PRESENTED BY
Cyber AI Chronicle
By Simon Ganiere · 23rd February 2025
Welcome back!
Project Overwatch is a cutting-edge newsletter at the intersection of cybersecurity, AI, technology, and resilience, designed to navigate the complexities of our rapidly evolving digital landscape. It delivers insightful analysis and actionable intelligence, empowering you to stay ahead in a world where staying informed is not just an option, but a necessity.
Table of Contents
TL;DR
It’s been a busy week and I wasn’t able to really focus on a long article this week. That being said still sharing a few shorter thoughts or comments on what I read or found interesting this week:
Last week I covered the DASF v2 framework, one of the challenges with those frameworks is where to start as no matter what there is still quite a level of complexity to get started. I took a Quick Look at the mitigation listed and which one will give you the biggest impact » READ MORE
Microsoft announced Majorana 1, the first quantum processor using topological qubits, promising more stable and scalable quantum computing. While the breakthrough could accelerate quantum error correction and pose future cryptographic risks, experts caution it’s still early—claims of practical use remain unproven and scaling to millions of qubits is a massive challenge. CISOs should monitor developments but focus on long-term quantum resilience and post-quantum cryptography planning, rather than reacting to the hype. » READ MORE
Still on the topic of quantum, check out this article on the potential impact of Quantum Computing on Bitcoin.
Apple has removed its Advanced Data Protection feature for iCloud in the United Kingdom following a secret government order demanding backdoor access to encrypted user data, sparking concerns about privacy rights and setting a precedent for global data security. » READ MORE
I’m exploring Perplexity to get summary and context on key topics. Here is a page that I have curated in a few minutes about the hack of Bybit which is one of the biggest heist in the crypto industry…a whopping $1.5 billion was stolen! That being said it looks like Bybit handle the whole situation pretty well. Get the CEO to communicate, live stream with the CEO setup quickly, transparent information, etc. Impressive! and very much align with the fact that a company is not judged on the incident itself but on how they are handling it.
I’m also still exploring some of the workflow tools like n8n to automate some of the work for this newsletter and other ideas…just not enough time in the week to be able to do everything 😅
DASF: Not Sure Where to Begin? Start Here!
I’ve been spending more time with the DASF framework lately, and I have to say—it’s impressive. It pulls together elements from other frameworks in a way that’s simple to understand. But let’s be honest: the big question is always the same—where do you start? What should you prioritize? And most importantly, what actions will give you the biggest security payoff?
While the answer depends on your specific environment, I crunched the numbers and took a close look at which DASF mitigation controls cover the most ground. The results? Access controls top the list—and by a significant margin.
It might seem obvious, but this highlights a crucial point: AI security starts with the basics. If your access controls are shaky, deploying AI will only amplify those issues. Fixing them early on isn’t just a good idea—it’s essential.
Here are the top mitigations with the broadest risk coverage:
Mitigation | Potential Security |
|---|---|
DASF-5 Control access to data and other objects | 27 |
DASF-1 SSO with IdP and MFA | 23 |
DASF-2 Sync users and groups | 22 |
DASF-3 Restric access using IP access lists | 22 |
DASF-4 Restrict access using private link | 22 |
DASF-31 Secure model serving endpoints | 15 |
DASF-55 Monitor audit logs | 15 |
DASF-42 Employ data-centric MLOps and LLMOps | 14 |
DASF-32 Streamline the usage and management of various large language model (LLM) providers | 12 |
DASF-37 Set up inference tables for monitoring and debugging models | 12 |
DASF-24 Control access to models and model assets | 11 |
DASF-30 Encrypt models | 11 |
DASF-60 Rate limit number of inference queries | 10 |
Access control measures (DASF-1 through DASF-5) dominate the top five. It’s a reminder that no matter how advanced the technology, strong fundamentals still win the day.
👉 Pro tip: Fixing access controls now will save you headaches later—especially as your AI initiatives scale.
If you missed last week’s newsletter, I covered more on DASF and how it stacks up against other frameworks. Check it out if you want to dive deeper »
Microsoft’s Quantum Leap: Major Breakthrough or Hype?
This week, Microsoft made waves by unveiling Majorana 1, the world’s first quantum processor powered by topological qubits. The headlines were flashy—“new state of matter” and promises of fault-tolerant quantum computing that could one day revolutionize industries from cryptography to drug discovery. Big claims! But let’s cut through the buzz and focus on what CISOs should really take away from this.
What Was Announced?
Topological Qubits: Microsoft claims to have created qubits based on a new state of matter—achieved by engineering materials called topoconductors. These qubits are designed to be more stable and less prone to error than traditional ones
Majorana Zero Modes (MZMs): The building blocks of these qubits are exotic quasiparticles known as MZMs. They hold promise for improving quantum error correction, a major hurdle in scaling quantum computers.
Roadmap: Microsoft’s ambitious goal? Scale to one million qubits and build a fault-tolerant quantum computer—not in decades, but in a matter of years.
Why Should You Care?
Quantum computing threatens the foundations of modern encryption. A fully functional quantum machine could, in theory, break widely used cryptographic systems like RSA. This makes the race toward quantum supremacy more than just a tech curiosity—it’s a national security issue
But… Is This the Game-Changer It Sounds Like?
Let’s pump the brakes. While the announcement is exciting, there are key limitations and skepticism to consider:
Early Days: Microsoft has built only a handful of topological qubits. They aren’t performing useful calculations yet
Peer Review Nuances: Although a paper was published in Nature, the peer-reviewed data doesn’t fully confirm the existence of the Majorana modes that Microsoft claims. Some experts argue the announcement is ahead of the science
Scaling Challenges: Moving from a few qubits to millions is a massive engineering feat. The history of quantum computing is littered with “breakthroughs” that took decades to materialize—if they did at all
What Should CISOs Do Now?
Stay Informed: Quantum is no longer just a theoretical threat. Track developments and assess your cryptographic posture.
Start Quantum-Resistant Planning: Standards bodies are working on post-quantum cryptography (PQC). CISOs should consider how to transition critical systems.
Beware the Hype: While the tech press loves bold claims, practical, enterprise-impacting quantum computers are still years away. Don’t let panic drive premature investments.
👉 Bottom Line: Microsoft’s announcement is a significant scientific milestone—but not a reason to overhaul your security roadmap just yet. Focus on quantum resilience and long-term cryptographic planning while keeping an eye on how this unfolds.
SPONSORED BY
There’s a reason 400,000 professionals read this daily.
Join The AI Report, trusted by 400,000+ professionals at Google, Microsoft, and OpenAI. Get daily insights, tools, and strategies to master practical AI skills that drive results.
Worth a full read
Will Quantum Computing Break Bitcoin?
Key Takeaways
Quantum computing's potential impact on Bitcoin is a concern for its users.
Quantum computers could significantly speed up problem-solving times.
Quantum computing could pose a threat to Bitcoin wallets and their related addresses.
Quantum computers are not yet a significant threat to Bitcoin mining.
Quantum computers could enhance the effectiveness of network attacks.
Quantum computing is still in its infancy and its full potential is unknown.
Wallet migration is a short-term solution to Bitcoin's quantum vulnerability.
Long-term solutions to Bitcoin's quantum vulnerability involve changing consensus to update the signature algorithm.
Lost or forgotten bitcoins pose a unique challenge in the face of potential quantum attacks.
The complete migration of Bitcoin addresses to quantum-resistant types will be a long and arduous process.
Agentic AI – Threats and Mitigations
Key Takeaways
Agent autonomy transforms traditional vulnerabilities, requiring novel security approaches and adaptive defenses.
Memory access controls are vital to prevent knowledge corruption and persistent exploitation.
Tool misuse risks highlight the need for strict input validation and permission constraints.
Cascading agent interactions amplify minor manipulations into systemic failures.
Rogue agents pose significant risks in distributed systems, emphasizing continuous monitoring.
Identity spoofing challenges demand robust multi-factor authentication and behavioral profiling.
Human oversight, while essential, can be a bottleneck vulnerable to decision fatigue exploits.
Overconfidence in AI outputs without independent validation fosters dangerous complacency.
Privilege management must be dynamic, context-aware, and minimize excessive access rights.
Trust relationships between agents must be validated to prevent inter-agent exploitation.
PWC: Agentic AI - the new frontier in GenAI
Key Takeaways
Agentic AI transforms industries by enabling autonomous decision-making and real-time problem-solving.
Multimodal AI enhances human-AI interactions, driving personalised and efficient customer experiences.
Outcome-based pricing models shift focus from tools to results, aligning costs with tangible business value.
Early AI adoption creates competitive barriers, positioning leaders as market innovators.
Ethical AI deployment requires proactive governance, transparency, and stakeholder involvement.
Human-in-the-loop systems ensure balance between autonomy and accountability.
AI-driven process automation frees employees to focus on strategic, creative tasks.
Robust data infrastructure and quality underpin effective agentic AI solutions.
Cross-functional collaboration is essential for comprehensive AI strategy execution.
Continuous learning and adaptability enable organisations to stay ahead in the evolving AI landscape.
Wisdom of the week
“Which is more important,” asked Big Panda, “the journey or the destination? “The company.” said Tiny Dragon.
Contact
Let me know if you have any feedback or any topics you want me to cover. You can ping me on LinkedIn or on Twitter/X. I’ll do my best to reply promptly!
Thanks! see you next week! Simon