Resilient Cyber

Itamar Apelblat was called back to reserve duty with Unit 8200 and spent hours waiting on access approvals. Then he remembered a service account his old team had set up seven years earlier. The password was a funny one, which is why he still knew it. He typed it in and became an admin on one of the most secure networks in the world.

Nobody had changed that password. They hadn't forgotten about it. They were scared to touch it, because no one knew what would break.

Now put that same fear in front of AI agents. When Token Security's research team checked its enterprise customers, more than 16 percent already had Muse running locally against corporate email and other resources. Muse launched three weeks before we recorded and is already the most popular productivity app in the App Store. Itamar expects agents to outnumber humans in the enterprise a hundred to one.

His argument is that you can't secure agents by predicting every way they'll go wrong. An agent chases its goal like it's possessed and takes any access path that gets it there. What you can control is how much autonomy it has and what it's allowed to touch. Both come down to identity.

In this episode I sit down with Itamar, co-founder and CEO of Token Security, and we get specific:

→ The three things companies find the first time they look: agents wired straight into production on raw API keys, abandoned experiment agents still holding privileges, and agents still running after their owner left the company
→ What intent-based access management actually means, and why an agent built to send a weekly report should never be able to delete from your database
→ Why IGA and PAM fall apart here: they treat Okta or Active Directory as the source of truth, while agents live on laptops, in SaaS, and in the cloud
→ What happens when a coding agent tries to connect to a production Snowflake environment, and how to decide where a human still has to approve
→ Who should own agent security, and why Itamar puts most of it with the IAM team

I also put the hard question to him. We've had twenty years to get least privilege right for humans and mostly failed, so why would agents go any better? Then, at the end, he gives his advice on where security teams should start. It isn't the usual "get visibility first," and it's the part I'd most want a security leader to hear.

⏱️ CHAPTERS

0:00 Introduction
1:49 Unit 8200 and "Which Product Do You Hate the Most?"
3:58 The Seven-Year-Old Password That Still Worked
6:32 Autonomy and Access: The Two Dimensions of Agent Risk
8:19 Why Agents Act Like They're Possessed
10:14 16% Already Run Muse: What Companies Find
12:46 Intent-Based Access Management, Explained
15:15 Twenty Years of Failed Least Privilege
17:15 Why Your IGA Won't Scale to Agents
19:04 Where the Human Stays in the Loop
21:12 Who Should Own Agent Security
23:33 The Security Teams That Get This Right
26:35 Closing Thoughts

SOURCES AND FURTHER READING

Token Security: https://www.token.security/
Itamar Apelblat on LinkedIn: https://www.linkedin.com/in/itamar-apelblat/

📬 SUBSCRIBE TO RESILIENT CYBER
https://www.resilientcyber.io

If this was useful, subscribe. I cover AI security, agentic AI, AppSec, software supply chain, vulnerability management, and cyber policy for practitioners every week.

#AISecurity #AgenticAI #IdentitySecurity #NonHumanIdentities #IAM #AIAgents #LeastPrivilege #Cybersecurity

Matt Hartley, co-founder and Chief Product Officer of BreachRx, on why the once-a-year tabletop no longer matches the incidents organizations will actually face, and how AI-accelerated micro tabletops let teams train like they fight.

In this episode I sit down with Matt Hartley, co-founder and Chief Product Officer of BreachRx. Matt has spent 28 years in cybersecurity. He started in government offensive security, then moved to the cyber threat intel startup iSight Partners, where he ran product through its acquisition by FireEye, and later ran Mandiant Engineering. He watched consulting teams parachute into companies that were compromised and meaningfully unprepared, and that convinced him the real gap was the response process, not detection or tooling.

We get into what a micro tabletop actually looks like and why exercises should run in the same system you will use during a real incident. We also cover how notification clocks as short as 30 minutes change what legal teams need to practice, and why the floor is shifting toward multiple concurrent incidents.

In this episode:
● Why the industry over-invests in technology and under-invests in the processes and procedures that carry a response
● What a micro tabletop looks like, from a one-hour cross-functional session to an exercise that runs asynchronously over a week with a 30-minute hot wash
● Train like you fight, and why rehearsing in your actual response platform lowers stress and folds lessons learned back into procedures
● More than 200 laws with incident notification requirements, including China's 30-minute window and India's six-hour rule covering even denial of service
● Starting a legal tabletop partway into a ransomware scenario to pressure test DORA, SEC, and other regulatory timelines
● Why privileging everything means nothing is privileged, and separating the factual record from privileged conversations
● The risk of 10 to 100 Slack or Teams channels spinning up around a single incident, and how one grouchy message can end up in evidence
● Keeping exercises from becoming check-the-box by matching frequency to team size and threat, pulling scenarios from the headlines, and running two incidents at once
● Why AI is raising the attack floor, from smarter background malicious traffic to lower-tier nation states making big capability jumps
● A workforce of AI agents that cuts incident reporting from 20 hours to 30 minutes and turns NIST, CISA, or Gartner guidance into tailored procedures
● A crawl, walk, run playbook for your first micro tabletop: three injects, one ornery stakeholder relationship, and a ten-minute debrief
● What regulators and plaintiffs will ask for, and why the SEC may look at five years of incidents, not just one

Chapters:
0:00 – Intro
1:42 – Matt's Path from Offensive Security to Mandiant
2:14 – Why Response, Not Detection, Was the Real Problem
3:40 – What a Micro Tabletop Actually Looks Like
7:37 – Train Like You Fight in the Real Platform
9:33 – Regulatory Clocks, DORA, SEC, and 30-Minute Notifications
11:56 – Privilege, Discoverability, and the 100 Slack Channel Problem
15:03 – Keeping Exercises from Becoming Check-the-Box
17:48 – Concurrent Incidents and a Rising Attack Floor
19:53 – Where AI Agents Fit in Incident Response
22:01 – Running Your First Micro Tabletop Next Month
24:12 – Proving Continuous Readiness to Regulators and Plaintiffs
26:55 – Closing

Connect with Matt Hartley:
LinkedIn: https://www.linkedin.com/in/wmhartl/

BreachRx: https://www.breachrx.com
BreachRx IR Exercises: https://www.breachrx.com/rex-platform/incident-response-exercises/

Read my piece on BreachRx and incident response, Chaos, Clocks, and 

Command Centers: https://www.resilientcyber.io/p/chaos-clocks-and-command-centers

Resilient Cyber: https://www.resilientcyber.io

Subscribe for more conversations with security practitioners and leaders.
#incidentresponse #tabletopexercise #aisecurity #ciso #cyberresilience #securityleadership

James Berthoty of Latio on why the endpoint is now the center of AI security, and why most of the real agent risk traces back to blind spots we already had.

I sit down with James Berthoty, who started Latio as an engineer frustrated by how hard it was to figure out what security tools actually did. That turned into a practitioner-focused, buyer-oriented analyst firm. We dig into Latio's second AI security market report. Last year the story was browsers and proxies. This year it's the endpoint, where coding agents now run with the permissions and credentials on developer laptops.

We get into what roughly 400 AI security startups are actually differentiating on and why runtime detection claims need hands-on testing. We also cover why James pushes back on the runtime-first narrative, where the new AI security budget is coming from, and how security leaders should narrow the vendor field before they ever take a demo.

In this episode:
● Why we are in the "pre-HTTPS era" of AI security, and how coding agents turned developer endpoints into the new pressure point
● Dedicated AI security budget jumping from 8% to 37% in a year, and how it flows out of AI governance
● What is already commoditized, and why intent-based runtime detection needs hands-on testing
● Agent blast radius, and why your agent shouldn't be able to drop the prod DB because you shouldn't be able to either
● Why recent AI safety incidents look more like misconfigurations than emergent behavior
● The next M&A wave as a fight over the laptop
● Anthropic's inference hooks, and why frontier labs likely won't absorb the security market
● MCPs and skills as a software supply chain problem with a new delivery mechanism

Chapters:
0:00 Intro
0:31 Why James started Latio
1:00 From browsers to the endpoint
8:15 Do you need a purpose-built AI security platform?
11:52 Where the AI security budget comes from
13:53 400 startups and testing runtime claims
19:28 Blast radius and runtime vs. posture
23:35 M&A and the fight over the endpoint
33:22 Inference hooks and frontier labs vs. vendors
36:03 MCPs, skills, and software supply chain
39:30 Advice for security leaders

Connect with James:
LinkedIn: https://www.linkedin.com/in/james-berthoty
Latio: https://www.latio.com/
AI Security Market Report: https://pulse.latio.tech/p/the-2026-latio-ai-security-market

Resilient Cyber: https://www.resilientcyber.io

Subscribe for more conversations with security practitioners and leaders.
#aisecurity  #agenticai #endpointsecurity #appsec #supplychainsecurity #mcp #cybersecurity

Zack Korman on why the AI safety debate has been captured by an extinction narrative, and why the incident everyone is calling a watershed moment looks a lot more like a preventable security failure.

In this episode I sit down with Zack Korman, co-founder of Embroidery, where he uses AI to monitor AI agents, and former CTO of the cybersecurity company Pistachio. I first came across Zack on X during the controversy over fake SOC 2 reports, and he has since become one of the sharpest critics of how AI safety is being framed, funded, and investigated.

We get into the effective altruism roots of the existential risk movement, why he argues METR's review of the Hugging Face incident was not independent oversight, and what an actual incident response firm would have done differently. Zack makes the case that alignment is one control among many, that there is no "safe" path in AI, only trade-offs, and that the real risk for most organizations is enterprise environments that were never ready for agents in the first place.

In this episode:

● How building an AI insider threat product pulled a developer and CTO into the cybersecurity community, and into picking fights on X
● Effective altruism, longtermism, and how a focus on preventing extinction came to dominate AI safety
● Why "just be careful" misses the point when every path involves trading one risk for another
● Dario Amodei's Pacing the Frontier, independent auditors, and why Zack calls bringing in METR "bring your friend to work day"
● The funding and relationships connecting METR, Redwood Research, Coefficient Giving, and the labs
● True believers versus IPO hype, and why genuine belief does not make someone right
● Why many AI doomers believed we were all going to die before they ever learned about computers
● The Hugging Face investigation, context drop in AI-analyzed transcripts, and treating knowable facts as unknowable
● What Unit 42 or Mandiant would have demanded before putting their name on the report
● Eight layers of failure, from a single package proxy at egress to missing monitoring, classifiers, and kill switches
● Alignment failure versus containment failure, and why alignment belongs inside defense in depth
● The real risks: threat actors misusing AI, enterprise agent deployments, and new attack chains
● Why security's risk-averse, laggard culture may be its biggest vulnerability

Chapters:

0:00 Intro
0:54 Zack's Background
1:43 What Turned a CTO Into an AI Safety Critic
3:31 Effective Altruism and the Extinction Narrative
5:53 No Safe Path, Only Trade-offs
6:40 METR, Redwood, and Bring Your Friend to Work Day
10:23 True Believers, Hype, or Both
12:49 How Doomers Find Their Way Into AI
14:39 Taking AI Risk Seriously When It Is Ideological
16:56 What an IR Firm Would Have Asked
21:10 Eight Layers of Failure in the Hugging Face Incident
22:49 Alignment Failure Versus Containment Failure
24:32 Alignment as One Layer of Defense in Depth
26:22 Why Enterprise Environments Are Not Ready for Agents
29:10 Security's Laggard Culture
29:19 Closing

Connect with Zack Korman:

X: https://x.com/ZackKorman
Embroidery: https://embroidery.io
Website: https://zkorman.com

Resilient Cyber: https://www.resilientcyber.io

Subscribe for more conversations with security practitioners and leaders.

#aisafety #aisecurity #agenticai #incidentresponse #effectivealtruism #cybersecurity

Jonathan Rende of Checkmarx on why the model writing your code cannot also be the control that validates it, and why rules-based and AI-driven scanning turn out to find almost entirely different bugs.

In this episode I sit down with Jonathan Rende of Checkmarx. Jonathan worked with Fortify and SPI Dynamics back in the day, spent most of the last decade leading product teams in developer and DevOps tooling, and came back to security eighteen months ago because, as he puts it, this is the heart of the hurricane. His argument is that AI is a bigger disruption than the internet, SaaS, or mobile were, not because of any single capability, but because it hits roles, process, and productivity all at once.

We get into the two waves he has watched play out with CISOs and their CEOs, why the pendulum has swung back toward program and posture questions in the last quarter, what his research team found when they benchmarked deterministic and probabilistic scanning side by side, and why he thinks agentic AppSec raises the profile of the security team rather than automating it away.

In this episode:

● Why the first half of 2026 became a real inflection point rather than another AI talking point

● The two waves: engineering told to run at any cost, then the pendulum swinging back toward posture and program design

● Why functional AI-generated code and secure AI-generated code are still two different things

● Separation of church and state, and the conflict of interest in letting the model that generates code also validate it

● Benchmarking deterministic and AI-based scanning across dozens of open source projects, and why the overlap stayed consistently under 10%

● Fidelity, F1 scores, and the absence of real standards or shared benchmarks in AppSec

● Why an incentive to reduce risk and an incentive to sell tokens are not the same incentive

● Why agents free AppSec professionals for higher-order work, and why this is not a dark factory

● Shadow IT becoming shadow AI, and early scans where half surfaced models, agents, and MCP servers security teams did not know existed

● Why new threat vectors show up first in fast-moving unregulated companies while regulated ones see more code-level issues

● Low-priority vulnerabilities chained into real impact, and why backlogs now matter as much as incoming code

● What to change first: metrics defined up front, in-workflow AppSec, and security reviews that went from annual to monthly

Chapters:

0:00 Intro
0:18 Fortify, SPI Dynamics, and a decade in developer tooling
1:24 Why he came back to AppSec
2:54 Why the first half of 2026 was the inflection point
5:27 Two waves, and the pendulum swinging back
9:08 Functional AI code versus secure AI code
11:58 Layered defense and the condensed lifecycle
15:04 What agents free AppSec teams to actually do
17:50 Separation of church and state
20:15 Fidelity, F1 scores, and not selling tokens
23:25 New threat vectors and organizational maturity
25:47 Shadow AI and what the inventory scans found
27:58 What to change first in your AppSec program
30:44 Closing

Connect with Jonathan:

LinkedIn: https://www.linkedin.com/in/jonathanrende/

Checkmarx: https://checkmarx.com

Resilient Cyber: https://www.resilientcyber.io

Subscribe for more conversations with security practitioners and leaders.

#appsec #aisecurity #devsecops #shadowai #vulnerabilitymanagement #ciso