| ๐ก๏ธ What I Share Here | Thoughts on cybersecurity trends & real-world defense strategies Insights from open source projects and community engagement Notes on programming (learning Python & tackling The Odin Project) Reflections from Stoicism & philosophy applied to daily life |
Most developers are trained to write code โ but senior engineers are defined by how well they can read and understand it.
The infographic breaks down why code comprehension is becoming the real differentiator in modern software systems.
Cross posted in LinkedIn:
https://www.linkedin.com/feed/update/activity:7461389515111170048
#SecureCoding #CodeReview #AIinSoftware #EngineeringLeadership
๐ฆ Firefox Hardening: The AI Breakthrough
Mozilla just revealed how they patched 423 vulnerabilities in a single monthโmore than the previous 15 months combined. Here is the "too long; didn't read" on their new security strategy:
The Tool: Mozilla built an "agentic harness" using Claude Mythos Preview. It doesn't just scan code; it acts like a security researcher to find, reproduce, and verify bugs.
The "Cold Cases": The AI unearthed bugs that survived for 20 years, including deep-rooted flaws in XSLT and legacy HTML elements that traditional tools missed.
Sandbox Safety: A huge win for privacyโthe system focused on "sec-high" issues to prevent attackers from escaping the browser sandbox to access your OS.
The Verdict: This isn't just hype. Mozilla is now integrating this AI-driven auditing directly into their daily build process to catch bugs before they ever reach a user.
The Bottom Line: Firefox just got significantly harder to hack, thanks to a massive AI-led cleanup of two decades of legacy code.
Read the full technical breakdown: ๐ https://hacks.mozilla.org/2026/05/behind-the-scenes-hardening-firefox
#Firefox #CyberSecurity #Mozilla #InfoSec #AI #ClaudeMythos #OpenSource #WebDev
A kernel bug sat in plain sight for 8 years. AI found it in an hour.
Wrong takeaway: AI is making attackers faster.
Better takeaway: our security model assumes too much about patching.
Assume latent flaws exist.
Design around containment, isolation, and resilience.
AI isnโt changing vulnerability physics.
Itโs exposing reality faster.
More thoughts here:
LinkedIn: ๐ https://www.linkedin.com/posts/dinesh-mr_73-sounds-impressive-until-you-ask-what-activity-7458128840872349696-kpVc
Cyber governance is not solved by hiring one โboard cyber expert.โ
The bigger shift: โ executive accountability
โ AI security by design
โ resilience over checkbox compliance
Boards shouldnโt ask, โAre we compliant?โ
They should ask, โCan we take a hit and keep operating?โ
Extract from๐ https://hbr.org/2026/04/boards-are-falling-short-on-cybersecurity
Despite boards placing greater emphasis on cyber risk, their ability to mitigate it is improving slowly and marginally. There are three prominent factors driving this problem: 1) thereโs a lack of cybersecurity expertise; 2) board-level conversations about AI ignore security; and 3) boards mistake regulatory compliance for security. There are concrete steps boards can take to address each factor. First, rather than increasing the number of directors with cybersecurity expertise, boards should concentrate their cybersecurity responsibilities on selecting and overseeing effective cybersecurity executives. Second, boards must treat AI as both a strategic opportunity as well as a cybersecurity and governance risk. Finally, boards should view cybersecurity less as a compliance-driven regulatory issue and more as a competitive, operational resilience issue, where market incentives and organizational accountability drive stronger security outcomes than government-imposed rules.
73% sounds impressive โ until you ask what it measures.
UK AISI tested Claude Mythos Preview on cyber tasks. Headline: 73% on expert CTFs. But CTFs are puzzles, not networks.
The real test โ a 32-step simulated attack โ was solved 3/10 times against an undefended range, with operator direction and heavy compute.
Four questions the report doesn't answer: noise, cost, operator guidance, OT pivot.
Full breakdown: [https://www.linkedin.com/posts/dinesh-mr_73-sounds-impressive-until-you-ask-what-activity-7458128840872349696-kpVc]
Bad Connection
Uncovering Global Telecom Exploitation by Covert Surveillance Actors https://citizenlab.ca/research/uncovering-global-telecom-exploitation-by-covert-surveillance-actors/
An investigation by the Citizen Lab Team, which uncovers two sophisticated telecom surveillance campaigns and, for the first time, directly links real-world attack traffic to mobile operator signalling infrastructure.
#CyberSecurity #ThreatIntelligence #Surveillance #TelecomSecurity #MobileSecurity #SS7 #NetworkSecurity #CyberEspionage #CitizenLab #Infosec #Privacy #DigitalRights #CyberResearch #SignalInfrastructure #Telecom
โSignal Liveโ โ a live social feed concept I explored as part of my VibeCoding series.
Idea: real-time, lightweight sharing.
Question: does this actually add utility, or just accelerate noise?
Curious how others think about this trade-off.
Signal/live is at ๐: https://signal-feed-8pl.pages.dev/
Blog post:
๐https://mrdee.in/vibecoding/vibecoding-021-signal-live/
