Reasoning · 0:58</p>
    <h2><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/01-que-es-razonar/">What It Means for an LLM to Reason</a></h2>
    <p>What reasoning means for a language model, what o1 and DeepSeek R1 added, and why evaluating reasoning requires looking at steps, cost and failure modes.</p><small>1. What &quot;reasoning&quot; means for a human · 2. What LLMs can do that looks like reasoning</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/01-que-es-razonar/">Watch video</a><a href="https://5sigmas.com/en/series/modelos-razonadores/01-que-es-razonar/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="reasoning" data-search="Reasoning Models Five chapters on how LLMs reason: test-time compute, failure modes, latency, cost and risks when models use tools. Reasoning Reasoning reasoning models test-time compute chain of thought LLM reasoning OpenAI o1 DeepSeek R1 RLVR GRPO sycophancy PRM ORM AI latency TTFT AI overthinking specification gaming budget forcing extended reasoning LLM failures Contents 1. What &quot;reasoning&quot; means for an LLM 2. How these systems fail">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/modelos-razonadores/00_presentacion_serie/" aria-label="Watch Reasoning Models">
    <img src="https://5sigmas.com/en/series/modelos-razonadores/00_presentacion_serie.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:45</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Reasoning · 0:45</p>
    <h2><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/00_presentacion_serie/">Reasoning Models</a></h2>
    <p>Five chapters on how LLMs reason: test-time compute, failure modes, latency, cost and risks when models use tools.</p><small>Contents · 1. What &quot;reasoning&quot; means for an LLM</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/modelos-razonadores/00_presentacion_serie/">Watch video</a><a href="https://5sigmas.com/en/series/modelos-razonadores/00_presentacion_serie/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="impact" data-search="AI as an Electrical Technology What AI implies in compute and energy terms, why demand can grow even as hardware improves, and where the real bottlenecks are. Economics, energy and well-being Economics, energy and well-being AI electrical technology artificial intelligence energy consumption AI data centers energy AI rebound effect GPU TPU compute LLM inference energy data center PUE AI energy bottlenecks 1. What &quot;compute&quot; means in practice Training Inference">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/02-ia-tecnologia-electrica/" aria-label="Watch AI as an Electrical Technology">
    <img src="https://5sigmas.com/en/series/ia-pib-bienestar-energia/02-ia-tecnologia-electrica.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Economics, energy and well-being · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/02-ia-tecnologia-electrica/">AI as an Electrical Technology</a></h2>
    <p>What AI implies in compute and energy terms, why demand can grow even as hardware improves, and where the real bottlenecks are.</p><small>1. What &quot;compute&quot; means in practice · Training</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/02-ia-tecnologia-electrica/">Watch video</a><a href="https://5sigmas.com/en/series/ia-pib-bienestar-energia/02-ia-tecnologia-electrica/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="impact" data-search="Electricity and Well-being Why reliable, affordable electricity enables real gains in health, logistics and industry, and the difference between having kilowatts and having supply quality. Economics, energy and well-being Economics, energy and well-being electricity and well-being energy access development rural electrification energy MTF electricity consumption per capita HDI and energy electricity supply quality energy human development nexus 1. The four main channels Health Logistics and industry">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/01-electricidad-bienestar/" aria-label="Watch Electricity and Well-being">
    <img src="https://5sigmas.com/en/series/ia-pib-bienestar-energia/01-electricidad-bienestar.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Economics, energy and well-being · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/01-electricidad-bienestar/">Electricity and Well-being</a></h2>
    <p>Why reliable, affordable electricity enables real gains in health, logistics and industry, and the difference between having kilowatts and having supply quality.</p><small>1. The four main channels · Health</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/01-electricidad-bienestar/">Watch video</a><a href="https://5sigmas.com/en/series/ia-pib-bienestar-energia/01-electricidad-bienestar/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="impact" data-search="AI, GDP, Well-being and Energy Quantitative analysis of AI&#x27;s impact on energy, productivity and well-being using real World Bank, IEA and Penn World Table data rather than speculative projections. Economics, energy and well-being Economics, energy and well-being AI energy artificial intelligence GDP AI productivity technology well-being AI electricity Contents 1. Electricity → well-being: the real mechanisms 2. AI as an electrical technology">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/00_presentacion_serie/" aria-label="Watch AI, GDP, Well-being and Energy">
    <img src="https://5sigmas.com/en/series/ia-pib-bienestar-energia/00_presentacion_serie.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Economics, energy and well-being · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/00_presentacion_serie/">AI, GDP, Well-being and Energy</a></h2>
    <p>Quantitative analysis of AI&#x27;s impact on energy, productivity and well-being using real World Bank, IEA and Penn World Table data rather than speculative projections.</p><small>Contents · 1. Electricity → well-being: the real mechanisms</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/00_presentacion_serie/">Watch video</a><a href="https://5sigmas.com/en/series/ia-pib-bienestar-energia/00_presentacion_serie/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="multimodality" data-search="Risks of Multimodal AI Systems Visual prompt injection, privacy, context leakage and tool manipulation when a multimodal system can observe and act. Multimodality Multimodality multimodal AI risks visual prompt injection AI system security AI image privacy multimodal LLM attacks safety alignment responsible multimodal AI 1. Visual prompt injection 2. System leakage and tool manipulation 3. Privacy: images, documents and metadata">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/multimodalidad-iag/05-riesgos/" aria-label="Watch Risks of Multimodal AI Systems">
    <img src="https://5sigmas.com/en/series/multimodalidad-iag/05-riesgos.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Multimodality · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/05-riesgos/">Risks of Multimodal AI Systems</a></h2>
    <p>Visual prompt injection, privacy, context leakage and tool manipulation when a multimodal system can observe and act.</p><small>1. Visual prompt injection · 2. System leakage and tool manipulation</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/05-riesgos/">Watch video</a><a href="https://5sigmas.com/en/series/multimodalidad-iag/05-riesgos/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="multimodality" data-search="Evaluating Multimodal Systems How to evaluate multimodal systems without confusing benchmarks with real capability: OCR, audio, grounding, reasoning and metric failures. Multimodality Multimodality multimodal model evaluation multimodal benchmarks OCRBench MMAU VQA MMMU real AI capabilities multimodal LLM evaluation generative AI metrics 1. What it means to evaluate grounding 2. The problem of benchmark contamination 3. Language bias: answering from probability, not evidence">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/multimodalidad-iag/04-evaluacion/" aria-label="Watch Evaluating Multimodal Systems">
    <img src="https://5sigmas.com/en/series/multimodalidad-iag/04-evaluacion.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Multimodality · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/04-evaluacion/">Evaluating Multimodal Systems</a></h2>
    <p>How to evaluate multimodal systems without confusing benchmarks with real capability: OCR, audio, grounding, reasoning and metric failures.</p><small>1. What it means to evaluate grounding · 2. The problem of benchmark contamination</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/04-evaluacion/">Watch video</a><a href="https://5sigmas.com/en/series/multimodalidad-iag/04-evaluacion/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="multimodality" data-search="Multimodal System Architectures Four multimodal architecture families, their differences in quality, cost and latency, and when each way of combining modalities makes sense. Multimodality Multimodality multimodal architectures early fusion late fusion ViT multimodal encoder multimodal decoder LLaVA GPT-4V generative AI architecture vision transformer 1. Visual encoder + connector + language model 2. Fusion through cross-attention 3. Native multimodal tokenization">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/multimodalidad-iag/03-arquitecturas/" aria-label="Watch Multimodal System Architectures">
    <img src="https://5sigmas.com/en/series/multimodalidad-iag/03-arquitecturas.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Multimodality · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/03-arquitecturas/">Multimodal System Architectures</a></h2>
    <p>Four multimodal architecture families, their differences in quality, cost and latency, and when each way of combining modalities makes sense.</p><small>1. Visual encoder + connector + language model · 2. Fusion through cross-attention</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/03-arquitecturas/">Watch video</a><a href="https://5sigmas.com/en/series/multimodalidad-iag/03-arquitecturas/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="multimodality" data-search="Alignment: From Pairs to Interactions How models learn that different signals describe the same content, and why data quality determines the robustness of multimodal alignment. Multimodality Multimodality multimodal alignment CLIP contrastive learning ImageBind multimodal embedding image text representation multimodal training visual instruction tuning 1. Image–text pairs: the foundation and its limits 2. Beyond the pair: aligning multiple modalities 3. Visual instruction: the next level">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/multimodalidad-iag/02-alineamiento/" aria-label="Watch Alignment: From Pairs to Interactions">
    <img src="https://5sigmas.com/en/series/multimodalidad-iag/02-alineamiento.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Multimodality · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/02-alineamiento/">Alignment: From Pairs to Interactions</a></h2>
    <p>How models learn that different signals describe the same content, and why data quality determines the robustness of multimodal alignment.</p><small>1. Image–text pairs: the foundation and its limits · 2. Beyond the pair: aligning multiple modalities</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/02-alineamiento/">Watch video</a><a href="https://5sigmas.com/en/series/multimodalidad-iag/02-alineamiento/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="multimodality" data-search="The Real Problem of Multimodality What it means to integrate text, images, audio and other modalities, and how perception, alignment, reasoning, generation and action fit together. Multimodality Multimodality multimodal AI multimodal systems multimodal LLM text image audio AI multimodal perception multimodal generative AI GPT-4V Gemini multimodal 1. A modality is not only an input type 2. The problem is not adding modalities, but crossing them without destroying them 3. The shared space matters, but it is not the only possible architecture">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/multimodalidad-iag/01-el-problema/" aria-label="Watch The Real Problem of Multimodality">
    <img src="https://5sigmas.com/en/series/multimodalidad-iag/01-el-problema.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Multimodality · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/01-el-problema/">The Real Problem of Multimodality</a></h2>
    <p>What it means to integrate text, images, audio and other modalities, and how perception, alignment, reasoning, generation and action fit together.</p><small>1. A modality is not only an input type · 2. The problem is not adding modalities, but crossing them without destroying them</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/01-el-problema/">Watch video</a><a href="https://5sigmas.com/en/series/multimodalidad-iag/01-el-problema/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="multimodality" data-search="Multimodality in Generative AI What it means to build systems that can perceive, align, reason, generate and act across text, image, audio, video, documents and other signals from the world. Multimodality Multimodality multimodal AI multimodal models CLIP Flamingo Gemini vision language audio text Contents 1. The real problem: what counts as multimodality 2. Alignment: from pairs to interactions">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/multimodalidad-iag/00_presentacion_serie/" aria-label="Watch Multimodality in Generative AI">
    <img src="https://5sigmas.com/en/series/multimodalidad-iag/00_presentacion_serie.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:02</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Multimodality · 1:02</p>
    <h2><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/00_presentacion_serie/">Multimodality in Generative AI</a></h2>
    <p>What it means to build systems that can perceive, align, reason, generate and act across text, image, audio, video, documents and other signals from the world.</p><small>Contents · 1. The real problem: what counts as multimodality</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/multimodalidad-iag/00_presentacion_serie/">Watch video</a><a href="https://5sigmas.com/en/series/multimodalidad-iag/00_presentacion_serie/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="history" data-search="Beyond the Transformer How the field is trying to move beyond pure Transformer scaling by combining tools, search, inference-time memory, world models and robotics. History of AI History of AI beyond the Transformer test-time compute AI memory world models Mamba SSM AI robotics AI agents AI search future of AI 1. Why the Transformer is no longer a complete map 1.1 Truth, uncertainty and hallucination 2. From next-token prediction to search over solution spaces">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/from-cave-to-agi/05-mas-alla/" aria-label="Watch Beyond the Transformer">
    <img src="https://5sigmas.com/en/series/from-cave-to-agi/05-mas-alla.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">History of AI · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/05-mas-alla/">Beyond the Transformer</a></h2>
    <p>How the field is trying to move beyond pure Transformer scaling by combining tools, search, inference-time memory, world models and robotics.</p><small>1. Why the Transformer is no longer a complete map · 1.1 Truth, uncertainty and hallucination</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/05-mas-alla/">Watch video</a><a href="https://5sigmas.com/en/series/from-cave-to-agi/05-mas-alla/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="history" data-search="Learn — from rules to data How AI moved from hand-written rules to learning from data: expert systems, statistics, neural networks and the AlexNet breakthrough. History of AI History of AI machine learning history expert systems MYCIN history of machine learning backpropagation AlexNet AI winters symbolic AI perceptron deep learning history 1. The age of rules: when intelligence was written by hand The first symbolic systems Expert systems: the mature symbolic paradigm">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/from-cave-to-agi/03-aprender/" aria-label="Watch Learn — from rules to data">
    <img src="https://5sigmas.com/en/series/from-cave-to-agi/03-aprender.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">History of AI · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/03-aprender/">Learn — from rules to data</a></h2>
    <p>How AI moved from hand-written rules to learning from data: expert systems, statistics, neural networks and the AlexNet breakthrough.</p><small>1. The age of rules: when intelligence was written by hand · The first symbolic systems</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/03-aprender/">Watch video</a><a href="https://5sigmas.com/en/series/from-cave-to-agi/03-aprender/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="history" data-search="Mechanize — from calculation to computing How humanity automated calculation: from the first physical mechanisms to the separation of program and hardware, and the theoretical foundations of modern computing. History of AI History of AI history of computing Alan Turing Charles Babbage Turing machine ENIAC Von Neumann computer history automated calculation theoretical computing 1. From automating calculations to programming procedures The first calculators: automation is not programming Jacquard: when instructions separated from the mechanism">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/from-cave-to-agi/02-mecanizar/" aria-label="Watch Mechanize — from calculation to computing">
    <img src="https://5sigmas.com/en/series/from-cave-to-agi/02-mecanizar.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">History of AI · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/02-mecanizar/">Mechanize — from calculation to computing</a></h2>
    <p>How humanity automated calculation: from the first physical mechanisms to the separation of program and hardware, and the theoretical foundations of modern computing.</p><small>1. From automating calculations to programming procedures · The first calculators: automation is not programming</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/02-mecanizar/">Watch video</a><a href="https://5sigmas.com/en/series/from-cave-to-agi/02-mecanizar/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="history" data-search="Represent — from counting to calculus How symbols, numbers, algebra and calculus made it possible to represent and manipulate the world, building the mathematical foundation of modern AI. History of AI History of AI history of mathematics symbolic representation algebra differential calculus history of computing Leibniz Newton mathematical notation origins of artificial intelligence 1. We invented languages to describe the world Before writing, we were already counting Numbers, position and zero">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/from-cave-to-agi/01-representar/" aria-label="Watch Represent — from counting to calculus">
    <img src="https://5sigmas.com/en/series/from-cave-to-agi/01-representar.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">History of AI · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/01-representar/">Represent — from counting to calculus</a></h2>
    <p>How symbols, numbers, algebra and calculus made it possible to represent and manipulate the world, building the mathematical foundation of modern AI.</p><small>1. We invented languages to describe the world · Before writing, we were already counting</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/01-representar/">Watch video</a><a href="https://5sigmas.com/en/series/from-cave-to-agi/01-representar/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="history" data-search="From the Caves to AGI An intellectual history of AI: from the first mathematical abstractions to foundation models. Mathematics, philosophy and computing in context. History of AI History of AI artificial intelligence history AI history Transformer foundation models AGI deep learning Contents 1. Represent (≈ 40,000 BCE – 1700) 2. Mechanize (≈ 1700 – 1956)">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/from-cave-to-agi/00_presentacion_serie/" aria-label="Watch From the Caves to AGI">
    <img src="https://5sigmas.com/en/series/from-cave-to-agi/00_presentacion_serie.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">History of AI · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/00_presentacion_serie/">From the Caves to AGI</a></h2>
    <p>An intellectual history of AI: from the first mathematical abstractions to foundation models. Mathematics, philosophy and computing in context.</p><small>Contents · 1. Represent (≈ 40,000 BCE – 1700)</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/00_presentacion_serie/">Watch video</a><a href="https://5sigmas.com/en/series/from-cave-to-agi/00_presentacion_serie/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="foundations" data-search="AGI: Artificial General Intelligence AGI means artificial general intelligence. This chapter explains its definitions, DeepMind&#x27;s and OpenAI&#x27;s levels, and what would still be required to reach it. Foundations Foundations agi artificial general intelligence what is agi agi definition deepmind agi levels superintelligence ai alignment openai agi agi risks 1. The definition problem 2. The definitions in dispute 2.1 The cognitive definition">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/04-agi/" aria-label="Watch AGI: Artificial General Intelligence">
    <img src="https://5sigmas.com/en/series/fundamentos-ia-iag/04-agi.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Foundations · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/04-agi/">AGI: Artificial General Intelligence</a></h2>
    <p>AGI means artificial general intelligence. This chapter explains its definitions, DeepMind&#x27;s and OpenAI&#x27;s levels, and what would still be required to reach it.</p><small>1. The definition problem · 2. The definitions in dispute</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/04-agi/">Watch video</a><a href="https://5sigmas.com/en/series/fundamentos-ia-iag/04-agi/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="foundations" data-search="Classical AI vs Generative AI Technical comparison between classical AI and generative AI: inputs, outputs, determinism, explainability, and when to use rules, ML, LLMs, RAG, or agents. Foundations Foundations classical AI vs generative AI AI differences when to use LLM RAG AI agents AI determinism AI explainability AI vs LLM AI operational matrix 1. The five differences 1.1 Inputs and outputs 1.2 Determinism">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/03-ia-vs-ia-generativa/" aria-label="Watch Classical AI vs Generative AI">
    <img src="https://5sigmas.com/en/series/fundamentos-ia-iag/03-ia-vs-ia-generativa.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Foundations · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/03-ia-vs-ia-generativa/">Classical AI vs Generative AI</a></h2>
    <p>Technical comparison between classical AI and generative AI: inputs, outputs, determinism, explainability, and when to use rules, ML, LLMs, RAG, or agents.</p><small>1. The five differences · 1.1 Inputs and outputs</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/03-ia-vs-ia-generativa/">Watch video</a><a href="https://5sigmas.com/en/series/fundamentos-ia-iag/03-ia-vs-ia-generativa/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="foundations" data-search="What is Generative AI? How generative AI works: from embeddings and the Transformer to foundation models. Scaling laws, LLMOps, and differences between LLMs, RAG, and agents. Foundations Foundations generative AI what is generative AI transformer language models LLM embedding scaling laws GPT foundation models RLHF fine-tuning 1. Embeddings: Translating text into numbers 2. The Transformer: The architecture that changes everything 3. Scaling laws: when quantity becomes quality">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/02-que-es-ia-generativa/" aria-label="Watch What is Generative AI?">
    <img src="https://5sigmas.com/en/series/fundamentos-ia-iag/02-que-es-ia-generativa.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Foundations · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/02-que-es-ia-generativa/">What is Generative AI?</a></h2>
    <p>How generative AI works: from embeddings and the Transformer to foundation models. Scaling laws, LLMOps, and differences between LLMs, RAG, and agents.</p><small>1. Embeddings: Translating text into numbers · 2. The Transformer: The architecture that changes everything</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/02-que-es-ia-generativa/">Watch video</a><a href="https://5sigmas.com/en/series/fundamentos-ia-iag/02-que-es-ia-generativa/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="foundations" data-search="What is AI? What Artificial Intelligence is, how it works and how it evolved: from heuristics and Machine Learning to neural networks and foundation models. Foundations Foundations artificial intelligence what is artificial intelligence machine learning deep learning generative AI neural networks foundation models 1. The General Framework: AI, ML, DL and GenAI 2. How do these systems learn? 3. How these systems are adjusted">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/01-que-es-ia/" aria-label="Watch What is AI?">
    <img src="https://5sigmas.com/en/series/fundamentos-ia-iag/01-que-es-ia.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Foundations · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/01-que-es-ia/">What is AI?</a></h2>
    <p>What Artificial Intelligence is, how it works and how it evolved: from heuristics and Machine Learning to neural networks and foundation models.</p><small>1. The General Framework: AI, ML, DL and GenAI · 2. How do these systems learn?</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/01-que-es-ia/">Watch video</a><a href="https://5sigmas.com/en/series/fundamentos-ia-iag/01-que-es-ia/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="foundations" data-search="AI and Generative AI Foundations Introductory series on AI and generative AI: what they are, how they work, how they differ and what AGI means. For technical professionals and decision-makers. Foundations Foundations artificial intelligence generative AI LLMs AGI machine learning AI foundations Contents 1. What AI is and how it evolved 2. What generative AI actually is">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/00_presentacion_serie/" aria-label="Watch AI and Generative AI Foundations">
    <img src="https://5sigmas.com/en/series/fundamentos-ia-iag/00_presentacion_serie.jpg" alt="" loading="lazy" width="1280" height="720"><span>0:53</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">Foundations · 0:53</p>
    <h2><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/00_presentacion_serie/">AI and Generative AI Foundations</a></h2>
    <p>Introductory series on AI and generative AI: what they are, how they work, how they differ and what AGI means. For technical professionals and decision-makers.</p><small>Contents · 1. What AI is and how it evolved</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/fundamentos-ia-iag/00_presentacion_serie/">Watch video</a><a href="https://5sigmas.com/en/series/fundamentos-ia-iag/00_presentacion_serie/">Read article</a></div>
  </div>
</article>


<article class="s5-video-card" data-s5-video-card data-topic="history" data-search="Scale — deep learning to foundation models A short video explanation of Scale — deep learning to foundation models. History of AI History of AI 1. 2012: when scale stopped being a detail 2. The Transformer and massive pretraining 3. Scale became a methodology">
  <a class="s5-video-card__poster" href="https://5sigmas.com/en/videos/series/from-cave-to-agi/04-escalar/" aria-label="Watch Scale — deep learning to foundation models">
    <img src="https://5sigmas.com/en/series/from-cave-to-agi/04-escalar.jpg" alt="" loading="lazy" width="1280" height="720"><span>1:03</span>
  </a>
  <div class="s5-video-card__body">
    <p class="s5-video-card__meta">History of AI · 1:03</p>
    <h2><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/04-escalar/">Scale — deep learning to foundation models</a></h2>
    <p>A short video explanation of Scale — deep learning to foundation models.</p><small>1. 2012: when scale stopped being a detail · 2. The Transformer and massive pretraining</small>
    <div class="s5-video-card__links"><a href="https://5sigmas.com/en/videos/series/from-cave-to-agi/04-escalar/">Watch video</a><a href="https://5sigmas.com/en/series/from-cave-to-agi/04-escalar/">Read article</a></div>
  </div>
</article>
</section>
  <p class="s5-video-library__empty" data-s5-video-empty hidden>No videos match this search.</p>
</div>