KEY TAKEAWAYS (an AI-Output with Hallucinations)

MY #1 USED AI TOOL

• Wispr Flow - I don’t type anymore, I just voice everything

AI AGENTS & AGentic workflows

• Atlas - AI Browser by OpenAI

• n8n - Workflow automation

• Cursor - Coding agent

• Comet by Perplexity - Best Agentic Browser - browse the internet with AI and get stuff done

• Lindy - Build an AI agent - great for AI automation

LLM

• Gemini 3 - The Best LLM on the planet right now

• OpenAI’s Deep Research - Absolutely the best research tool out there right now

• Manus - Most comprehensive AI agent in my opinion

• Claude Skills - Claude Skills is the best skills agent - this allows to be more deterministic with your agents.

• Grok 4 - Equivalent to o3 from ChatGPT - awesome for X data

Text-to-App

• Lovable - Build any app with agentic reasoning (best at front-end)

• v0 by Vercel - Text-to-Webpage (really good front-end)

• Replit - Idea-to-App (best for full app)

Text-to-IMAGE/VIDEO

• Midjourney - Best for pictures and video

• HeyGen - AI Video Avatar

• Runway - Text to Video

• Google’s Nano Banana Pro - Best Image Generator

Research & other

• NotebookLM - Research and the two AI podcast hosts that you see above :)

• Google AI Studio - Real-time AI co-presence

• Clay - AI Outbound Sales | Generating lists of emails and sending cold outreach

Briefing: Innovation, AI, and the Agentic Era

Executive Summary

ESummary: Shawn Kanungo's Agentic AI Presentation

This was your keynote at McAfee, focused on agentic AI as the most disruptive technology of the century. Here are the detailed takeaways:

Core Thesis

Agentic AI represents a fundamental shift from content creation to action. Since ChatGPT launched in late 2022, AI has mainly helped with summarizing, refining, and creating content. In 2026 and beyond, the conversation moves to AI that actually gets stuff done—software that can think like us and execute autonomously.

Key Concepts

The Spectrum of Agency

You distinguished between two ends:

• LLM Workflows: Predefined recipes (AI summarizes a conversation, sends an email)

• True Autonomous Agents: Always-on, 24/7, making decisions independently

The reality is that platforms are becoming more agentic over time—gaining reasoning capabilities, tool access, and decision-making autonomy incrementally.

The Perplexity/Harvard Study Finding

People aren't yet using agents to execute tasks. They're primarily using them to outsource research and thinking. This is the current adoption phase before full action delegation.

The New Internet Paradigm

• Old model: Search → Decide → Act (you do everything)

• New model: Delegate to agent → Agent researches and purchases → You adjust

Memorable Frameworks

"The Kelly Rowland Future"

The future of technology is that it just works without users understanding how. Like Mariah Carey apparently never hearing of Microsoft Excel—technology should handle everything invisibly.

The Restaurant Analogy for AI Comfort Levels

1. Collaborator: AI helps you create recipes

2. Sous Chef: AI assists with execution

3. Main Chef: AI handles core operations

4. Co-owner: AI makes autonomous decisions

System of Record → System of Intelligence

For decades, software was just an elegant way to retrieve, collect, and maintain data. Agentic AI creates a system of intelligence—capable of 24/7 reconciliation, auditing, and tasks humans could never sustain.

The Security Opportunity (McAfee-Specific)

New Threat Categories for Agents

• Prompt injection: Manipulating agents to reveal confidential information

• Tool poisoning: Agents leveraging nefarious or fake tools

• Delegation overreach: Agents exceeding their intended scope

• Context poisoning: Corrupted memory/context that agents pull from

"AI Marks the End of Trust"

When you can't distinguish real from AI-generated, the fundamental question shifts from "Is this real?" to "Who made it?" Identity becomes the firewall—everything created online should have identity attached.

The McAfee Positioning

Consumers don't care about background technology. They care about: Is my stuff secure? Is my identity protected? The future of security is frictionless, agentic, and omnipresent—a safety companion guiding users through an increasingly synthetic internet.

The Human Constants

Despite radical technological change, these have never changed in 300,000 years of human existence:

• Love for stories

• Desire for status

• Need for safety and security

• Craving for intimacy and relationships

• Delight and surprise

Technology hasn't changed these needs—it's raised expectations around them.

The Friction Paradox

You argued for bringing back "the F word"—friction, meaning humanity.

Two Winning Categories

1. Fast and frictionless: Amazon, Temu, McDonald's—push button, get result

2. Slow and experiential: Four Seasons, Hermès, Apple Genius Bar—craftsmanship, one-on-one interaction

The danger zone is the middle—the "black hole of mediocrity." Organizations need to choose which extreme they're optimizing for.

The Disruption Imperative

"100 to Zero" vs "Zero to 100"

Going from zero to growth is sexy. The harder, more rewarding work is disrupting yourself—reimagining, reskilling, becoming a rookie again.

The Carl Jung Quote

"The first half of your life is developing a healthy ego; the last half is letting go."

The Coca-Cola Executive's Insight

Their biggest disruptor isn't Prime or any competitor—it's good leadership. In the pursuit of optimization, they forgot how to reimagine. Good leaders optimize the past; remarkable leaders focus on innovation and ask: "Am I willing to disrupt myself?"

The Fear Antidote

When someone told you the AI presentation scared them, your response: "The reason you're scared is because you haven't put in the work yet."

Everyone who immerses themselves in the technology, builds proficiency, develops the muscle—none of them say AI will take their job. They all say: "What can we create? How can we help people?"

Your advice: Work scared until you become scary. Work scared until you become dangerous.

Closing Line

"The most dangerous person in the room is the person who is most afraid, yet bold enough to move forward."

Embracing Innovation's Challenges

The Willingness to Look Foolish

• True innovation requires embracing uncertainty and discomfort ("the darkness")

• Focuses on identity rather than just outcomes

• Requires qualities like curiosity, resilience, and risk-taking

Innovation as Universal Responsibility

• Not a separate department but a mindset for everyone

• Continuously asking how to improve

The Most Valuable Question

• "What will you start today that scares you?"

• Begins changing your story and challenging limits

Self-Disruption

• The greatest challenge isn't scaling up (0 to 100)

• It's having the courage to disrupt yourself and what's working (100 to 0)

Beyond Mountains

• Success shouldn't breed complacency

• There are always new heights after reaching what seemed like the summit

To Reach 5×, You Can’t Think 5 % Bigger—You Have to Go to Zero

Take any core process—finance approvals, R&D cycles, vessel maintenance—and ask: If we rebuilt this from scratch with today’s tech, what vanishes? What automates? What now demands more human warmth? Going “to zero” prevents incrementalism from smothering breakthroughs.

The Most Dangerous Person in the Room

It’s not the algorithm or the credentialed expert—it’s the individual who feels the fear, yet moves anyway. Bold + scared beats brilliant + stuck. Be that person, hire that person, partner with that person, and the future tips in your favor.