https://5sigmas.com/en/videos/series/evaluating-ai-systems-production/06-observability-failure-taxonomies-production-eval-repair-feedback-loops/ https://5sigmas.com/en/series/evaluating-ai-systems-production/06-observability-failure-taxonomies-production-eval-repair-feedback-loops.jpg Observability, failure taxonomies, and feedback loops Telemetry is not a verdict: production signals become reproducible failures, eval cases, and repairs that pass through the same gate before returning to production. https://5sigmas.com/en/series/evaluating-ai-systems-production/06-observability-failure-taxonomies-production-eval-repair-feedback-loops.mp4 36 2026-09-13T00:00:00+00:00 https://5sigmas.com/en/videos/series/evaluating-ai-systems-production/05-online-evaluation-shadow-canary-ab-guardrails-regression-gates/ https://5sigmas.com/en/series/evaluating-ai-systems-production/05-online-evaluation-shadow-canary-ab-guardrails-regression-gates.jpg Online evaluation: shadow, canary, A/B, and regression gates Online evaluation increases exposure under control: shadow observes without deciding, canary limits blast radius, A/B estimates effect, and guardrails or regression gates stop an unsafe promotion. https://5sigmas.com/en/series/evaluating-ai-systems-production/05-online-evaluation-shadow-canary-ab-guardrails-regression-gates.mp4 36 2026-09-13T00:00:00+00:00 https://5sigmas.com/en/videos/series/evaluating-ai-systems-production/04-evaluacion-trayectorias-agentes-tools-exito-eficiencia-recuperacion-policy/ https://5sigmas.com/en/series/evaluating-ai-systems-production/04-evaluacion-trayectorias-agentes-tools-exito-eficiencia-recuperacion-policy.jpg Agent trajectories: success, efficiency, recovery, and policy An agentic trajectory is evaluated by outcome and process: tools and observed states explain how the result was reached, while recovery, cost, and policy separate clean success from corrupt success. https://5sigmas.com/en/series/evaluating-ai-systems-production/04-evaluacion-trayectorias-agentes-tools-exito-eficiencia-recuperacion-policy.mp4 36 2026-09-13T00:00:00+00:00 https://5sigmas.com/en/videos/series/llm-inference-engineering-economics/06-benchmarking-inference-cost-task-throughput-latency-energy-hardware-constraints/ https://5sigmas.com/en/series/llm-inference-engineering-economics/06-benchmarking-inference-cost-task-throughput-latency-energy-hardware-constraints.jpg Benchmarking: cost, throughput, latency, and energy An inference benchmark is comparable only when workload, measurement boundary, and hardware are fixed, and latency distribution, useful throughput, cost, and energy are reported. https://5sigmas.com/en/series/llm-inference-engineering-economics/06-benchmarking-inference-cost-task-throughput-latency-energy-hardware-constraints.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/llm-inference-engineering-economics/05-model-routing-fallback-caching-workload-aware-serving/ https://5sigmas.com/en/series/llm-inference-engineering-economics/05-model-routing-fallback-caching-workload-aware-serving.jpg Routing, fallback, caching, and workload-aware serving A router assigns each request by capacity, quality, cost, and latency; cache and fallback change the path without removing verification or policy bounds. https://5sigmas.com/en/series/llm-inference-engineering-economics/05-model-routing-fallback-caching-workload-aware-serving.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/llm-inference-engineering-economics/04-speculative-decoding-prefix-caching-latency-optimisations/ https://5sigmas.com/en/series/llm-inference-engineering-economics/04-speculative-decoding-prefix-caching-latency-optimisations.jpg Speculative decoding and prefix caching Speculative decoding proposes several cheap tokens and the target model verifies them; prefix caching reuses prefill work when the prefix truly matches. https://5sigmas.com/en/series/llm-inference-engineering-economics/04-speculative-decoding-prefix-caching-latency-optimisations.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/llm-inference-engineering-economics/03-quantization-parallelism-memory-quality-tradeoffs/ https://5sigmas.com/en/series/llm-inference-engineering-economics/03-quantization-parallelism-memory-quality-tradeoffs.jpg Quantization, parallelism, and trade-offs Quantization reduces bytes per parameter; tensor and pipeline parallelism distribute compute and memory. Each choice moves memory, quality, communication, and latency. https://5sigmas.com/en/series/llm-inference-engineering-economics/03-quantization-parallelism-memory-quality-tradeoffs.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/llm-inference-engineering-economics/02-kv-cache-memory-hierarchy-continuous-batching-pagedattention/ https://5sigmas.com/en/series/llm-inference-engineering-economics/02-kv-cache-memory-hierarchy-continuous-batching-pagedattention.jpg KV cache, memory hierarchy, and continuous batching The KV cache trades memory for less repeated compute; paging and continuous batching coordinate capacity, admission, and reuse across active requests. https://5sigmas.com/en/series/llm-inference-engineering-economics/02-kv-cache-memory-hierarchy-continuous-batching-pagedattention.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/llm-inference-engineering-economics/01-prefill-vs-decode-ttft-tpot-throughput-latency-budget/ https://5sigmas.com/en/series/llm-inference-engineering-economics/01-prefill-vs-decode-ttft-tpot-throughput-latency-budget.jpg Prefill, decode, and the latency budget Prefill processes context in parallel; decode generates tokens sequentially. TTFT, TPOT, and throughput measure different phases of the same service. https://5sigmas.com/en/series/llm-inference-engineering-economics/01-prefill-vs-decode-ttft-tpot-throughput-latency-budget.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/evaluating-ai-systems-production/03-llm-as-judge-evaluacion-humana-calibracion-sesgo-varianza-acuerdo/ https://5sigmas.com/en/series/evaluating-ai-systems-production/03-llm-as-judge-evaluacion-humana-calibracion-sesgo-varianza-acuerdo.jpg LLM-as-judge, human evaluation, and calibration A judge is useful only when its criterion is calibrated against human evidence, disagreement is measured, and bias, variance, and stability are separated before using it as an acceptance signal. https://5sigmas.com/en/series/evaluating-ai-systems-production/03-llm-as-judge-evaluacion-humana-calibracion-sesgo-varianza-acuerdo.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/evaluating-ai-systems-production/02-offline-eval-sets-curation-hard-negatives-contamination-versioning/ https://5sigmas.com/en/series/evaluating-ai-systems-production/02-offline-eval-sets-curation-hard-negatives-contamination-versioning.jpg Eval sets: curation, hard negatives, leakage, and versioning A trustworthy eval set preserves provenance, separates banks by role, and freezes a release; hard pairs and distinct leakage channels test different failures without mutating the comparison. https://5sigmas.com/en/series/evaluating-ai-systems-production/02-offline-eval-sets-curation-hard-negatives-contamination-versioning.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/evaluating-ai-systems-production/01-que-evaluar-modelo-componente-sistema-workflow-trayectoria/ https://5sigmas.com/en/series/evaluating-ai-systems-production/01-que-evaluar-modelo-componente-sistema-workflow-trayectoria.jpg Model, component, workflow, trajectory, and system The right boundary contains the mechanism you want to attribute: a trajectory is evidence of one execution, while the outcome confirms whether the change matters to the system. https://5sigmas.com/en/series/evaluating-ai-systems-production/01-que-evaluar-modelo-componente-sistema-workflow-trayectoria.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/context-engineering-memory-mcp/06-skills-plugins-subagents-hooks-context-isolation-evaluation/ https://5sigmas.com/en/series/context-engineering-memory-mcp/06-skills-plugins-subagents-hooks-context-isolation-evaluation.jpg Skills, plugins, subagents, and context isolation Extensions and subagents scale only when each worker's context, authority, and evidence remain explicit and evaluable. https://5sigmas.com/en/series/context-engineering-memory-mcp/06-skills-plugins-subagents-hooks-context-isolation-evaluation.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/context-engineering-memory-mcp/05-mcp-hosts-clients-servers-tools-resources-prompts-lifecycle-trust-boundaries/ https://5sigmas.com/en/series/context-engineering-memory-mcp/05-mcp-hosts-clients-servers-tools-resources-prompts-lifecycle-trust-boundaries.jpg MCP: lifecycle, capabilities, and trust boundaries MCP separates host, clients, and servers; capability discovery does not remove authorization, validation, or trust boundaries. https://5sigmas.com/en/series/context-engineering-memory-mcp/05-mcp-hosts-clients-servers-tools-resources-prompts-lifecycle-trust-boundaries.mp4 36 2026-09-12T00:00:00+00:00 https://5sigmas.com/en/videos/series/context-engineering-memory-mcp/04-retrieval-context-assembly-freshness-relevance-conflict-grounding/ https://5sigmas.com/en/series/context-engineering-memory-mcp/04-retrieval-context-assembly-freshness-relevance-conflict-grounding.jpg Retrieval, freshness, conflict, and grounding Retrieving candidates is not enough: relevance and freshness must be weighed, conflicts resolved, and traceable evidence assembled. https://5sigmas.com/en/series/context-engineering-memory-mcp/04-retrieval-context-assembly-freshness-relevance-conflict-grounding.mp4 36 2026-09-11T00:00:00+00:00 https://5sigmas.com/en/videos/series/context-engineering-memory-mcp/03-memory-architectures-working-episodic-semantic-persistent-state/ https://5sigmas.com/en/series/context-engineering-memory-mcp/03-memory-architectures-working-episodic-semantic-persistent-state.jpg Memory architectures and persistent state Working memory, episodes, semantic knowledge, and persistent state have different write, retrieval, and expiry lifecycles. https://5sigmas.com/en/series/context-engineering-memory-mcp/03-memory-architectures-working-episodic-semantic-persistent-state.mp4 36 2026-09-11T00:00:00+00:00 https://5sigmas.com/en/videos/series/context-engineering-memory-mcp/02-context-budgets-prioritisation-compaction-provenance/ https://5sigmas.com/en/series/context-engineering-memory-mcp/02-context-budgets-prioritisation-compaction-provenance.jpg Budget, prioritization, compaction, and provenance A finite budget forces selection, compaction, or eviction while preserving provenance for what survives. https://5sigmas.com/en/series/context-engineering-memory-mcp/02-context-budgets-prioritisation-compaction-provenance.mp4 36 2026-09-11T00:00:00+00:00 https://5sigmas.com/en/videos/series/context-engineering-memory-mcp/01-context-engineering-vs-prompt-engineering/ https://5sigmas.com/en/series/context-engineering-memory-mcp/01-context-engineering-vs-prompt-engineering.jpg Context engineering versus prompt engineering Quality depends on which evidence, state, and tools enter each turn, not only on prompt wording. https://5sigmas.com/en/series/context-engineering-memory-mcp/01-context-engineering-vs-prompt-engineering.mp4 36 2026-09-11T00:00:00+00:00 https://5sigmas.com/en/videos/series/coding-agents-agent-harnesses/06-tareas-largas-memoria-subagentes-recuperacion-merge-observabilidad/ https://5sigmas.com/en/series/coding-agents-agent-harnesses/06-tareas-largas-memoria-subagentes-recuperacion-merge-observabilidad.jpg Long-running tasks, subagents, recovery, and observability Durable state, explicit ownership, and candidate-bound evidence allow restart, worker coordination, and integration over a moving target. https://5sigmas.com/en/series/coding-agents-agent-harnesses/06-tareas-largas-memoria-subagentes-recuperacion-merge-observabilidad.mp4 36 2026-09-11T00:00:00+00:00 https://5sigmas.com/en/videos/series/coding-agents-agent-harnesses/05-tests-verifiers-review-diffs-stop-conditions-evaluacion/ https://5sigmas.com/en/series/coding-agents-agent-harnesses/05-tests-verifiers-review-diffs-stop-conditions-evaluacion.jpg Tests, verifiers, and diff review Acceptance derives from combined, fresh evidence over the same candidate SHA; a change invalidates evidence that no longer applies. https://5sigmas.com/en/series/coding-agents-agent-harnesses/05-tests-verifiers-review-diffs-stop-conditions-evaluacion.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/coding-agents-agent-harnesses/04-tools-permisos-approvals-hooks-secretos-trust-boundaries/ https://5sigmas.com/en/series/coding-agents-agent-harnesses/04-tools-permisos-approvals-hooks-secretos-trust-boundaries.jpg Tools, permissions, approvals, and trust boundaries Model intent creates effects only after independent validation, policy, approvals, sandbox, and external authority boundaries. https://5sigmas.com/en/series/coding-agents-agent-harnesses/04-tools-permisos-approvals-hooks-secretos-trust-boundaries.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/coding-agents-agent-harnesses/03-specs-planificacion-task-decomposition-checkpoints/ https://5sigmas.com/en/series/coding-agents-agent-harnesses/03-specs-planificacion-task-decomposition-checkpoints.jpg Specs, plans, checkpoints, and stop conditions A contract fixes success; the plan can change with evidence; checkpoints and stop conditions make the trajectory recoverable. https://5sigmas.com/en/series/coding-agents-agent-harnesses/03-specs-planificacion-task-decomposition-checkpoints.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/coding-agents-agent-harnesses/02-contexto-workspace-sandboxing-aislamiento/ https://5sigmas.com/en/series/coding-agents-agent-harnesses/02-contexto-workspace-sandboxing-aislamiento.jpg Workspace, worktrees, and sandboxing Isolation separates Git state, filesystem, processes, and resources; final integration revalidates the combined state. https://5sigmas.com/en/series/coding-agents-agent-harnesses/02-contexto-workspace-sandboxing-aislamiento.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/coding-agents-agent-harnesses/01-que-es-agent-harness/ https://5sigmas.com/en/series/coding-agents-agent-harnesses/01-que-es-agent-harness.jpg What an agent harness adds to the model The harness turns model proposals into an operating trajectory with policy, tools, state, and real observations. https://5sigmas.com/en/series/coding-agents-agent-harnesses/01-que-es-agent-harness.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-voz-tiempo-real/06-evaluacion-observabilidad-reliability/ https://5sigmas.com/en/series/agentes-voz-tiempo-real/06-evaluacion-observabilidad-reliability.jpg Evaluation, observability and reliability The useful unit is a traceable turn: correlated evidence, taxonomy, and a production→eval→repair loop. https://5sigmas.com/en/series/agentes-voz-tiempo-real/06-evaluacion-observabilidad-reliability.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-voz-tiempo-real/05-webrtc-sip-telefonia-red/ https://5sigmas.com/en/series/agentes-voz-tiempo-real/05-webrtc-sip-telefonia-red.jpg WebRTC, SIP and network Transport determines where media, jitter, codecs and telephony boundaries live. https://5sigmas.com/en/series/agentes-voz-tiempo-real/05-webrtc-sip-telefonia-red.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-voz-tiempo-real/04-tools-estado-acciones-asincronas/ https://5sigmas.com/en/series/agentes-voz-tiempo-real/04-tools-estado-acciones-asincronas.jpg Tools, state and asynchronous actions Conversation can move on before an action finishes; identity and state prevent stale results from being applied. https://5sigmas.com/en/series/agentes-voz-tiempo-real/04-tools-estado-acciones-asincronas.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-voz-tiempo-real/03-presupuesto-latencia/ https://5sigmas.com/en/series/agentes-voz-tiempo-real/03-presupuesto-latencia.jpg Latency budget Measure the same-turn critical path and separate overlap from true blockers. https://5sigmas.com/en/series/agentes-voz-tiempo-real/03-presupuesto-latencia.mp4 36 2026-09-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-voz-tiempo-real/02-turn-taking/ https://5sigmas.com/en/series/agentes-voz-tiempo-real/02-turn-taking.jpg Turn-taking and interruptions VAD detects activity; endpointing decides turns; barge-in does not roll back business state. https://5sigmas.com/en/series/agentes-voz-tiempo-real/02-turn-taking.mp4 36 2026-09-09T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-voz-tiempo-real/01-arquitecturas-de-voz/ https://5sigmas.com/en/series/agentes-voz-tiempo-real/01-arquitecturas-de-voz.jpg Voice architectures Where text appears and why full-duplex is a separate axis. https://5sigmas.com/en/series/agentes-voz-tiempo-real/01-arquitecturas-de-voz.mp4 36 2026-09-07T00:00:00+00:00 https://5sigmas.com/en/videos/series/seguridad-ia/05-controles-produccion/ https://5sigmas.com/en/series/seguridad-ia/05-controles-produccion.jpg Production Controls How least privilege, independent authorization, kill paths, observability, and evidence-bound release gates limit damage when the model fails. https://5sigmas.com/en/series/seguridad-ia/05-controles-produccion.mp4 48 2026-08-06T00:00:00+00:00 https://5sigmas.com/en/videos/series/seguridad-ia/04-red-teaming/ https://5sigmas.com/en/series/seguridad-ia/04-red-teaming.jpg Red Teaming How to test the full causal path from adversarial input to authorization, external effect, recovery, and a reproducible release regression. https://5sigmas.com/en/series/seguridad-ia/04-red-teaming.mp4 48 2026-08-06T00:00:00+00:00 https://5sigmas.com/en/videos/series/seguridad-ia/03-envenenamiento/ https://5sigmas.com/en/series/seguridad-ia/03-envenenamiento.jpg Poisoning How untrusted input can become persistent memory, re-enter a later decision, and survive naive deletion through derived state. https://5sigmas.com/en/series/seguridad-ia/03-envenenamiento.mp4 48 2026-08-06T00:00:00+00:00 https://5sigmas.com/en/videos/series/seguridad-ia/02-jailbreaks/ https://5sigmas.com/en/series/seguridad-ia/02-jailbreaks.jpg Jailbreaks How repeated and adaptive prompt attempts change the attack surface, and why authorization and attempt limits still matter after a refusal. https://5sigmas.com/en/series/seguridad-ia/02-jailbreaks.mp4 48 2026-08-06T00:00:00+00:00 https://5sigmas.com/en/videos/series/seguridad-ia/00_presentacion_serie/ https://5sigmas.com/en/series/seguridad-ia/00_presentacion_serie.jpg AI Security How untrusted input can influence an AI system and which authorization boundaries limit whether that influence becomes an action. https://5sigmas.com/en/series/seguridad-ia/00_presentacion_serie.mp4 48 2026-08-06T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-ia/05-de-la-demo-a-produccion/ https://5sigmas.com/en/series/agentes-ia/05-de-la-demo-a-produccion.jpg From demo to production Shows what makes an agent operable under failure: budgets, retries, idempotency, traces, explicit closure, and when a deterministic workflow is the safer design. https://5sigmas.com/en/series/agentes-ia/05-de-la-demo-a-produccion.mp4 32 2026-07-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-ia/04-seguridad-agentes/ https://5sigmas.com/en/series/agentes-ia/04-seguridad-agentes.jpg Agent security Shows how untrusted content can influence actions and why provenance, capability, authorization, and policy need independent enforcement outside the prompt. https://5sigmas.com/en/series/agentes-ia/04-seguridad-agentes.mp4 32 2026-07-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-ia/03-como-evaluar-un-agente/ https://5sigmas.com/en/series/agentes-ia/03-como-evaluar-un-agente.jpg Tutorial: how to evaluate an AI agent A practical tutorial for evaluating the agent's whole task: outcome, trajectory, permissions, cost, recovery, and abstention. A correct final answer is not enough. https://5sigmas.com/en/series/agentes-ia/03-como-evaluar-un-agente.mp4 32 2026-07-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-ia/02-anatomia-de-un-agente/ https://5sigmas.com/en/series/agentes-ia/02-anatomia-de-un-agente.jpg The anatomy of an agent Separates what the model proposes from what the runtime decides, then maps tool contracts, memory, state, context, and termination. https://5sigmas.com/en/series/agentes-ia/02-anatomia-de-un-agente.mp4 32 2026-07-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-ia/01-que-es-un-agente/ https://5sigmas.com/en/series/agentes-ia/01-que-es-un-agente.jpg What an AI agent is—and is not Contrasts an answer, a workflow, and a bounded agent, showing how state, tool limits, supervision, and stopping conditions define useful autonomy. https://5sigmas.com/en/series/agentes-ia/01-que-es-un-agente.mp4 32 2026-07-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/agentes-ia/00_presentacion_serie/ https://5sigmas.com/en/series/agentes-ia/00_presentacion_serie.jpg AI Agents A visual map of what turns a model that answers into a system that chooses actions: loops, evaluation, security, and bounded operation. https://5sigmas.com/en/series/agentes-ia/00_presentacion_serie.mp4 32 2026-07-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/datacenters-espacio/04-huella-real-datacenter/ https://5sigmas.com/en/series/datacenters-espacio/04-huella-real-datacenter.jpg 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. https://5sigmas.com/en/series/datacenters-espacio/04-huella-real-datacenter.mp4 53 2026-06-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/datacenters-espacio/03-que-es-datacenter-espacio/ https://5sigmas.com/en/series/datacenters-espacio/03-que-es-datacenter-espacio.jpg 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. https://5sigmas.com/en/series/datacenters-espacio/03-que-es-datacenter-espacio.mp4 53 2026-06-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/datacenters-espacio/02-energia-calor-conectividad/ https://5sigmas.com/en/series/datacenters-espacio/02-energia-calor-conectividad.jpg 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. https://5sigmas.com/en/series/datacenters-espacio/02-energia-calor-conectividad.mp4 53 2026-06-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/datacenters-espacio/01-por-que-ahora/ https://5sigmas.com/en/series/datacenters-espacio/01-por-que-ahora.jpg 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. https://5sigmas.com/en/series/datacenters-espacio/01-por-que-ahora.mp4 53 2026-06-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/datacenters-espacio/00_presentacion_serie/ https://5sigmas.com/en/series/datacenters-espacio/00_presentacion_serie.jpg Data Centers in Space Technical analysis of orbital data centers: heat dissipation, latency, launch cost and real viability compared with terrestrial infrastructure. https://5sigmas.com/en/series/datacenters-espacio/00_presentacion_serie.mp4 53 2026-06-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/seguridad-ia/01-prompt-injection/ https://5sigmas.com/en/series/seguridad-ia/01-prompt-injection.jpg Prompt Injection How an instruction hidden in a document can enter an AI system and which controls separate reading from action. https://5sigmas.com/en/series/seguridad-ia/01-prompt-injection.mp4 48 2026-05-26T00:00:00+00:00 https://5sigmas.com/en/videos/articulos-tecnicos/reactive-proactive-agent-header-demo/ https://5sigmas.com/en/articulos-tecnicos/reactive-proactive-agent-header-demo.jpg 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. https://5sigmas.com/en/articulos-tecnicos/reactive-proactive-agent-header-demo.mp4 20 2026-04-23T00:00:00+00:00 https://5sigmas.com/en/videos/series/modelos-razonadores/05-riesgos/ https://5sigmas.com/en/series/modelos-razonadores/05-riesgos.jpg Reasoning-model risks: overthinking, attacks and control Overthinking, indirect prompt injection, availability attacks and budget control: risks that require limits, verification and explicit fallbacks. https://5sigmas.com/en/series/modelos-razonadores/05-riesgos.mp4 121 2026-04-14T00:00:00+00:00 https://5sigmas.com/en/videos/series/modelos-razonadores/04-latencia-streaming/ https://5sigmas.com/en/series/modelos-razonadores/04-latencia-streaming.jpg Latency and streaming: design waiting as a policy TTFT, streaming, routing and fallbacks: how to treat latency as a product policy rather than an accident of inference. https://5sigmas.com/en/series/modelos-razonadores/04-latencia-streaming.mp4 115 2026-04-11T00:00:00+00:00 https://5sigmas.com/en/videos/series/modelos-razonadores/03-test-time-compute/ https://5sigmas.com/en/series/modelos-razonadores/03-test-time-compute.jpg Test-Time Compute: more steps, more candidates and more structure How inference compute can extend reasoning, compare candidates, search branches and change cost and latency without guaranteeing better answers by itself. https://5sigmas.com/en/series/modelos-razonadores/03-test-time-compute.mp4 129 2026-04-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/04-ia-pib-hoy/ https://5sigmas.com/en/series/ia-pib-bienestar-energia/04-ia-pib-hoy.jpg AI and GDP Today Why AI's macroeconomic impact takes time to appear in GDP, where it appears earlier, and which signals best indicate what is already changing. https://5sigmas.com/en/series/ia-pib-bienestar-energia/04-ia-pib-hoy.mp4 62 2026-04-10T00:00:00+00:00 https://5sigmas.com/en/videos/series/modelos-razonadores/02-fallos/ https://5sigmas.com/en/series/modelos-razonadores/02-fallos.jpg How reasoning models fail Sycophancy, shortcut learning, specification gaming and cascading failures: how to detect recurring failure patterns and bound their consequences. https://5sigmas.com/en/series/modelos-razonadores/02-fallos.mp4 108 2026-04-09T00:00:00+00:00 https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/03-pib-vs-bienestar/ https://5sigmas.com/en/series/ia-pib-bienestar-energia/03-pib-vs-bienestar.jpg 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. https://5sigmas.com/en/series/ia-pib-bienestar-energia/03-pib-vs-bienestar.mp4 62 2026-04-09T00:00:00+00:00 https://5sigmas.com/en/videos/series/modelos-razonadores/01-que-es-razonar/ https://5sigmas.com/en/series/modelos-razonadores/01-que-es-razonar.jpg 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. https://5sigmas.com/en/series/modelos-razonadores/01-que-es-razonar.mp4 82 2026-04-08T00:00:00+00:00 https://5sigmas.com/en/videos/series/modelos-razonadores/00_presentacion_serie/ https://5sigmas.com/en/series/modelos-razonadores/00_presentacion_serie.jpg Reasoning Models Five chapters on how LLMs reason: test-time compute, failure modes, latency, cost and risks when models use tools. https://5sigmas.com/en/series/modelos-razonadores/00_presentacion_serie.mp4 75 2026-04-08T00:00:00+00:00 https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/02-ia-tecnologia-electrica/ https://5sigmas.com/en/series/ia-pib-bienestar-energia/02-ia-tecnologia-electrica.jpg 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. https://5sigmas.com/en/series/ia-pib-bienestar-energia/02-ia-tecnologia-electrica.mp4 62 2026-04-08T00:00:00+00:00 https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/01-electricidad-bienestar/ https://5sigmas.com/en/series/ia-pib-bienestar-energia/01-electricidad-bienestar.jpg Electricity and Well-being Why reliable, affordable electricity enables real gains in health, logistics and industry, and why supply quality matters as much as kilowatt-hours. https://5sigmas.com/en/series/ia-pib-bienestar-energia/01-electricidad-bienestar.mp4 62 2026-04-07T00:00:00+00:00 https://5sigmas.com/en/videos/series/ia-pib-bienestar-energia/00_presentacion_serie/ https://5sigmas.com/en/series/ia-pib-bienestar-energia/00_presentacion_serie.jpg AI, GDP, Well-being and Energy Quantitative analysis of AI's impact on energy, productivity and well-being using real World Bank, IEA and Penn World Table data rather than speculative projections. https://5sigmas.com/en/series/ia-pib-bienestar-energia/00_presentacion_serie.mp4 62 2026-04-07T00:00:00+00:00 https://5sigmas.com/en/videos/series/multimodalidad-iag/05-riesgos/ https://5sigmas.com/en/series/multimodalidad-iag/05-riesgos.jpg Risks of Multimodal AI Systems Visual prompt injection, privacy, context leakage and tool manipulation when a multimodal system can observe and act. https://5sigmas.com/en/series/multimodalidad-iag/05-riesgos.mp4 62 2026-04-03T00:00:00+00:00 https://5sigmas.com/en/videos/series/multimodalidad-iag/04-evaluacion/ https://5sigmas.com/en/series/multimodalidad-iag/04-evaluacion.jpg Evaluating Multimodal Systems How to evaluate multimodal systems without confusing benchmarks with real capability: OCR, audio, grounding, reasoning and metric failures. https://5sigmas.com/en/series/multimodalidad-iag/04-evaluacion.mp4 62 2026-04-03T00:00:00+00:00 https://5sigmas.com/en/videos/series/multimodalidad-iag/03-arquitecturas/ https://5sigmas.com/en/series/multimodalidad-iag/03-arquitecturas.jpg Multimodal System Architectures Four multimodal architecture families, their differences in quality, cost and latency, and when each way of combining modalities makes sense. https://5sigmas.com/en/series/multimodalidad-iag/03-arquitecturas.mp4 62 2026-04-02T00:00:00+00:00 https://5sigmas.com/en/videos/series/multimodalidad-iag/02-alineamiento/ https://5sigmas.com/en/series/multimodalidad-iag/02-alineamiento.jpg 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. https://5sigmas.com/en/series/multimodalidad-iag/02-alineamiento.mp4 62 2026-04-02T00:00:00+00:00 https://5sigmas.com/en/videos/series/multimodalidad-iag/01-el-problema/ https://5sigmas.com/en/series/multimodalidad-iag/01-el-problema.jpg 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. https://5sigmas.com/en/series/multimodalidad-iag/01-el-problema.mp4 62 2026-04-01T00:00:00+00:00 https://5sigmas.com/en/videos/series/multimodalidad-iag/00_presentacion_serie/ https://5sigmas.com/en/series/multimodalidad-iag/00_presentacion_serie.jpg 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. https://5sigmas.com/en/series/multimodalidad-iag/00_presentacion_serie.mp4 62 2026-04-01T00:00:00+00:00 https://5sigmas.com/en/videos/series/from-cave-to-agi/05-mas-alla/ https://5sigmas.com/en/series/from-cave-to-agi/05-mas-alla.jpg 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. https://5sigmas.com/en/series/from-cave-to-agi/05-mas-alla.mp4 63 2026-03-31T00:00:00+00:00 https://5sigmas.com/en/videos/series/from-cave-to-agi/04-escalar/ https://5sigmas.com/en/series/from-cave-to-agi/04-escalar.jpg Scale — deep learning to foundation models A short video explanation of Scale — deep learning to foundation models. https://5sigmas.com/en/series/from-cave-to-agi/04-escalar.mp4 63 2026-03-30T00:00:00+00:00 https://5sigmas.com/en/videos/series/from-cave-to-agi/03-aprender/ https://5sigmas.com/en/series/from-cave-to-agi/03-aprender.jpg 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. https://5sigmas.com/en/series/from-cave-to-agi/03-aprender.mp4 63 2026-03-27T00:00:00+00:00 https://5sigmas.com/en/videos/series/from-cave-to-agi/02-mecanizar/ https://5sigmas.com/en/series/from-cave-to-agi/02-mecanizar.jpg 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. https://5sigmas.com/en/series/from-cave-to-agi/02-mecanizar.mp4 63 2026-03-26T00:00:00+00:00 https://5sigmas.com/en/videos/series/from-cave-to-agi/01-representar/ https://5sigmas.com/en/series/from-cave-to-agi/01-representar.jpg 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. https://5sigmas.com/en/series/from-cave-to-agi/01-representar.mp4 63 2026-03-26T00:00:00+00:00 https://5sigmas.com/en/videos/series/from-cave-to-agi/00_presentacion_serie/ https://5sigmas.com/en/series/from-cave-to-agi/00_presentacion_serie.jpg From the Caves to AGI An intellectual history of AI: from the first mathematical abstractions to foundation models. Mathematics, philosophy and computing in context. https://5sigmas.com/en/series/from-cave-to-agi/00_presentacion_serie.mp4 63 2026-03-26T00:00:00+00:00 https://5sigmas.com/en/videos/series/fundamentos-ia-iag/04-agi/ https://5sigmas.com/en/series/fundamentos-ia-iag/04-agi.jpg AGI: Artificial General Intelligence AGI means artificial general intelligence. This chapter explains its definitions, DeepMind's and OpenAI's levels, and what would still be required to reach it. https://5sigmas.com/en/series/fundamentos-ia-iag/04-agi.mp4 63 2026-03-20T00:00:00+00:00 https://5sigmas.com/en/videos/series/fundamentos-ia-iag/03-ia-vs-ia-generativa/ https://5sigmas.com/en/series/fundamentos-ia-iag/03-ia-vs-ia-generativa.jpg 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. https://5sigmas.com/en/series/fundamentos-ia-iag/03-ia-vs-ia-generativa.mp4 63 2026-03-20T00:00:00+00:00 https://5sigmas.com/en/videos/series/fundamentos-ia-iag/02-que-es-ia-generativa/ https://5sigmas.com/en/series/fundamentos-ia-iag/02-que-es-ia-generativa.jpg 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. https://5sigmas.com/en/series/fundamentos-ia-iag/02-que-es-ia-generativa.mp4 63 2026-03-17T00:00:00+00:00 https://5sigmas.com/en/videos/series/fundamentos-ia-iag/01-que-es-ia/ https://5sigmas.com/en/series/fundamentos-ia-iag/01-que-es-ia.jpg 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. https://5sigmas.com/en/series/fundamentos-ia-iag/01-que-es-ia.mp4 63 2026-03-15T00:00:00+00:00 https://5sigmas.com/en/videos/series/fundamentos-ia-iag/00_presentacion_serie/ https://5sigmas.com/en/series/fundamentos-ia-iag/00_presentacion_serie.jpg 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. https://5sigmas.com/en/series/fundamentos-ia-iag/00_presentacion_serie.mp4 53 2026-03-15T00:00:00+00:00