Video summary

Día 1 | Curso gratuito de inteligencia artificial con Certificado

Main summary

Key takeaways

Educational

Main ideas, concepts, and lessons

  • Purpose of the training (4-day free live course)

    • Days 1–4 are designed to help participants become AI professionals through live, practical exercises.
    • The course focuses on building virtual employees / AI agents that can delegate day-to-day work and increase productivity.

  • Reframing AI for non-technical people

    • AI is presented as not only for technical/digital-native people—anyone can master it professionally.
    • The main advantage is combining AI capability with the user’s experience and judgment in their domain.

  • Urgency and “exponential” change

    • AI adoption is described as exponential, not gradual (e.g., “multiplying by 10 every year”).
    • Not using AI is framed as a disadvantage: jobs and skills will evolve, and people who don’t adapt may be left behind.

  • Dual reality: opportunity + risk

    • Opportunity
      • AI can multiply productivity and potentially help people earn more and improve professional standing.
    • Risk
      • AI-driven layoffs and workplace transformation are described as already happening in major companies.
      • AI skills are described as non-optional for employability and market value.

  • What AI models can do (capabilities)

    • AI models are framed as “digital brains” with abilities such as:
      • Programming
      • Data analysis
      • Writing on any topic
      • Conversation (speaking)
      • Memory/recall (context about the user)
      • Creating images/videos
      • Perceiving/understanding images (image input)

  • How to get better results: prompt engineering (“priming”)

    • AI outputs depend heavily on how users write instructions.
    • Prompt Engineering is introduced via a method called priming, structured as:
      1. Role
      2. Context
      3. Instruction
      4. Format (e.g., PDF/Excel/website)

  • Practical organization inside ChatGPT: Projects + Scheduled Tasks

    • Projects
      • A structured space for shared sources/documents to keep conversations consistent within a theme.
      • Supports project-level instructions, reducing the need to repeat context.
    • Scheduled tasks
      • Automates recurring actions (e.g., daily research at a set time).

  • Day 1 focus

    • Learn core model usage and basic workflows:
      • In-depth research
      • Image analysis
      • Document analysis and transformation
      • Voice/talk-to-the-model role simulation
    • Prepare for agents on Day 2.

  • Preview of “agents”

    • Agents are described as the next step beyond chat:
      • Like moving from Alexa (talking) to a humanoid robot (doing tasks end-to-end).
    • Agents are framed as virtual employees that can use tools (email, calendar, etc.) and perform tasks autonomously.

Methodologies / instructions presented (detailed)

A) “Priming” methodology for better prompts (Prompt Engineering)

Use this structure to get higher-quality outputs from AI models:

  • 1) Role

    • Specify what kind of expert the AI should be.
    • Example: “Act as a business consulting and growth strategy expert.”

  • 2) Context

    • Define where the expert “works,” who the audience is, and domain details.
    • Example: “You work in the consulting division of [program/university/company] supporting students.”

  • 3) Instruction

    • Describe the exact task(s), goals, constraints, and what the AI must produce.
    • Example: “Create a growth plan for Volkswagen Mexico City to increase sales by 15% by end of year.”

  • 4) Format

    • Request the output structure/type explicitly (e.g., “return as PDF”).
    • Example: “Provide results in a PDF” (optionally “use thinking mode”).

Core lesson: Without priming, the AI may generate a generic plan and a “dump” of information. With priming, it produces clearer diagnoses, structured action plans, and goal-aligned results.

B) ChatGPT hands-on workflows taught on Day 1 (4 “big things”)

Participants practice these main functions:

  • 1) In-depth information search

    • Use Advanced Research (described as available in paid accounts).
    • Process shown:
      • Enable advanced research
      • Ask using dictation naturally
      • Generate a plan, then run the investigation for many minutes
      • Receive a structured report with diagrams/flowcharts
      • Download/export (e.g., Word/PDF)

  • 2) Image analysis

    • Upload photos and ask the model to interpret them.
    • Examples:
      • Domestic: analyzing washing machine programs and selecting safe detergents
      • Work: translating menus, interpreting invoices, understanding photographed documents

  • 3) Talking/speaking simulation (voice role-play)

    • Use voice interaction rather than typing to simulate real-world performance pressure.
    • Examples:
      • Training for sales conversations
      • Practicing interviews
      • Teaching kids complex concepts (e.g., bedtime-story explanations)

  • 4) Document analysis and transformation

    • Upload spreadsheets/statements and ask for:
      • Categorization of transactions
      • Summaries (income/expenses/savings)
      • Recommendations
      • Output as transformed spreadsheets (e.g., multiple tabs/sheets)
    • Follow-up iteration:
      • Ask for projections (e.g., 10-year investment scenarios)
      • Request a more visual output format (e.g., a “website”/HTML-style explanation)

C) Using ChatGPT interface effectively (as taught)

Key steps and concepts mentioned:

  • Create a free account

    • Access via chat.com
    • Use “sign up for free” and sign in with Google

  • Understand free vs paid tiers

    • Paid accounts are described as offering advanced features and modes.
    • Mentioned options:
      • “Go” / “Pro” style paid levels (pricing varies by country)
      • More advanced research and “thinking” behavior

  • Switch between model modes

    • Instant mode: faster, less “thinking”
    • Thinking/professional modes: better for building and planning
    • Dictation/voice is highlighted as a productivity tool.

  • Use the left sidebar features

    • Projects
    • Scheduled tasks
    • Library/storage for generated assets (e.g., images)

D) Cloud/other models and “connectors” (conceptual workflow)

  • Introduce Cloude/Cloud as a model with:

    • Multiple model tiers (e.g., Opus/Sonet/Haiku)
    • Connectors to integrate tools such as:
      • Gmail/email
      • WhatsApp
      • Google Calendar and other work applications (conceptually)

  • Example workflow shown:

    • An AI agent:
      • Understands a company from its website
      • Finds potential clients in specific cities
      • Builds a CRM in Excel format
      • Drafts sales emails and places them into email drafts (or schedules them)

Overall “Day 1” progression (what participants are meant to do)

  • Learn the capabilities of AI models (what they can do).
  • Practice the four main tasks:
    • In-depth research
    • Image analysis
    • Voice role-play
    • Document analysis/transformation
  • Learn priming to improve prompt quality.
  • Learn Projects and Scheduled Tasks to organize work for recurring themes and automation.
  • Complete today’s assignment:
    • Create a professional prompt for a real challenge.

Assignment / required practice (explicit)

  • Task 1 (by tomorrow)

    • Create your first professional prompt:
      • Use Role + Context + Instructions (+ desired Format).
    • The prompt should address either:
      • A work challenge (strategy, competitor analysis, research, etc.), or
      • A personal challenge (e.g., difficult family conversation, travel planning).
    • Submit via the provided support contact/email.

  • Optional extension

    • If you already have a project idea and time:
      • Create a Project and upload relevant documents as context, such as:
        • “Best practices” PDFs/guides
        • Company information
        • Example materials/templates already available
      • The goal is to prevent the AI from “making up” facts by grounding it in your uploaded context.

Speakers / sources mentioned (and who they are in the video)

Speakers (featured people)

  • Arnau Ramoso — co-founder of Learning Heroes (host/introducer)
  • Javier Sanz — Director of Artificial Intelligence at Learning Heroes; consulting firm; works with clients and OpenAI/ChatGPT
  • Carolina — appears at the end during the “tomorrow” agent teaser and schedule/cartoon-like phone agent example (voice/host segment)
  • Eduardo — described as a student participant whose story is narrated (podcast testimony format)

Sources / referenced organizations / public figures

  • OpenAI (ChatGPT creator)
  • Learning Heroes (training provider; presented as a “university center for disruptive technologies”)
  • OpenChatGPT (mentioned as launching to the world)
  • Google (for access via chat.com and model comparison)
  • Gemini (Google’s model)
  • Grok (Elon Musk’s model; connected to Twitter/real-time data in the narration)
  • Cloud/Clot (model referenced as “Cloud/Cloude/Claudio”; includes “connectors”)
  • Layoffs/layoff tracking site: Layoffs.fyi (spelled in subtitles as “Lay of Hedge”)
  • Harvard simulation study (quality improvement with AI assistance)
  • PricewaterhouseCoopers (PwC) (study on wages/value increases with AI skills)
  • Oxford University (study referenced similarly)
  • Satya Nadella — CEO of Microsoft (quoted/paraphrased about AI as a must-have skill)
  • Jensen Huang (NVIDIA leader) — referenced in the context of AI skill importance
  • Dario Amodei (Anthropic leader) — referenced in the context of AI skill importance
  • Mustafa Suleyman (AI leader) — referenced similarly
  • Microsoft Copilot (mentioned as a related tool; integration with connectors discussed)
  • Excel / Word / PDF (document export formats referenced in practice)

  • Microsoft Teams (mentioned as a context for exporting/sending)

  • YouTube/Zoom chat participants

    • Multiple named attendees are briefly greeted (e.g., Cristina, Carlos, Francisco, Lalia, Marita, Ari(st)óes, Ariel), but they are not speakers in the instructional sense.

Original video