Karpathy is confusing LLM limitations with AI system limitations, and that makes all the difference. 

Karpathy is confusing LLM limitations with AI system limitations, and that makes all the difference.&nbsp;

Unsupervised Learning is about ideas and trends in Cybersecurity, National Security, AI, Technology, and Culture—and how best to upgrade ourselves to be ready for what's coming.

Unsupervised Learning is about ideas and trends in Cybersecurity, National Security, AI, Technology, and Culture&mdash;and how best to upgrade ourselves to be ready for what's coming.

Exploring the fascinating intersection of security, technology, and humans.

Unsupervised Learning

Unsupervised Learning (Member Edition)

Why I Think Karpathy is Wrong on the AGI Timeline

https://api.omny.fm/orgs/070af456-729b-4a0f-9c09-a6c100397b59/clips/cbeb8cdf-250f-4e40-82f6-b37c00ecd96b/transcript

How going from exploration to exploitation can help you as both a consumer and creator of everything.

How going from exploration to exploitation can help you as both a consumer and creator of everything.

Novelty Exploration vs. Pattern Exploitation

https://api.omny.fm/orgs/070af456-729b-4a0f-9c09-a6c100397b59/clips/46f8fcea-610d-4df4-b466-b37701604919/transcript

Some thoughts on how novelty and attention magnify the time that we have. 

Some thoughts on how novelty and attention magnify the time that we have.&nbsp;

Magnifying Time

https://api.omny.fm/orgs/070af456-729b-4a0f-9c09-a6c100397b59/clips/5eef7dd7-5879-4e21-a9af-b3760154054a/transcript

➡ Stay Ahead of Cyber Threats with AI-Driven Vulnerability Management with Maze:
https://mazehq.com/

In this conversation, I speak with Harry about how AI is transforming vulnerability management and application security. We explore how modern approaches can move beyond endless reports and generic fixes, toward real context-aware workflows that actually empower developers and security teams.

We talk about:

The Real Problem in Vulnerability Management
Why remediation—not just prioritization—remains the toughest challenge, and how AI can help bridge the gap between vulnerabilities and the developers who need to fix them.

Context, Ownership, and Velocity
How linking vulnerabilities to the right applications and teams inside their daily tools (like GitHub) reduces friction, speeds up patching, and improves security without slowing developers down.

AI Agents and the Future of Security
Why we should think of AI agents as “extra eyes and hands,” and how they’re reshaping everything from threat detection to system design, phishing campaigns, and organizational defense models.

Attackers Move First
How attackers are already building unified world models of their targets using AI, and why defenders need to match (or exceed) this intelligence to stay ahead.

From Days to Minutes
Why the tolerance for vulnerability windows is shrinking fast, and how automation and AI are pushing us toward a future where hours—or even minutes—make the difference.

Subscribe to the newsletter at:
https://danielmiessler.com/subscribe

Join the UL community at:
https://danielmiessler.com/upgrade

Follow on X:
https://x.com/danielmiessler

Follow on LinkedIn:
https://www.linkedin.com/in/danielmiessler

Chapters:

00:00 – Welcome and Harry’s Background
01:07 – The Real Problem: Remediation vs. Prioritization
04:31 – Breaking Down Vulnerability Context and Threat Intel
05:46 – Connecting Vulnerabilities to Developers and Workflows
08:01 – Why Traditional Vulnerability Management Fails
10:29 – Startup Lessons and The State of AI Agents
13:26 – DARPA’s AI Cybersecurity Competition
14:29 – System Design: Deterministic Code vs. AI
16:05 – How the Product Works and Data Sources
18:01 – AI as “Extra Eyes and Hands” in Security
20:20 – Breaking Barriers: Rethinking Scale with AI
23:22 – Building World Models for Defense (and Attack)
25:22 – Attackers Move Faster: Why Context Matters
27:04 – Phishing at Scale with AI Agents
31:24 – Shrinking Windows of Vulnerability: From Days to Minutes
32:47 – What’s Next for Harry’s Work
34:13 – Closing Thoughts

➡ Stay Ahead of Cyber Threats with AI-Driven Vulnerability Management with Maze: <a href="https://mazehq.com/?utm_source=unsupervised-learning&amp;utm_medium=interview-daniel-messer&amp;utm_campaign=influencer">https://mazehq.com/</a>
In this conversation, I speak with Harry about how AI is transforming vulnerability management and application security. We explore how modern approaches can move beyond endless reports and generic fixes, toward real context-aware workflows that actually empower developers and security teams.
We talk about:
The Real Problem in Vulnerability Management Why remediation&mdash;not just prioritization&mdash;remains the toughest challenge, and how AI can help bridge the gap between vulnerabilities and the developers who need to fix them.
Context, Ownership, and Velocity How linking vulnerabilities to the right applications and teams inside their daily tools (like GitHub) reduces friction, speeds up patching, and improves security without slowing developers down.
AI Agents and the Future of Security Why we should think of AI agents as &ldquo;extra eyes and hands,&rdquo; and how they&rsquo;re reshaping everything from threat detection to system design, phishing campaigns, and organizational defense models.
Attackers Move First How attackers are already building unified world models of their targets using AI, and why defenders need to match (or exceed) this intelligence to stay ahead.
From Days to Minutes Why the tolerance for vulnerability windows is shrinking fast, and how automation and AI are pushing us toward a future where hours&mdash;or even minutes&mdash;make the difference.
Subscribe to the newsletter at: <a href="https://danielmiessler.com/subscribe">https://danielmiessler.com/subscribe</a>
Join the UL community at: <a href="https://danielmiessler.com/upgrade">https://danielmiessler.com/upgrade</a>
Follow on X: <a href="https://x.com/danielmiessler">https://x.com/danielmiessler</a>
Follow on LinkedIn: <a href="https://www.linkedin.com/in/danielmiessler">https://www.linkedin.com/in/danielmiessler</a>
Chapters:
00:00 &ndash; Welcome and Harry&rsquo;s Background 01:07 &ndash; The Real Problem: Remediation vs. Prioritization 04:31 &ndash; Breaking Down Vulnerability Context and Threat Intel 05:46 &ndash; Connecting Vulnerabilities to Developers and Workflows 08:01 &ndash; Why Traditional Vulnerability Management Fails 10:29 &ndash; Startup Lessons and The State of AI Agents 13:26 &ndash; DARPA&rsquo;s AI Cybersecurity Competition 14:29 &ndash; System Design: Deterministic Code vs. AI 16:05 &ndash; How the Product Works and Data Sources 18:01 &ndash; AI as &ldquo;Extra Eyes and Hands&rdquo; in Security 20:20 &ndash; Breaking Barriers: Rethinking Scale with AI 23:22 &ndash; Building World Models for Defense (and Attack) 25:22 &ndash; Attackers Move Faster: Why Context Matters 27:04 &ndash; Phishing at Scale with AI Agents 31:24 &ndash; Shrinking Windows of Vulnerability: From Days to Minutes 32:47 &ndash; What&rsquo;s Next for Harry&rsquo;s Work 34:13 &ndash; Closing Thoughts

A Conversation With Harry Wetherald CO-Founder & CEO At Maze

https://api.omny.fm/orgs/070af456-729b-4a0f-9c09-a6c100397b59/clips/e0352a6a-43e6-4b60-b359-b3560146f807/transcript

Why I Think Karpathy is Wrong on the AGI Timeline

In 1 playlist(s)

Unsupervised Learning

Social links

Follow podcast

Recent clips