The conversations enterprise IT leaders actually listen to.
Volume 342, Issue 3 · 19 articles from 2 issues
Security news roundup covering agentic AI incidents, a Cisco ISE zero-day, and the Plugin4Shell flaw.
This Security Weekly News episode with Doug White and Joshua Marpet covers OpenAI detailing more cases of AI agents taking unauthorized actions, a maximum-severity Cisco ISE zero-day exploited in attacks, the Settra ransomware group's use of MeshAgent RMM, and unauthenticated RCE flaws in The Events Calendar WordPress plugin.
Marpet details Plugin4Shell, in which four AI coding agents (Claude Code, OpenAI Codex, GitHub Copilot, and Gemini CLI) fetch plugins pinned to a commit hash without verifying what they received, letting a repository owner name a branch to look like the pinned hash. Claude Code and Codex are fixed; Copilot has no fix.
Most companies deploy AI in customer service, but lack orchestration needed for measurable returns.
While 98% of companies have deployed AI in customer service operations, only 15% combine agentic AI with cross-departmental orchestration.
Pedro Andrade from Talkdesk explains that orchestration, meaning the coordination of data, systems and workflows, is the critical divider between measurable returns and merely counting deflected calls.
CXA leaders report four times higher net promoter score improvements than less mature adopters. The challenge lies in governance, integration and compliance rather than raw AI capability.
Weekly HPC news roundup covering AI trends, power constraints, fusion energy, mathematics, and Taiwan servers.
This edition of HPC News Bytes examines five significant developments in high-performance computing and AI.
Topics include expanding concerns about AI doomerism, global power limitations affecting AI infrastructure expansion, artificial intelligence applications in managing fusion energy processes, recent AI achievements in solving complex mathematical problems, and Taiwan's growing dominance in AI server production.
The podcast episode is available across major platforms including iTunes, Spotify, and YouTube.
Oxy's Chief AI Officer on running hundreds of AI systems across physical oil and gas operations.
Patrick Bangert leads Occidental Petroleum's Applied AI Center of Excellence, overseeing hundreds of AI projects across physical operations from live camera feeds to predictive equipment maintenance.
Most of Oxy's AI runs on classic machine learning and computer vision rather than large language models, with generative AI reserved for tasks like enterprise search.
Bangert argues a credible AI business case must show material return; a few minutes saved per employee does not meet that bar. He emphasises that AI initiatives succeed or fail on adoption, meaning whether the people doing the work want the tool and can use it daily.
Set up a shared knowledge base so multiple AI agents can access the same notes and decisions.
AI engineer Sajal Sharma demonstrated how to build a shared workspace for multiple agents so they can access common tasks, notes, and decisions.
The approach involves creating an AGENTS.md navigation file, layering daily notes into summaries, bridging files for different AI tools, packaging repeatable tasks as skills, syncing across machines, and ensuring agents reread current state before making changes.
Projects like LangChain's OpenWiki and Y Combinator's GBrain show this pattern is gaining adoption among developers managing multiple AI systems.
At Black Hat USA, OpenAI engineers reconstruct the Hugging Face incident and explore lessons learned about AI safeguards and cyber resilience.
At this Black Hat USA 2026 talk, OpenAI security engineers and researchers will reconstruct the OpenAI-Hugging Face incident and examine its implications for AI security, cyber resilience, and alignment. Throughout the session, they will share insights that address key topics raised by the Black Hat Review Board, including model safeguards, evaluation and containment practices, defensive use cases for AI, and the broader implications of increasingly autonomous systems for the cybersecurity community.
CISOs debate whether AI reduces or worsens alert fatigue in security operations.
Clayton C. Peddy (ABBYY CISO) and Erika Dean (Tricentis CISO) discuss AI's dual role in cybersecurity alerting systems.
While AI can improve threat detection and reduce false positives through machine learning, it may simultaneously increase alert volume if not properly tuned.
The podcast explores how CISOs balance hypervigilance with noise reduction in security operations centers, examining whether AI ultimately helps or exacerbates alert fatigue among security teams.
Horizon3 CEO on automated pentesting for exposure management, plus smart TV privacy and security news.
Horizon3 CEO and co-founder Snehal Antani joins Enterprise Security Weekly to discuss how automated validation and autonomous penetration testing help teams prioritise as vulnerability counts spike, separating exploitable flaws from the rest.
A topic segment examines smart TV privacy in light of allegations that LG collects customer data, and how viewers can opt out.
The weekly enterprise news covers funding and acquisitions, an update on Anthropic's Glasswing project, how fast a mobile phone worm could spread, exposed SSH, large cryptocurrency thefts, SOC 2 assurance gaps, and employers reading AI chat history.
Public sector CISO on risk-based vulnerability prioritisation and the limits of AI in patch management.
Chuck Loring of the Lee County Clerk of Circuit Court & Comptroller joins CISO Stories to discuss how budgets, staffing, legacy technology, and mission priorities shape vulnerability management across public and private sectors.
The conversation challenges the idea that AI alone will solve the problem and identifies where it can genuinely help. Loring argues for risk-based prioritisation as vulnerability volumes continue to surge, and explores what public and private organisations can learn from each other.
The core message: effective vulnerability management is about addressing what poses the greatest risk, not patching everything.
Scott Holleran, CIO of Altus Fire & Life Safety, joins Phil Howard to discuss what gives technology leaders real influence inside the business.
Scott shares lessons from the Army and three private equity-backed technology roles, including why leaders should walk the floor, ride with frontline teams, and map the full process before choosing a system.
He explains how to frame difficult requests through cost, time, return, and tradeoffs; why business-specific data expertise belongs close to operations; how clear behavioral scorecards help teams perform; and why AI will strengthen specialists rather than erase the value of expertise.
Oxide team solves a PCIe clock buffer supply chain crisis through detective work and microscopic rework.
When Oxide discovered their Microchip PCIe clock buffer alternative required termination resistors unlike the original part, a critical flaw emerged during manufacturing.
Engineers identified the issue through oscilloscope measurements, then faced a seemingly impossible task: hand-soldering thirty 0402-sized resistors (roughly pepper-grain sized) on each board.
Supply chain constraints, part availability crises, and a probe-effect measurement error nearly halted production, but skilled rework technicians and precise engineering problem-solving saved the manufacturing timeline.
Jeff Hobbs, Director of IT at SV Labs Corporation, joins Mike Kelley to discuss the leadership skills that matter beyond technical expertise.
Jeff explains how he uses Copilot to tailor communication for different stakeholders, examine unfamiliar security questions, troubleshoot Intune issues, and help a small team close knowledge gaps.
He also discusses the limits: AI output needs review, paid licenses need measures, and easy access to public tools creates data-protection risk. The conversation returns to one idea throughout: technical ability can get a leader noticed, but influence, relationships, and business judgment move the work forward.
IT leader Mikkel Finsen discusses rebuilding core technology systems at Open Door Family Medical Center, emphasizing that healthcare organizations must prioritize simplified infrastructure, business resilience, and strict governance over market hype when implementing AI.
Technology leader Mikkel Finsen joins host Mike Kelley to share insights from his career transition from corporate IT at KPMG to leading digital infrastructure in non-profit healthcare.
He details how his team simplified their network and established clear application recovery priorities after overcoming a major ransomware attack and post-COVID challenges.
Turning to artificial intelligence, Finsen stresses that healthcare AI must focus on practical, patient-centered use cases—such as ambient documentation—governed by human validation, controlled data access, and rigorous staff training.
CIO Gabriel Chappell on balancing ERP modernization, AI governance, and organizational change management.
Gabriel Chappell, CIO at Fabricut Inc., shares how he leads digital transformation in a complex manufacturing environment. He recommends shadowing users to learn actual processes, involving subject matter experts in design and testing, and using business analysts to bridge technical and operational gaps.
Chappell favors AI governance boards that enable responsible experimentation over blanket prohibitions, while tackling shadow IT sprawl and data security risk. His approach prioritizes protecting revenue streams, limiting unnecessary customization, and building collaborative partnerships across the C-suite.
IT leader Rajvir Madan on how self-awareness, culture, and operations close the transformation failure gap.
Rajvir Madan argues that leaders must separate personal worth from project outcomes and stop applying familiar playbooks to unfamiliar contexts.
The conversation lays out three pillars for closing the transformation gap: examining internal patterns, verifying that organizational culture is genuinely ready for change, and aligning operations to measurable business outcomes rather than activity metrics.
Practical advice includes turning abstract values into specific operating principles, building safe feedback loops, and treating unmeasured data as a signal about organizational readiness.
ThreatLocker's Rob Allen on letting employees use AI agents safely, plus the legal risk of AI pentesting agents.
ThreatLocker Chief Product Officer Rob Allen joins Enterprise Security Weekly to discuss how businesses can let employees experiment with agentic AI safely, after OpenClaw became the wake-up call that pushed many to block agents outright.
The topic segment examines AI pentesting agents and the legal liability that arises when an autonomous tester attacks the wrong company. The news segment covers a new MCP standard, AI text watermarking, Cl0p's pattern of targeting file transfer products, the single third party behind several AI sandbox escapes, and the Blue Report 2026 security benchmarks.
Thermal expert Tim Shedd on data center liquid cooling from HPC origins to enterprise AI deployments.
StorageReview interviews thermal management specialist Tim Shedd about data center cooling technologies, tracing the lineage from historical Cray systems to modern Dell PowerEdge servers.
The conversation covers single-phase direct-to-chip cooling, negative-pressure systems, two-phase architectures, and the continued relevance of air cooling.
Topics include cold plate design, contamination management, standardization challenges across vendors, and how regional environmental factors shape facility operations.
The through-line is that optimizing thermal physics and manufacturing practice is what makes high-density accelerator workloads deployable.
A talk on reverse brain drain and what losing top researchers means for US startups and innovation.
Michael Burtov argues that America's decades-long advantage in attracting the world's best scientists, engineers, and entrepreneurs may be reversing.
He examines the forces behind the shift: aggressive international competition for researchers, immigration and research-policy friction in the US, the globalization of venture capital and technology infrastructure, and the growing ability to build world-class companies anywhere.
The talk connects research funding to startup creation years later, suggests scientists may soon be recruited like professional athletes, and argues venture capital will increasingly follow researchers. His closing point is that the real loss is the company never founded because its founder built elsewhere.
Marsh, Mimecast, and Zscaler leaders on board-level cyber risk, agentic insiders, and ransomware targeting managers.
Dan Bowden, Global Business CISO at Marsh, joins Business Security Weekly to discuss connecting cyber risk to the outcomes boards care about - resilience, financial exposure, regulatory impact, and reputation - positioning cyber as an enterprise risk rather than a technical one, drawing on Marsh's Cyber Catalyst research and Global Cyber Claims Report.
In a Black Hat interview, Mimecast CISO Leslie Nielsen argues agentic AI is an extension of human risk rather than a standalone category, so existing people-focused controls are the right foundation.
Zscaler's Brett Stone-Gross presents ThreatLabz research on which employees ransomware crews compromise first and why managers make valuable initial targets.
More from CxO Podcasts · All IT section newsletters · This week