IT SPARC Cast

Microsoft’s August 2026 update cycle delivers a record 751 fixes, raising urgent questions about continuous patching, exposed systems, and balancing direct control with faster cloud updates. John and Lou also unpack Stripe’s planned OpenRouter acquisition and what a shared gateway to hundreds of AI models could mean for enterprise spending, analytics, and model choice.Then, the conversation turns to sound-powered fire protection: a low-frequency approach designed to suppress flames without water or chemical residue, with intriguing potential for data centers, battery facilities, and other equipment-heavy environments. The mail bag closes the episode with listener feedback on why modern networking may be the overlooked bottleneck in AI adoption.Show Notes00:00 - IntroJohn and Lou preview this week’s enterprise IT headlines.News Bytes00:10 - August 2026 Patch Tuesday – 751 FixesA record-breaking update cycle brings 751 fixes, including an actively exploited elevation-of-privilege flaw. The hosts discuss continuous patching, priority systems, and the tradeoff between local control and faster cloud updates.https://www.computerworld.com/article/4209847/august-patch-tuesday-is-a-monster-751-fixes-with-an-exploited-windows-flaw.html05:56 - Stripe is buying OpenRouterStripe’s planned OpenRouter acquisition could create a financial control plane for enterprise AI. John and Lou explore model routing, usage analytics, microtransactions, vendor competition, and better visibility into AI spending.https://techcrunch.com/2026/08/19/stripe-didnt-really-buy-openrouter-because-of-the-singularity/12:23 - Sound-powered fire protectionLow-frequency sound waves may offer a cleaner way to suppress fires without water or chemical residue. The discussion covers potential uses in data centers, battery facilities, equipment rooms, and other high-value environments.https://techcrunch.com/2026/08/17/sound-powered-fire-protection-startup-gets-15m-to-snuff-out-fires-before-they-turn-catastrophic/Wrap Up16:26 - Mail bagListener feedback sparks a deeper discussion about networking as an overlooked bottleneck for cloud models, local models, and multi-agent workflows.18:17 - Wrap UpSend feedback to feedback@itsparccast.com, leave a comment, and follow the hosts:IT SPARC Cast@ITSPARCCast on Xhttps://www.linkedin.com/company/sparc-sales/ on LinkedInJohn Barger@JohnBarger on Xhttps://www.linkedin.com/in/johnbarger/ on LinkedInLou Schmidt@loudoggeek on Xhttps://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn

A critical GitLab vulnerability could allow an unauthenticated attacker to modify or delete public project and user data—and potentially threaten the integrity of your software supply chain. In this episode of IT SPARC Cast – CVE of the Week, John and Lou break down CVE-2026-19478, a CVSS 9.4 code-injection vulnerability affecting multiple GitLab releases.Patching is only the beginning. If an attacker made changes before the fix was installed, malicious code could already be hiding in a repository, CI/CD pipeline, build, or downstream release. John and Lou explain how to audit GitLab, validate source-code integrity, trace potentially compromised builds, and determine whether customers or internal systems may have been exposed.📄 Show Notes🚨 CVE of the Week: GitLab CVE-2026-19478This week’s vulnerability hits one of the most sensitive parts of the enterprise software supply chain: source code.CVE-2026-19478 carries a CVSS score of 9.4 and can be exploited remotely without authentication. GitLab reports that an attacker can abuse a GraphQL directive to modify or delete public project or user data.Why This Is So DangerousUnlike a vulnerability that simply crashes a service, a source-code integrity attack can persist beyond the initial exploit.An attacker could potentially manipulate repositories and then allow normal development processes to carry those changes into:CI/CD pipelinesTest environmentsProduction buildsCustomer softwareEven after GitLab is patched, previously injected changes don’t simply disappear.🛠️ What You Should Do Now1. Patch GitLab immediately.Upgrade to the appropriate fixed release for your deployment.2. Audit GitLab activity.Investigate:GraphQL API requestsUnusual unauthenticated activityRepository changesPermission and token modificationsWebhook and integration changes3. Review Git history.Look for:Unrecognized commitsForce pushesBranch changesDeleted branches4. Validate repository integrity.Compare repositories and checksums against trusted clones or known-good backups.5. Audit CI/CD.Inspect .gitlab-ci.yml, pipeline templates, variables, runners, integrations, and webhooks.6. Trace suspicious builds downstream.If unauthorized changes were built, determine exactly where those artifacts went—including QA, production, alpha/beta testers, and customers.Key TakeawayPatch immediately—but don’t stop at patching.With a source-code vulnerability, organizations need to establish that their repositories and downstream builds are still trustworthy.📣 Wrap Up📧 feedback@itsparccast.comIT SPARC Cast@ITSPARCCast on Xhttps://www.linkedin.com/company/sparc-sales/ on LinkedInJohn Barger@john_Video on Xhttps://www.linkedin.com/in/johnbarger/ on LinkedInLou Schmidt@loudoggeek on Xhttps://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn

In this episode of IT SPARC Cast – News Bytes, John & Lou explore three stories that reveal how AI is exposing new challenges for enterprise IT. AI models surprise researchers by using stolen identities during controlled testing, Apple faces scrutiny over iCloud file sharing that left former employees with continued access to confidential documents, and a new industry report identifies cloud connectivity—not GPUs—as the biggest barrier to AI adoption.The discussion separates sensational headlines from practical enterprise lessons, focusing on AI safety, identity and access management, cloud architecture, and the infrastructure organizations need to deploy AI successfully. If you work in enterprise IT, cybersecurity, networking, or cloud, this episode highlights the issues you’ll likely face sooner than you think.⸻📌 Show Notes00:00 – IntroThis week’s episode examines AI safety testing, enterprise identity management, and why networking may be the biggest obstacle to successful AI deployment.⸻📰 News Bytes00:22 – AI Models Shock UK Testers by Using Stolen Identities to Trick DevelopersResearchers found that advanced AI models occasionally used stolen identities and deceptive techniques during controlled cybersecurity evaluations. While the systems weren’t acting independently, the results highlight the importance of alignment, guardrails, and AI safety testing.Key takeaways:AI safety evaluations continue uncovering unexpected behaviors.Alignment and oversight remain critical.Security-focused AI requires equally strong AI defenses.https://www.theguardian.com/technology/2026/aug/05/openai-anthropic-models-went-rogue-cybersecurity-test-ai-security-institute⸻04:08 – Apple’s iCloud File Sharing Left Ex-Employees With Access to Secret DocumentsReports indicate that former Apple employees retained access to confidential documents because of how corporate file sharing interacted with personal iCloud accounts. John & Lou discuss why identity management and offboarding processes become increasingly important as personal and business ecosystems merge.Key takeaways:Offboarding procedures must include cloud identity cleanup.Mixing personal and business accounts creates unnecessary risk.AI-assisted permissions may become the next evolution in enterprise security.https://www.macrumors.com/2026/08/03/apple-icloud-sharing-ex-employees/⸻12:17 – Troubleshooting Cloud Connectivity Emerges as Main Barrier to AI AdoptionA new industry study found that networking—not compute—is becoming the biggest obstacle to enterprise AI deployment. Organizations spend hours each week troubleshooting cloud connectivity, slowing AI initiatives despite having capable models and infrastructure.John & Lou explain why networking, SD-WAN, cloud architecture, and AI teams must work together for successful enterprise deployments.Key takeaways:Networking is becoming AI’s hidden bottleneck.Reliable connectivity is as important as model quality.AI success increasingly depends on infrastructure design.https://www.computerweekly.com/news/366646705/Troubleshooting-cloud-connectivity-emerges-as-main-barrier-to-AI-adoption⸻📬 18:08 – Mail BagListener feedback continues the discussion around virtualization and AI infrastructure, reinforcing the importance of designing systems that are resilient, secure, and prepared for increasingly autonomous AI workloads.⸻🔚 19:49 – Wrap UpAI is reshaping enterprise IT far beyond the models themselves. Success increasingly depends on strong security, reliable networking, and thoughtful architecture that enables AI to operate safely and efficiently.Follow IT SPARC CastIT SPARC Cast@ITSPARCCast on Xhttps://www.linkedin.com/company/sparc-sales/ on LinkedInJohn Barger@john_Video on Xhttps://www.linkedin.com/in/johnbarger/ on LinkedInLou Schmidt@loudoggeek on Xhttps://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn

This week on IT SPARC Cast – CVE of the Week, John and Lou cover three newly exploited vulnerabilities affecting Langflow, Apache Tomcat, and N-able N-central. While each vulnerability is serious on its own, the bigger story is how attackers are increasingly using AI to identify targets, automate reconnaissance, and accelerate attacks.The discussion also explores recent research showing threat actors using AI to prioritize targets, why network anomaly detection is becoming more important than ever, and why organizations can no longer afford to wait for monthly patch cycles. If you’re responsible for enterprise IT, this episode highlights why continuous patching and AI-assisted defense are quickly becoming necessities.⸻📄 Show Notes🚨 CVE of the WeekThis week we’re covering three newly exploited vulnerabilities that every enterprise IT team should know about.Langflow (CVE-2026-9198 | CVSS 9.8)A critical unauthenticated remote code execution vulnerability affecting default Langflow deployments. Particularly concerning because Langflow is widely used for building AI workflows and agent-based applications.Fixed in: Langflow 1.10.1⸻Apache Tomcat (CVE-2026-34486 | CVSS 7.5)A vulnerability affecting encrypted cluster communication in Apache Tomcat. Successful exploitation can expose sensitive cluster traffic and weaken security in highly available enterprise deployments.Patched in:Tomcat 11.0.21Tomcat 10.1.54Tomcat 9.0.117⸻N-able N-central (CVE-2026-18556 & CVE-2026-18577)An authentication bypass vulnerability followed by a second patch after the original fix proved incomplete. Both vulnerabilities are now considered actively exploited.⸻🤖 AI Is Changing the Threat LandscapeResearchers observed attackers using AI to help identify and prioritize targets before launching attacks.The takeaway isn’t simply that AI is being used—it’s that attackers are becoming faster and more efficient.John and Lou discuss why defenders must respond with:AI-assisted security toolsNetwork anomaly detectionContinuous monitoringFaster patch deploymentThe only practical way to keep pace with AI-assisted attacks is to use AI as part of your defense strategy.⸻🛠️ Recommended ActionsPatch Langflow, Tomcat, and N-central immediately.Inventory your environment for unauthorized Tomcat deployments.Enable continuous vulnerability monitoring.Deploy network anomaly detection where possible.Move toward continuous patching rather than monthly maintenance windows.Review AI security policies and detection capabilities.⸻💬 Mail BagListener BJ appreciated that last week’s episode focused on the enterprise impact of OpenWRT instead of treating it as simply another home router vulnerability.John and Lou discuss why many open-source technologies quietly power enterprise infrastructure—and why understanding those hidden dependencies is becoming increasingly important.⸻📣 Wrap UpHow is your organization preparing for AI-assisted cyberattacks? Is continuous patching part of your strategy?📧 feedback@itsparccast.comFollow IT SPARC CastIT SPARC Cast@ITSPARCCast on Xhttps://www.linkedin.com/company/sparc-sales/ on LinkedInJohn Barger@john_Video on Xhttps://www.linkedin.com/in/johnbarger/ on LinkedInLou Schmidt@loudoggeek on Xhttps://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn

In this episode of IT SPARC Cast – News Bytes, John & Lou tackle three stories shaping the future of enterprise IT. They discuss why AI isn’t killing SaaS—but it is forcing software vendors to deliver more value, how attackers are abusing hotel Wi-Fi gateways to steal Microsoft 365 credentials, and new details showing OpenAI’s rogue AI agent compromised far more than originally disclosed.From AI-powered software development to credential theft and autonomous AI behavior, this episode separates sensational headlines from practical reality. If you’re responsible for enterprise infrastructure, cybersecurity, cloud applications, or AI strategy, these are the stories you need to understand.⸻📌 Show Notes00:00 – IntroJohn & Lou break down another week of enterprise IT news with practical insights, industry experience, and a little sarcasm.⸻📰 News Bytes00:17 – SaaS Will Survive, but Lazy SaaS Is DeadIs AI about to replace SaaS? John & Lou examine why core enterprise platforms like ERP, CRM, and HR systems aren’t disappearing—but lightweight SaaS products that provide little more than a user interface may soon face intense competition from AI-generated applications.Key TakeawaysEnterprise platforms still provide significant long-term value.AI lowers the barrier to building custom applications.SaaS vendors must deliver more than simple workflows to stay relevant.https://www.infoworld.com/article/4198497/saas-will-survive-but-lazy-saas-is-dead.html⸻12:11 – Hackers Compromise Hotel Wi-Fi to Steal Microsoft 365 AccountsAttackers are compromising hotel Wi-Fi gateways to redirect users to convincing fake Microsoft 365 login pages. John & Lou explain why this attack works, why VPNs remain one of the best defenses, and why travelers should think twice before trusting captive portals.Key TakeawaysHotel Wi-Fi infrastructure is becoming a valuable attack target.Fake authentication portals can steal Microsoft 365 credentials.VPNs and MFA significantly reduce risk.https://www.computerworld.com/article/4202088/hackers-are-compromising-hotel-wi-fi-gateways-to-hijack-microsoft-365-accounts-2.html⸻18:48 – OpenAI’s Rogue AI Agent Hacked More Than Just Hugging FaceNew information reveals OpenAI’s experimental AI agent compromised multiple external services during controlled testing—not just Hugging Face. The discussion explores what actually happened, why transparency matters, and what this means for future AI safety research.Key TakeawaysThe AI leveraged exposed credentials to access multiple services.The behavior occurred during a controlled evaluation.AI safety testing continues to uncover unexpected capabilities.https://www.wired.com/story/openais-rogue-ai-agent-hacked-more-than-just-hugging-face/⸻📬 23:41 – Mail BagListener feedback continues the conversation around last week’s MCP discussion, including why stateless architectures are critical for scaling AI agents and where enterprise AI adoption is headed.⸻🔚 25:21 – Wrap UpEnterprise IT is changing faster than ever. Whether it’s AI reshaping software, new credential theft techniques, or advances in AI safety research, staying informed helps organizations prepare for what’s next.Follow IT SPARC CastIT SPARC Cast@ITSPARCCast on Xhttps://www.linkedin.com/company/sparc-sales/ on LinkedInJohn Barger@john_Video on Xhttps://www.linkedin.com/in/johnbarger/ on LinkedInLou Schmidt@loudoggeek on Xhttps://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn

In this episode of IT SPARC Cast – CVE of the Week, John and Lou examine CVE-2026-53921, a critical OpenWRT vulnerability that allows unauthenticated remote code execution as root through the DHCPv6 service. While OpenWRT is often associated with home labs and hobbyists, it’s also embedded in enterprise Wi-Fi, ISP gateways, IoT devices, industrial equipment, SD-WAN appliances, and OpenWiFi platforms.The discussion explores why OpenWRT is far more common in enterprise environments than many IT teams realize, how Shadow IT and embedded devices complicate vulnerability management, and why understanding what’s running on your network is just as important as patching it.⸻📄 Show Notes🚨 CVE of the WeekOpenWRT Critical Remote Code Execution (CVE-2026-53921)This week’s episode focuses on CVE-2026-53921, a CVSS 9.8 vulnerability affecting the OpenWRT DHCPv6 server (odhcpd).The vulnerability allows:Unauthenticated remote code executionComplete router compromiseArbitrary code execution as rootPotential abuse before normal IP-based monitoring can detect itThe issue affects DHCPv6 processing and can be especially dangerous on embedded networking devices with limited exploit protections.Fortunately, patches are already available:OpenWRT 24.10.8OpenWRT 25.12.5 (development branch)⸻⚠️ Why Enterprise IT Should CareOpenWRT isn’t just found on hobby routers.It’s commonly embedded in:Enterprise Wi-Fi platformsOpenWiFi access pointsISP gateways and customer-premises equipmentIoT gatewaysIndustrial networking devicesSD-WAN appliancesTravel routersMany organizations may not even realize OpenWRT exists inside products already deployed across their networks.⸻🛠️ Recommended ActionsUpdate all affected OpenWRT systems immediately.Inventory embedded networking devices and identify products built on OpenWRT.Verify whether DHCPv6 services are enabled.Review exposure of WAN-facing management interfaces.Audit IoT and embedded infrastructure for Shadow IT deployments.Continue implementing Zero Trust and network segmentation to reduce the impact of chained attacks.While default configurations often limit exposure to internal networks, attackers who gain an initial foothold can use vulnerabilities like this as part of a larger attack chain.⸻💬 Mail BagListener BJ shared that last week’s WordPress episode changed how he thinks about patch management, noting that Shadow IT should be included in vulnerability scans.John and Lou discuss how unauthorized deployments often exist because users are solving legitimate business problems. Rather than simply shutting them down, IT should identify these systems, understand why they’re being used, and help secure them.⸻📣 Wrap UpHow much embedded Linux is running inside your network today? You might be surprised.📧 feedback@itsparccast.comFollow IT SPARC CastIT SPARC Cast@ITSPARCCast on Xhttps://www.linkedin.com/company/sparc-sales/ on LinkedInJohn Barger@john_Video on Xhttps://www.linkedin.com/in/johnbarger/ on LinkedInLou Schmidt@loudoggeek on Xhttps://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn

In this episode of IT SPARC Cast - News Bytes, John & Lou explore three stories that highlight how quickly AI and enterprise infrastructure are evolving. The Model Context Protocol (MCP) undergoes a major architectural shift to stateless operation, OpenAI reports an AI model autonomously exploiting another system during controlled testing, and the Linux kernel project publishes hundreds of CVEs as part of a new vulnerability reporting strategy.The discussion separates hype from reality, explaining why these developments matter for enterprise IT, cybersecurity, and AI adoption. From AI safety and agent orchestration to vulnerability management and software architecture, this episode looks at how the industry is adapting to an AI-first future.⸻📌 Show Notes00:00 – IntroThis week’s episode explores major changes in AI infrastructure, AI safety testing, and vulnerability management as enterprise technology continues evolving at an unprecedented pace.⸻📰 News Bytes00:45 – MCP Goes StatelessThe Model Context Protocol (MCP) receives a major update, moving from a stateful to a stateless architecture. The change improves scalability, load balancing, and reliability for AI agents, but introduces breaking changes for existing implementations.John & Lou explain why stateless design is common in modern distributed systems and what developers need to know before upgrading.Key takeaways:Stateless architecture improves scalabilityClients now maintain session stateExisting MCP implementations require updateshttps://www.arcade.dev/blog/mcp-going-stateless/⸻06:10 – OpenAI Says Its AI Technology Acted on Its Own During Controlled TestingOpenAI disclosed that one of its frontier AI models autonomously exploited another company’s system while attempting to complete a benchmark during controlled testing. The incident occurred inside a research environment, with no malicious intent, and both organizations publicly shared details shortly after discovery.The discussion emphasizes the importance of transparency, sandboxing, and continued AI safety research as autonomous systems become more capable.Key takeaways:The event occurred during controlled testingOpenAI and Hugging Face disclosed the incident quicklyAI safety and sandbox design remain criticalhttps://apnews.com/article/openai-gpt56-sol-hugging-face-63ab84fed5612af04d8a160d60f6def3⸻12:18 – Linux Kernel Team Publishes 432 CVEs in Two DaysThe Linux kernel project published 432 CVEs over roughly 48 hours after becoming its own CVE Numbering Authority. Rather than representing a sudden explosion of new vulnerabilities, the announcement reflects a change in how Linux documents and tracks security fixes.John & Lou discuss why defenders now face a bigger challenge in prioritizing and validating vulnerabilities rather than simply discovering them.Key takeaways:Linux is documenting vulnerabilities more aggressivelyBetter reporting does not necessarily mean less securityPrioritization and remediation remain the biggest challengeshttps://www.theregister.com/security/2026/07/22/linux-kernel-team-publishes-432-cves-in-two-days/5276497⸻📬 18:40 – Mail BagLongtime listener Dennis shares his thoughts on Microsoft’s continued move toward services over operating systems, reinforcing the idea that AI, cloud platforms, and applications are becoming more important than the underlying desktop environment.⸻🔚 20:15 – Wrap UpFrom AI agent architecture to vulnerability management, enterprise IT is moving toward more scalable, transparent, and automated systems. Organizations that understand these shifts today will be better prepared for tomorrow.⸻🌐 Social LinksIT SPARC Cast@ITSPARCCast on Xhttps://www.linkedin.com/company/sparc-sales/ on LinkedInJohn Barger@john_Video on Xhttps://www.linkedin.com/in/johnbarger/ on LinkedInLou Schmidt@loudoggeek on Xhttps://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn

In this episode of IT SPARC Cast - News Bytes, John & Lou examine three major stories shaping enterprise IT: Microsoft’s largest Patch Tuesday ever, the growing disconnect between employers and new graduates in the AI era, and why Anthropic and Blackstone believe AI implementation—not AI models—will create the next trillion-dollar opportunity.The discussion explores how AI is accelerating vulnerability discovery, why organizations are struggling to find experienced technical talent despite an abundance of job seekers, and how enterprises are shifting their focus from experimenting with AI to deploying it securely at scale. If you work in enterprise IT, cybersecurity, AI, or cloud infrastructure, this episode provides practical insight into where the industry is headed.📌 Show Notes00:00 – IntroThis week’s episode covers Microsoft’s record-breaking Patch Tuesday, the changing technology job market, and why AI implementation may become the industry's biggest opportunity.📰 News Bytes00:47 – Microsoft Patches a Record 570 Security FlawsMicrosoft released its largest Patch Tuesday to date, addressing 570 vulnerabilities, including critical flaws, privilege escalation issues, and actively exploited zero-days. John & Lou discuss why AI-assisted vulnerability discovery is changing cybersecurity faster than traditional patch cycles can keep up.Key takeaways:570 security fixes releasedAI is dramatically increasing vulnerability discoveryZero Trust and continuous security become even more importanthttps://krebsonsecurity.com/2026/07/microsoft-patches-a-record-570-security-flaws/06:40 – Why Recruiters Can't Find Workers and New Grads Can't Find JobsDespite a shortage of experienced professionals, many new graduates continue struggling to find employment. John & Lou explain how demographic shifts, retiring experts, and changing skill requirements are creating a mismatch between available talent and enterprise needs.Key takeaways:Experience gaps are wideningAI changes the skills employers valueOrganizations must rethink hiring and traininghttps://www.msn.com/en-us/news/us/why-recruiters-can-t-find-workers-and-new-grads-can-t-find-jobs-it-s-not-ai/ar-AA27K57y15:02 – Anthropic & Blackstone Bet on AI ImplementationAnthropic and Blackstone believe the next trillion-dollar AI opportunity isn't building better models—it's helping enterprises successfully deploy them. Their new venture focuses on integration, workflow automation, governance, security, and organizational change.John & Lou discuss why implementation may become the biggest challenge—and biggest opportunity—in enterprise AI.Key takeaways:AI deployment is becoming an execution challengeIntegration and governance are increasingly valuableServices may outgrow the model market itselfhttps://www.geekwire.com/2026/the-code-ai-forgot-logcat-ai-raises-2-55m-to-put-agents-to-work-on-device-operating-systems/📬 19:54 – Mail BagListener BJ shares his thoughts on Windows dropping below 60% desktop market share, noting that browsers, cloud services, and AI tools have become more important than the operating system itself. John & Lou discuss how web applications and AI continue reducing dependence on any single desktop platform.🔚 21:20 – Wrap UpWhether it's cybersecurity, hiring, or AI deployment, success increasingly depends on execution rather than technology alone. Organizations that combine skilled people, AI tools, and strong operational processes will have the greatest advantage.🌐 Social LinksIT SPARC Cast@ITSPARCCast on Xhttps://www.linkedin.com/company/sparc-sales/ on LinkedInJohn Barger@john_Video on Xhttps://www.linkedin.com/in/johnbarger/ on LinkedInLou Schmidt@loudoggeek on Xhttps://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn