Daily Tech Digest by Kannan Subbiah: metrics

Showing posts with label metrics. Show all posts

Daily Tech Digest - March 10, 2026

Quote for the day:

"A leader has the vision and conviction that a dream can be achieved. He inspires the power and energy to get it done." -- Ralph Nader

🎧 Listen to this digest on YouTube Music

▶ Play Audio Digest

Duration: 37 mins • Perfect for listening on the go.

Job disruption by AI remains limited — and traditional metrics may be missing the real impact

This article on computerworld explores the current state of artificial intelligence in the workforce. Despite widespread alarm, data from Challenger, Gray & Christmas indicates that AI accounted for roughly 8 to 10 percent of job cuts in early 2026. Researchers from Anthropic argue that traditional metrics fail to capture the nuances of AI integration, introducing an "observed exposure" methodology. This technique combines theoretical large language model capabilities with actual usage data, revealing that while certain roles—such as computer programmers and customer service representatives—have high exposure to automation, actual deployment lags significantly behind technical potential. Currently, AI functions primarily as a tool for task-based augmentation rather than full-scale replacement, which enhances worker productivity but complicates entry-level hiring. The report suggests that while immediate mass unemployment hasn't materialized, the long-term impact will require a fundamental re-engineering of workflows. This shift may disproportionately affect younger workers as companies struggle to balance AI efficiency with the necessity of maintaining a pipeline of human talent. Ultimately, the transition necessitates a strategic realignment of human roles to ensure sustainable growth in an intelligence-native era.

Why Password Audits Miss the Accounts Attackers Actually Want

This article on BleepingComputer highlights a critical disconnect between standard compliance-driven password audits and the actual tactics used by cybercriminals. While traditional audits prioritize technical requirements like complexity and rotation, they often overlook the context that makes an account vulnerable. For instance, a password can be statistically "strong" yet already compromised in a previous breach; research indicates that 83% of leaked passwords still meet regulatory standards. Furthermore, audits frequently neglect "orphaned" accounts belonging to former employees or contractors, which provide silent entry points for attackers. Service accounts—often over-privileged and exempt from expiry policies—represent another major blind spot. The piece argues that point-in-time snapshots are insufficient against continuous threats like credential stuffing. To be truly effective, security teams must shift toward continuous monitoring, incorporating breached-password screening and risk-based prioritization. By expanding the scope to include dormant, external, and service accounts, organizations can move beyond mere compliance to address the high-value targets that attackers prioritize. Ultimately, securing a digital environment requires recognizing that a compliant password is not necessarily a safe one in the face of modern, targeted exploitation.

AI is supercharging cloud cyberattacks - and third-party software is the most vulnerable

The latest Google Cloud Threat Report, as analyzed by ZDNET, highlights a significant escalation in cybersecurity risks where artificial intelligence is increasingly being used to "supercharge" cloud-based attacks. The report reveals a dramatic collapse in the window between the disclosure of a vulnerability and its mass exploitation, shrinking from weeks to mere days. Rather than targeting the highly secured core infrastructure of major cloud providers, threat actors are now focusing their efforts on unpatched third-party software and code libraries. This shift emphasizes that the modern supply chain remains a critical weak point for many organizations. Furthermore, the report notes a transition away from traditional brute force attacks toward more sophisticated identity-based compromises, including vishing, phishing, and the misuse of stolen human and non-human identities. Data exfiltration is also evolving, with "malicious insiders" increasingly using consumer-grade cloud storage services to move confidential information outside the corporate perimeter. To combat these AI-powered threats, Google’s experts recommend that businesses adopt automated, AI-augmented defenses, prioritize immediate patching of third-party tools, and strengthen identity management protocols. Ultimately, the report serves as a stark warning that in the current threat landscape, speed and automation are no longer optional but essential components of a robust cybersecurity strategy.

Change as Metrics: Measuring System Reliability Through Change Delivery Signals

This article highlights that system changes account for the vast majority of production incidents, necessitating their treatment as primary reliability indicators. To manage this risk, the author proposes a framework centered on three core business metrics: Change Lead Time, Change Success Rate, and Incident Leakage Rate. While aligned with DORA principles, this model specifically focuses on delivery quality by distinguishing between immediate deployment failures and latent defects that manifest as post-release incidents. To operationalize these goals, technical control metrics such as Change Approval Rate, Progressive Rollout Rate, and Change Monitoring Windows are introduced to provide actionable insights into pipeline friction and risk. The piece further advocates for a platform-agnostic, event-centric data architecture to collect these signals across diverse, distributed environments. This centralized approach avoids the brittleness of platform-specific logging and provides a unified view of system health. Ultimately, the framework empowers organizations to transform change management from a reactive necessity into a proactive, measurable engineering capability. By integrating these metrics, development teams can effectively balance the need for high-speed delivery with the imperative of system stability, ensuring that rapid innovation does not come at the expense of user experience or operational reliability.

The future of generative AI in software testing

In this article on Techzine, experts Hélder Ferreira and Bruno Mazzotta discuss the transformative shift of AI from a simple task accelerator to a fundamental structural layer within delivery pipelines. As global IT investment in AI is projected to surge toward $6.15 trillion by 2026, the software testing landscape is evolving beyond early challenges like hallucinations and "vibe coding" toward a sophisticated "quality intelligence layer." The authors outline four critical areas where AI adds strategic value: generating complex scenario-based datasets, suggesting high-risk exploratory prompts, automating defect triage to identify regression patterns, and enabling context-aware execution that prioritizes testing based on actual risk rather than volume. Crucially, the piece argues that while AI can significantly enhance velocity, sustainable success depends on maintaining "humans-in-the-loop" to ensure traceability and accountability. In this new era, the primary differentiator for enterprises will not be the sheer amount of AI deployed, but the effectiveness of their governance frameworks. By linking intent with execution and using AI as connective tissue across the lifecycle, organizations can achieve a balance where rapid delivery is supported by explainable automation and human-verified confidence in software quality.

CIOs cut IT corners to manufacture budget for AI

In this CIO.com article, author Esther Shein examines the aggressive strategies IT leaders are employing to fund artificial intelligence initiatives amidst stagnant overall budgets. Faced with intense pressure from boards and executive leadership to prioritize AI, many CIOs are being forced to make difficult trade-offs that jeopardize long-term stability. Common tactics include delaying non-critical infrastructure refreshes, such as server expansions and network improvements, which are often pushed out by twelve to eighteen months. Additionally, organizations are aggressively consolidating vendors, renegotiating contracts, and cutting legacy software subscriptions to free up capital. Some leaders have even implemented strict "self-funding" mandates where every new AI project must be offset by equivalent cuts elsewhere. Beyond technical sacrifices, the human element is also affected, with many departments reducing reliance on contractors or trimming internal staff to reallocate funds toward high-impact AI use cases. While these measures enable rapid deployment, they frequently lead to the accumulation of technical debt and a narrower scope for implementations. Ultimately, the piece warns that while these "corners" are being cut to fuel innovation, the resulting lack of focus on foundational maintenance could present significant operational risks in the future.

Beyond Prompt Injection: The Hidden AI Security Threats in Machine Learning Platforms

In the article "Beyond Prompt Injection: The Hidden AI Security Threats in Machine Learning Platforms," the focus of AI security shifts from headline-grabbing prompt injections to the critical vulnerabilities within MLOps infrastructure. While many security teams prioritize protecting chatbots from manipulation, the underlying platforms used to train and deploy models often present a far more dangerous attack surface. Through a red team engagement, researchers demonstrated how a simple self-registered trial account could be used to achieve remote code execution on a provider’s cloud infrastructure. By deploying a seemingly legitimate but malicious machine learning model, attackers can exploit the fact that these platforms must execute arbitrary code to function. The study highlights a significant risk: once RCE is achieved, weak network segmentation can allow adversaries to bypass trust boundaries and access sensitive internal databases or services. This effectively turns a managed ML environment into a gateway for lateral movement within a corporate network. To mitigate these threats, the article stresses that organizations must move beyond model-centric security and adopt robust infrastructure protections, including strict network isolation, continuous behavior monitoring, and a "zero-trust" approach to user-deployed artifacts, ensuring that the convenience of rapid AI development does not come at the cost of total system compromise.

Enterprise agentic AI requires a process layer most companies haven’t built

The VentureBeat article emphasizes that while 85% of enterprises aspire to implement agentic AI within the next three years, a staggering 76% acknowledge that their current operations are fundamentally unequipped for this transition. The core issue lies in the absence of a "process layer"—a critical foundation of optimized workflows and operational intelligence that provides AI agents with the necessary context to function effectively. Without this layer, agents are essentially "guessing," leading to a lack of reliability that causes 82% of decision-makers to fear a failure in return on investment. The piece argues that the primary hurdle is not merely technological but rather rooted in organizational structure and change management. Most companies suffer from siloed data and fragmented processes that hinder the seamless integration of autonomous systems. To overcome these barriers, businesses must prioritize process optimization and operational visibility, ensuring that AI-driven initiatives are linked to strategic executive outcomes. Simply layering advanced AI over inefficient, legacy frameworks will likely result in costly friction. Ultimately, for agentic AI to move beyond experimental pilots and deliver scalable value, organizations must first build a robust architectural bridge that connects sophisticated models with the complex, real-world logic of their daily business operations and high-stakes organizational decision cycles.

Building resilient foundations for India’s expanding Data Centre ecosystem

In "Building resilient foundations for India's expanding Data Centre ecosystem," Saurabh Verma explores the rapid evolution of India’s data infrastructure and the urgent necessity of prioritizing long-term resilience over mere capacity. As cloud adoption and 5G accelerate growth across hubs like Mumbai, Chennai, and Hyderabad, the sector faces escalating challenges that demand a sophisticated understanding of risk management. The article argues that modern data centres are no longer just IT assets but critical infrastructure whose failure directly impacts the digital economy. Beyond physical damage, business interruptions often result in massive financial losses, contractual penalties, and significant reputational harm. Climate change has emerged as a significant operational reality, with heatwaves and flooding stressing cooling systems and electrical grids. Furthermore, the convergence of cyber and physical risks means that digital disruptions can quickly translate into tangible infrastructure damage. Construction complexities and logistical interdependencies further amplify potential losses, making early risk engineering essential for success. Ultimately, the piece emphasizes that resilience must be a core design pillar rather than an afterthought. By integrating disciplined risk management from site selection through operations, Indian providers can gain a commercial advantage, securing better investment and insurance terms while building a sustainable, trustworthy backbone for the nation’s digital future.

CVE program funding secured, easing fears of repeat crisis

The Common Vulnerabilities and Exposures (CVE) program has successfully secured stable funding, alleviating industry-wide fears of a repeat of the 2025 crisis that nearly crippled global vulnerability tracking. As detailed in the CSO Online report, the Cybersecurity and Infrastructure Security Agency (CISA) and the MITRE Corporation have renegotiated their contract, transitioning the 26-year-old program from a discretionary expenditure to a protected line item within CISA's budget. This structural change effectively eliminates the "funding cliff" that previously required a last-minute emergency extension. While CISA leadership emphasizes that the program is now fully funded and evolving, some experts note that the specifics of the "mystery contract" remain opaque. The resolution comes at a critical time, as the cybersecurity community had already begun developing contingencies, such as the independent CVE Foundation, to reduce reliance on a single government source. Despite the financial stability, challenges regarding transparency, modernization, and international governance persist. The article underscores that while the immediate threat of a service lapse has faded, the incident served as a stark reminder of the global security ecosystem's fragility. Moving forward, the focus shifts toward ensuring this essential public resource remains resilient against future political or administrative shifts within the United States government.

Daily Tech Digest - March 03, 2026

Quote for the day:

“Appreciate the people who give you expensive things like time, loyalty and honesty.” -- Vala Afshar

Making sense of 6G: what will the ‘agentic telco’ look like?

6G will be the fundamental network for physical AI, promises Nvidia. Think of self-driving cars, robots in warehouses, or even AI-driven surgery. It’s all very futuristic; to actually deliver on these promises, a wide range of industry players will be needed, each developing the functionality of 6G. ... The ultimate goal for network operators is full automation, or “Level 5” automation. However, this seems too ambitious for now in the pre-6G era. Google refers to the twilight zone between Levels 4 and 5, with 4 assuming fully autonomous operation in certain circumstances. Currently, the obvious example of this type of automation is a partially self-driving car. As a user, you must always be ready to intervene, but ideally, the vehicle will travel without corrections. A Waymo car, which regularly drives around without a driver, is officially Level 4. ... Strikingly, most users hardly need this ongoing telco innovation. Only exceptionally extensive use of 4K streams, multiple simultaneous downloads, and/or location tracking can exceed the maximum bandwidth of most forms of 5G. Switch to 4G and in most use cases of mobile network traffic, you won’t notice the difference. You will notice a malfunction, regardless of the generation of network technology. However, the idea behind the latest 5G and future 6G networks is that these interruptions will decrease. Predictions for 6G assume a hundredfold increase in speed compared to 5G, with a similar improvement in bandwidth.

FinOps for agents: Loop limits, tool-call caps and the new unit economics of agentic SaaS

FinOps practitioners are increasingly treating AI as its own cost domain. The FinOps Foundation highlights token-based pricing, cost-per-token and cost-per-API-call tracking and anomaly detection as core practices for managing AI spend. Seat count still matters, yet I have watched two customers with the same licenses generate a 10X difference in inference and tool costs because one had standardized workflows and the other lived in exceptions. If you ship agents without a cost model, your cloud invoice quickly becomes the lesson plan ... In early pilots, teams obsess over token counts. However, for a scaled agentic SaaS running in production, we need one number that maps directly to value: Cost-per-Accepted-Outcome (CAPO). CAPO is the fully loaded cost to deliver one accepted outcome for a specific workflow. ... We calculate CAPO per workflow and per segment, then watch the distribution, not just the average. Median tells us where the product feels efficient. P95 and P99 tell us where loops, retries and tool storms are hiding. Note, failed runs belong in CAPO automatically since we treat the numerator as total fully loaded spend for that workflow (accepted + failed + abandoned + retried) and the denominator as accepted outcomes only, so every failure is “paid for” by the successes. Tagging each run with an outcome state and attributing its cost to a failure bucket allows us to track Failure Cost Share alongside CAPO and see whether the problem is acceptance rate, expensive failures or retry storms.

AI went from assistant to autonomous actor and security never caught up

The first is the agent challenge. AI systems have moved past assistants that respond to queries and into autonomous agents that execute multi-step tasks, call external tools, and make decisions without per-action human approval. This creates failure conditions that exist without any external attacker. An agent with overprivileged access and poor containment boundaries can cause damage through ordinary operation. ... The second category is the visibility challenge. Sixty-three percent of employees who used AI tools in 2025 pasted sensitive company data, including source code and customer records, into personal chatbot accounts. The average enterprise has an estimated 1,200 unofficial AI applications in use, with 86% of organizations reporting no visibility into their AI data flows. ... The third is the trust challenge. Prompt injection moved from academic research into recurring production incidents in 2025. OWASP’s 2025 LLM Top 10 list ranked prompt injection at the top. The vulnerability exists because LLMs cannot reliably separate instructions from data input. ... Wang recommended tiering agents by risk level. Agents with access to sensitive data or production systems warrant continuous adversarial testing and stronger review gates. Lower-risk agents can rely on standardized controls and periodic sampling. “The goal is to make continuous validation part of the engineering lifecycle,” she said.

A scorecard for cyber and risk culture

Cybersecurity and risk culture isn’t a vibe. It’s a set of actions, behaviors and attitudes you can point to without raising your voice. ... You can’t train people into that. You have to build an environment where that behavior makes sense, an environment based on trust and performance not one or the other ... Ownership is a design outcome. Treat it like product design. Remove friction. Clarify choices. Make it hard to do the wrong thing by accident and easy to make the best possible decision. ... If you can’t measure the behavior, you can’t claim the culture. You can claim a feeling. Feelings don’t survive audits, incidents or Board scrutiny. We’ve seen teams measure what’s easy and then call the numbers “maturity.” Training completion. Controls “done.” Zero incidents. Nice charts. Clean dashboards. Meanwhile, the real culture runs beneath the surface, making exceptions, working around friction and staying quiet when speaking up feels risky. ... One of the most dangerous culture metrics is silence dressed up as success. “Zero incidents reported” can mean you’re safe. It can also mean people don’t trust the system enough to speak up. The difference matters. The wrong interpretation is how organizations walk into breaches with a smile. Measure culture as you would safety in a factory. ... Metrics without governance create cynical employees. They see numbers. They never see action. Then they stop caring. Be careful not to make compliance ‘the culture’ as it’s what people do when no one is looking that counts.

Why encrypted backups may fail in an AI-driven ransomware era

For 20 years, I've talked up the benefits of the tech industry's best-practice 3-2-1 backup strategy. This strategy is just how it's done, and it works. Or does it? What if I told you that everything you know and everything you do to ensure quality backups is no longer viable? In fact, what if I told you that in an era of generative AI, when it comes to backups, we're all pretty much screwed? ... The easy-peasy assumption is that your data is good before it's backed up. Therefore, if something happens and you need to restore, the data you're bringing back from the backup is also good. Even without malware, AI, and bad actors, that's not always the way things turn out. Backups can get corrupted, and they might not have been written right in the first place, yada, yada, yada. But for this article, let's assume that your backup and restore process is solid, reliable, and functional. ... Even if the thieves are willing to return the data, their AI-generated vibe-coded software might be so crappy that they're unable to keep up their end of the bargain. Do you seriously think that threat actors who use vibe coding test their threat engines? ... Some truly nasty attacks specifically target immutable storage by seeking out misconfigurations. Here, they attack the management infrastructure, screwing with network data before it ever reaches the backup system. The net result is that before encryption of off-site backups begins, and before the backups even take place, the malware has suitably corrupted and infected the data.

How Deepfakes and Injection Attacks Are Breaking Identity Verification

Unlike social media deception, these attacks can enable persistent access inside trusted environments. The downstream impact is durable: account persistence, privilege-escalation pathways, and lateral movement opportunities that start with a single false verification decision. ... One practical problem for deepfake defense is generalization: detectors that test well in controlled settings often degrade in “in-the-wild” conditions. Researchers at Purdue University evaluated deepfake detection systems using their real-world benchmark based on the Political Deepfakes Incident Database (PDID). PDID contains real incident media distributed on platforms such as X, YouTube, TikTok, and Instagram, meaning the inputs are compressed, re-encoded, and post-processed in the same ways defenders often see in production. ... It’s important to be precise: PDID measures robustness of media detection on real incident content. It does not model injection, device compromise, or full-session attacks. In real identity workflows, attackers do not choose one technique at a time; they stack them. A high-quality deepfake can be replayed. A replay can be injected. An injected stream can be automated at scale. The best media detectors still can be bypassed if the capture path is untrusted. That’s why Deepsight goes even deeper than asking “Is this video a deepfake?”

Virtual twins and AI companions target enterprise war rooms

Organisations invest millions digitising processes and implementing enterprise systems. Yet when business leaders ask questions spanning multiple domains, those systems don’t communicate effectively. Teams assemble to manually cross-reference data, spending days producing approximations rather than definitive answers. Manufacturing experts at the conference framed this as decades of incomplete digitisation. ... Addressing this requires fundamentally changing how enterprise data is structured and accessed. Rather than systems operating independently with occasional data exchanges, the approach involves projecting information from multiple sources onto unified representations that preserve relationships and context. Zimmerman used a map analogy to explain the concept. “If you take an Excel spreadsheet with location of restaurants and another Excel spreadsheet with location of flower shops, and you try to find a restaurant nearby a flower shop, that’s difficult,” he said. “If it’s on the map, it is simple because the data are correlated by nature.” ... Having unified data representations solves part of the problem. Accessing them requires interfaces that don’t force users to understand complex data structures or navigate multiple applications. The conversational AI approach – increasingly common across enterprise software – aims to let users ask questions naturally rather than construct database queries or click through application menus.

The rise of the outcome-orchestrating CIO

Delivering technology isn’t enough. Boards and business leaders want results — revenue, measurable efficiency, competitive advantage — and they’re increasingly impatient with IT organizations that can’t connect their work to those outcomes. ... Funding models change, too. Traditional IT budgets fund teams to deliver features. When the business pivots, that becomes a change request — creating friction even when it’s not an adversarial situation. “Instead, fund a value stream,” Sample says. “Then, whatever the business needs, you absorb the change and work toward shared goals. It doesn’t matter what’s on the bill because you’re all working toward the same outcome.” It’s a fundamental reframing of IT’s role. “Stop talking about shared services,” says Ijam of the Federal Reserve. “Talk about being a co-owner of value realization.” That means evolving from service provider to strategic partner — not waiting for requirements but actively shaping how technology creates business results. ... When outcome orchestration is working, the boardroom conversation changes. “CIOs are presenting business results enabled by technology — not just technology updates — and discussing where to invest next for maximum impact,” says Cox Automotive’s Johnson. “The CFO begins to see technology as an investment that generates returns, not just a cost to be managed.” ... When outcome orchestration takes hold, the impact shows up across multiple dimensions — not just in business metrics, but in how IT is perceived and how its people experience their work.

The future of banking: When AI becomes the interface

Experiences must now adapt to people—not the other way around. As generative capabilities mature, customers will increasingly expect banking interactions to be intuitive, conversational, and personalized by default, setting a much higher bar for digital experience design. ... Leadership teams must now ask harder questions. What proprietary data, intelligence, or trust signals can only our bank provide? How do we shape AI-driven payment decisions rather than merely fulfill them? And how do we ensure that when an AI decides how money moves, our institution is not just compliant, but preferred? ... AI disruption presents both significant risk and transformative opportunity for banks. To remain relevant, institutions must decide where AI should directly handle customer interactions, how seamlessly their services integrate into AI-driven ecosystems, and how their products and content are surfaced and selected by AI-led discovery and search. This requires reimagining the bank’s digital assistant across seven critical dimensions: being front and centre at the point of intent, contextual in understanding customer needs, multi-modal across voice, text, and interfaces, agentic in taking action on the customer’s behalf, revenue-generating through intelligent recommendations, open and connected to broader ecosystems, and capable of providing targeted, proactive support.

The End of the ‘Observability Tax’: Why Enterprises are Pivoting to OpenTelemetry

For enterprises to reclaim their budget, they must first address inefficiency—the “hidden tax” of observability facing many DevOps teams. Every organization is essentially rebuilding the same pipeline from scratch, and when configurations aren’t standardized, engineers aren’t learning from each other; they’re actually repeating the same trial-and-error processes thousands of times over. This duplicated effort leads to a waste of time and resources. It often takes weeks to manually configure collectors, processors, and exporters, plus countless hours of debugging connection issues. ... If data engineers are stuck in a cycle of trial-and-error to manage their massive telemetry, then organizations are stuck drinking from a firehose instead of proactively managing their data in a targeted manner. In a world where AI demands immediate access to enormous volumes of data, this lack of flexibility becomes a fatal competitive disadvantage. If enterprises want to succeed in an AI-driven world, their data infrastructure must be able to handle the rapid velocity of data in motion without sacrificing cost-efficiency. Identifying and mitigating these hidden challenges and costs is imperative if enterprises want to turn their data into an asset rather than a liability. ... When organizations reclaim complete control of their data pipelines, they can gain a competitive edge.

Daily Tech Digest - February 26, 2026

Quote for the day:

"It is not such a fierce something to lead once you see your leadership as part of God's overall plan for his world." -- Calvin Miller

Boards don’t need cyber metrics — they need risk signals

Decision-makers want to know whether risk is increasing or decreasing, whether controls are effective, and whether the organization can limit damage when prevention fails. Metrics are therefore useful when they clarify those questions. “Time is really the universal metric because everyone can understand time,” Richard Bejtlich, strategist and author in residence at Corelight, tells CSO. “How fast do we detect problems, and how fast do we contain them. Dwell time, containment time. That’s the whole game for me.” Organizations cannot prevent every intrusion, Bejtlich argues, but they can measure how quickly they recognize and contain one. ... Wendy Nather, a longtime CISO who is now an advisor at EPSD, cautions against equating measurement with understanding. “When you are reporting to the board, there are some things you just cannot count that you have to report anyway,” she tells CSO. She points to incidents, near misses, and changes in assumptions as examples. “Anything that changes your assumptions about how you’re managing your security program, you should be bringing those to the board, even if you can’t count them,” Nather says. Regular metrics can create a rhythm of predictability, and that predictability could lull board members into a false sense of security. “Metrics are very seductive,” she says. “They lead us toward things that can be counted, that happen on a regular basis.” The result may be a steady flow of data that obscures structural risk or emerging weaknesses, Nather warns.

The Enterprise AI Postmortem Playbook: Diagnosing Failures at the Data Layer

Your first rule of the playbook is to treat AI incidents as data incidents – until proven otherwise. You should start by tagging the failure type. Document whether it’s a structure issue, retrieval misalignment, conflict with metric definition, or other categories. Ideally, you want to assign the issue to an owner and attach evidence to force some discipline into the review. Try to classify the issue into clearly defined buckets. For example, you can classify into these four buckets: structural failure, retrieval misalignment, definition conflict, or freshness failure. Once this part is clear, the investigation becomes more focused. The goal with this step is to isolate the data fault line. ... The next step is to move one layer deeper. Identify the source table behind the retrieved context. You also want to confirm the timestamp of the last refresh. Check whether any ingestion jobs failed, partially completed, or ran late. Silent failures are common. A job may succeed technically while loading incomplete data. As you go through the playbook continue tracing upstream. Find the transformation job that shaped the dataset. Look at recent schema changes. Check whether any business rules were updated. The idea here is to rebuild the exact path that led to the output. Try to not make any assumptions at this stage about model behavior – simply keep tracing until the process is complete. Don’t be surprised if the model simply worked with what it was given.

Top Attacks On Biometric Systems (And How To Defend Against Them)

Presentation attacks, often referred to as spoofing attacks, occur when an attacker presents a fake biometric sample to a sensor (like a camera or microphone) in an attempt to impersonate a legitimate user. Common examples include printed photos, video replays, silicone masks, prosthetics or synthetic fingerprints. More recently, high-quality deepfake videos have become a powerful new tool in the attacker’s arsenal. ... Passive liveness techniques, which analyze subtle physiological and behavioral signals without requiring user interaction, are particularly effective because they reduce friction while improving security. However, liveness detection must be resilient to unknown attack methods, not just tuned to detect known spoof types. ... Not all biometric attacks happen in front of the sensor. Replay and injection attacks target the biometric data pipeline itself. In these scenarios, attackers intercept, replay or inject biometric data, such as images or templates, directly into the system, bypassing the sensor entirely. ... Defensive strategies must extend beyond the biometric algorithm. Secure transmission, encryption in transit, device attestation, trusted execution environments and validation that data originates from an authorized sensor are all essential. ... Although less visible to end users, attacks targeting biometric templates and databases can pose long-term risks. If biometric templates are compromised, the impact extends far beyond a single breach.

Open-source security debt grows across commercial software

High and critical risk findings remain widespread. Most codebases contain at least one high risk vulnerability, and nearly half contain at least one critical risk issue. Those rates dipped slightly from the prior year even as total vulnerability counts rose. Supply chain attacks add another layer of risk. Sixty five percent of surveyed organizations experienced a software supply chain attack in the past year. ... “As AI reshapes software development, security teams will have to continue to adapt in turn. Security budgets and security guidelines should reflect this new reality. Leaders should continue to invest in tooling and education required to equip teams to manage the drastic increase in velocity, volume, and complexity of applications,” Mackey said. Board level reporting also requires adjustment as vulnerability volumes rise. ... Outdated components appear in nearly every audited environment. More than nine in ten codebases contain components that are several years out of date or show no recent development activity. A large share of components run many versions behind current releases. Only a small fraction operate on the latest available version. This maintenance debt intersects with regulatory obligations. The EU Cyber Resilience Act entered into effect in late 2024, with key reporting requirements taking effect in 2026 and broader enforcement following in 2027.

The agentic enterprise: Why value streams and capability maps are your new governance control plane

The enterprise is currently undergoing a seismic pivot from generative AI, which focuses on content creation, to agentic AI, which focuses on goal execution. Unlike their predecessors, these agents possess “structured autonomy”: the ability to perceive contexts, plan actions and execute across systems without constant human intervention. For the CIO and the enterprise architect, this is not merely an upgrade in automation speed; it is a fundamental shift in the firm’s economic equation. We are moving from labor-centric workflows to digital labor capable of disassembling and reassembling entire value chains. ... In an agentic enterprise, the value stream map is no longer just a diagram; it is the control plane. It must explicitly define the handoff protocols between human and digital agents. In my opinion, Value stream maps must move from static documents stored in a repository to context documents used to drive agentic automation. ... If a value stream does not exist, you cannot automate it. For new agentic workflows, do not map the current human process. Instead, use an outcome-backwards approach. Work backward from the concrete deliverable (e.g., customer onboarded) to identify the minimum viable API calls required. Before granting write access, run the new agentic stream in shadow mode to validate agent decisions against human outcomes.

Beyond compliance: Building a culture of data security in the digital enterprise

Cyber compliance is something organisations across industrial sectors take seriously, especially with new regulations getting introduced and non-compliance having consequences such as hefty penalties. Hence, businesses are placing compliance among their top priorities. However, hyper-focusing only on compliance can lead to tunnel vision, crippling creativity, and innovation. It fails to offer a comprehensive risk assessment due to the checklist approach it follows, exposing organizations to vulnerabilities and fast-evolving threats. Having a compliance-first mindset can lead to incomplete risk assessment, creating blind spots and security gaps in security provisions. ... With businesses relying on data for operations, customer engagement, and decision-making, ensuring data security protects both users and organisations. Data breaches have severe consequences, including financial losses, reputational damage, customer churn, and regulatory penalties. With data moving across on-premises data centers, cloud platforms, third-party ecosystems, remote work environments, and AI-driven applications, there is a need for a holistic, culture-driven approach to cybersecurity. ... Data protection traditionally was focused on safeguarding the perimeter by securing networks and systems within the physical boundaries where data was normally stored.

If you thought RTO battles were bad, wait until AI mandates start taking hold across the industry

With the advent of generative AI and the incessant beating of the drum by executives hellbent on unlocking productivity gains, we could see a revival of the dreaded workforce mandate –- only this time with AI. We’ve already had a glimpse of the same RTO tactics being used with AI over the last year. In mid-2025, Microsoft introduced new rules aimed at boosting AI use across the company, with an internal memo warning staff that “using AI is no longer optional”. ... As with RTO mandates, we’re now reaching a point where upward mobility within the enterprise could be at risk as a result of AI use. It’s a tactic initially touted by Dell in 2024 when enforcing its own hybrid work rules, which prompted a fierce backlash among staff. Forcing workers to use AI or risk losing out on promotions will have the desired effect executives want, namely that employees will use the technology, but that’s missing the point entirely. AI has been framed by many big tech providers as a prime opportunity to supercharge productivity and streamline enterprise efficiency. We’ve all heard the marketing jargon. If business leaders are at the point where they’re forcing staff to use the technology, it begs the question of whether it’s actually having the desired effect, which recent analysis suggests it’s not. ... Recent analysis from CompTIA found roughly one-third of companies now require staff to complete AI training.

In perfect harmony: How Emerald AI is turning data centers into flexible grid assets

At the core of Emerald AI is its Emerald Conductor platform. Described by Sivaram as “an AI for AI,” the system orchestrates thousands of AI workloads across one or more data centers, dynamically adjusting operations to respond to grid conditions while ensuring the facility maintains performance. The system achieves this through a closed-loop orchestration platform comprising an autonomous agent and a digital twin simulator. ... A point keenly pointed out by Steve Smith, chief strategy and regulation officer at National Grid, at the time of the announcement: “As the UK’s digital economy grows, unlocking new ways to flexibly manage energy use is essential for connecting more data centers to our network efficiently.” The second reason was National Grid's transatlantic stature - as an American company active in both the UK and US markets - and its commitment to the technology. “They’ve invested in the program and agreed to a demo, which makes them the ideal partner for our first international launch,” says Sivaram. The final, and most important, factor, notes Sivaram, was the access to the NextGrid Alliance, a consortium of 150 utilities worldwide. By gaining access to such a robust partner network, the deal could serve as a springboard for further international projects. This aligns with the company’s broader partnership approach. Emerald AI has already leveraged Nvidia’s cloud partner network to test its technology across US data centers, laying the groundwork for broader deployment and continued global collaboration.

7 ways to tame multicloud chaos with generative AI

Architects have the difficult job of understanding tradeoffs between proprietary cloud services and cross-cloud platforms. For example, should developers use AWS Glue, Azure Data Factory, or Google Cloud Data Fusion to develop data pipelines on the respective platforms, or should they adopt a data integration platform that works across clouds? ... “Managing multicloud is like learning multiple languages from AWS, Azure, Oracle, and others, and it’s rare to have teams that can traverse these environments fluidly and effectively. Plus, services and concepts are not portable among clouds, especially in cloud-native PaaS services that go beyond IaaS,” says Harshit Omar, co-founder and CTO at FluidCloud. One way to work around this issue is to assign an AI agent to support the developer or architect in evaluating platform selections. ... Standardizing infrastructure and service configurations across different clouds requires expertise in different naming conventions, architecture, tools, APIs, and other paradigms. Look for genAI tools to act as a translator to streamline configurations, especially for organizations that can templatize their requirements. ... CI/CD, infrastructure-as-code, and process automation are key tools for driving efficiency, especially when tasks span multiple cloud environments. Many of these tools use basic flows and rules to streamline tasks or orchestrate operations, which can create boundary cases that cause process-blocking errors.

It’s Time To Reinforce Institutional Crypto Key Management With MPC: Sodot CEO

For years, crypto security operations were almost exclusively focused on finding a way to protect the private keys to crypto wallets. It’s known as the “custody risk,” and it will always be a concern to anyone holding digital assets. However, Sofer believes that custody is no longer the weakest link. Cyberattackers have come to realize that secure wallets, often held in cold storage, are far too difficult to crack. ... Sodot has built a self-hosted infrastructure platform that leverages a pair of cutting-edge security techniques – namely, Multi-Party Computation or MPC and Trusted Execution Environments or TEEs. With Sodot’s platform, API keys are never reassembled in full plaintext, eliminating one of the main weaknesses of traditional secrets managers, which typically expose the entire key to any authenticated machine. Instead, Sodot uses MPC to split each key into multiple “shares” that are held by different partners on different technology stacks, Sofer explained. Distributing risk in this way makes an attacker’s job exponentially more difficult, as it means they would have to compromise multiple isolated systems to gain access. ... “Keys are here to stay, and they will control more value and become more sensitive as technology progresses,” Sofer concluded. “As financial institutions get more involved in crypto, we believe demand for self-hosted solutions that secure them will only grow, driven by performance requirements, operational resilience, and control over security boundaries.”

Daily Tech Digest - February 24, 2026

Quote for the day:

"Transparent reviews create fairness. Subjective reviews create frustration." -- Gordon Tredgold

AI agents and bad productivity metrics

The great promise of was that it would finally clear our backlogs. Coding agents would churn out boilerplate at superhuman speeds, and teams would finally ship exactly what the business wants. The reality, as we settle into 2026, is far more uncomfortable. Artificial intelligence is not going to save developer productivity because writing code was never the bottleneck in software engineering. ... For decades, one of the most common debugging techniques was entirely social. A production alert goes off. You look at the version control history, find the person who wrote the code, ask them what they were trying to accomplish, and reconstruct the architectural intent. But what happens to that workflow when no one actually wrote the code? What happens when a human merely skimmed a 3,000-line agent-generated pull request, hit merge, and moved on to the next ticket? When an incident happens, where is the deep knowledge that used to live inside the author? ... The metrics that matter are still the boring ones because they measure actual business outcomes. The remain the best sanity check we have because they tie delivery speed directly to system stability. They measure deployment frequency, lead time for changes, change failure rate, and time to restore service. None of those metrics cares about the number of commits your agents produced today. They only care about whether your system can absorb change without breaking.

How vertical SaaS is redefining enterprise efficiency

For the past decade, horizontal SaaS has been the defining force in enterprise technology. Platforms like CRMs, ERP suites and collaboration tools promised universality, offering a single platform to manage every business function across all industries. The strategy made sense: a large total addressable market, reusable architecture and marketing scale. flips that model. It is narrow by design but deep in impact. A report by Strategy& found that B2B vertical software companies are now growing faster than their horizontal peers, thanks to higher retention rates, lower churn rates and better unit economics. When software mirrors how a business already works, people stop treating it like a tool they tolerate and start relying on it like infrastructure. ... In regulated industries, isn’t a feature; it’s the baseline for trust. I learned early that trying to retrofit audit trails or data retention policies after go-live only creates technical debt. Instead, design for compliance as a first-class product layer: immutable logs, permission hierarchies and exportable compliance reports built into the system. ... Vertical products don’t thrive in isolation. Integration with industry hardware, marketplaces and regulatory systems drives adoption. In one case, we partnered with a hardware vendor to automatically sync manifest data from their devices, cutting onboarding time in half and unlocking co-marketing opportunities.

API Security Standards: 10 Essentials to Get You Started

Most API security flaws are created during the design phase. You're too late if you're waiting until deployment to think about threats. Shift-left principles mean integrating security early, especially at the design phase, where flawed assumptions become future exploits. Start by mapping out each endpoint's purpose, what data it touches, and who should access it. Identify where trust is assumed (not earned), roles blur, and inputs aren't validated. ... Every API has a breaking point. If you don't define it, attackers will. prevent denial-of-service (DoS) attacks, and they're also your first defense against scraping, brute-forcing, enumeration, and even accidental misuse by poorly built integrations. APIs, by nature, invite automation. Without guardrails, that openness turns into a floodgate. And in some cases, unchecked abuse opens the door to far worse issues, like remote code execution, where improperly scoped input or lack of throttling leads directly to exploitation. ... APIs are built to accept input. Attackers find ways to exploit it. The core rule is this - if you didn't expect it, don't process it. If you didn't define it, don't send it. Define request and response schemas explicitly using tools like or JSON Schema, as recommended by leading API security standards. Then enforce them — at the gateway, app layer, or both. Don't just use validation as linting; treat it as a runtime contract. If the payload doesn't match the spec, reject it.

Why AI Urgency Is Forcing a Data Governance Reset

The cost of weak governance shows up in familiar ways: teams can’t find data, requirements arrive late in the process, and launches stall when compliance realities collide with product timelines. Without governance, argues, organizations “ultimately suffer from higher cost basis,” with downstream consequences that “impact the bottom line.” ... McQuillan sees a clear step-change in executive urgency since generative AI (GenAI) became mainstream. “There’s been a rapid adoption, particularly since the advent of GenAI and the type of generative and agentic technologies that a lot of C-suites are taking on,” he says. But he also describes a common leadership gap: many executives feel pressure to become “AI-enabled” without a clear definition of what that means or how to build it sustainably. “There’s very much a well-understood need across all companies to become AI-enabled in some way,” he says. “But the problem is a lot of folks don’t necessarily know how to define that.” In the absence of clarity, organizations often fall into scattershot experimentation. What concerns McQuillan the most is how the pace of the “race” shapes priorities. ... When asked whether the long-running mantra “data is the new oil” still holds in the era of large language models and agentic workflows, McQuillan is direct. “It holds true now more than ever,” he says. He acknowledges why attention drifts: “It’s natural for people to gravitate toward things that are shiny,” and “AI in and of itself is an absolutely magnificent space.”

Building a Least-Privilege AI Agent Gateway for Infrastructure Automation with MCP, OPA, and Ephemeral Runners

An agent misinterpreting an instruction can initiate destructive infrastructure changes, such as tearing down environments or modifying production resources. A compromised agent identity can be abused to exfiltrate secrets, create unauthorized workloads, or consume resources at scale. In practice, teams often discover these issues late, because traditional logs record what happened, but not why an agent decided to act in the first place. For organizations, this liability creates operational and governance challenges. Incidents become harder to investigate, change approvals are bypassed unintentionally, and security teams are left with incomplete audit trails. Over time, this problem erodes trust in automation itself, forcing teams to either roll back agent usage or accept increasing levels of unmanaged risk. ... A more sustainable approach is to introduce an explicit control layer between agents and the systems they operate on. In this article, we focus on an AI Agent Gateway, a dedicated boundary that validates intent, enforces policy as code, and isolates execution before any infrastructure or service API is invoked. Rather than treating agents as privileged actors, this model treats them as untrusted requesters whose actions must be authorized, constrained, observed, and contained. ... In the context of AI-driven automation, defense in depth means that no single component, neither the agent, nor the gateway, nor the execution environment, has enough authority on its own to cause damage.

Demystifying CERT‑In’s Elemental Cyber Defense Controls: A Guide for MSMEs

For India’s Micro, Small, and Medium Enterprises (MSMEs), cybersecurity is no longer a “big company problem.” With digital payments, SaaS adoption, cloud-first operations, and supply‑chain integrations becoming the norm, MSMEs are now prime targets for cyberattacks. To help these organizations build a strong foundational security posture, the Indian Computer Emergency Response Team (CERT-In) has released CIGU-2025-0003, outlining a baseline of Cyber Defense Controls, which prescribes 15 Elemental Cyber Security Controls—a pragmatic, baseline set of safeguards designed to uplift the nation’s cyber hygiene. ... These controls, mapped to 45 recommendations, enable essential digital hygiene, protect against ransomware, ensure regulatory compliance, and are required for annual audits. CERT‑In’s Elemental Controls are designed as minimum essential practices that every Indian organization—regardless of size—should implement. ... The CERT-In guidelines offer a simplified, actionable starting point for MSMEs to benchmark their security. These controls are intentionally prescriptive, unlike ISO or NIST, which are more framework‑oriented. ... Because threats constantly evolve and MSMEs face unique risks depending on their industry and data sensitivity, organizations should view this framework not as an endpoint, but as the first critical step toward building a comprehensive security program akin to ISO 27001 or NIST CSF 2.0.

AI-fuelled cyber attacks hit in minutes, warns CrowdStrike

CrowdStrike reports a sharp acceleration in cyber intrusions, with attackers moving from initial access to lateral movement in less than half an hour on average as widely available artificial intelligence tools become embedded in criminal workflows. Its latest Global Threat Report puts average eCrime "breakout time" at 29 minutes in 2025, a 65% improvement on the prior year. ... Alongside generative AI use in preparation and execution, the report describes attempts to exploit AI systems directly. Adversaries injected malicious prompts into GenAI tools at more than 90 organisations, using them to generate commands associated with credential theft and cryptocurrency theft. ... Incidents linked to North Korea rose more than 130%, while activity by the group CrowdStrike tracks as FAMOUS CHOLLIMA more than doubled. The report says DPRK-nexus actors used AI-generated personas to scale insider operations. It also cites a large cryptocurrency theft attributed to the actor it calls PRESSURE CHOLLIMA, valued at USD $1.46 billion and described as the largest single financial heist ever reported. The report also references AI-linked tooling used by other state and criminal groups. Russia-nexus FANCY BEAR deployed LLM-enabled malware, which it named LAMEHUG, for automated reconnaissance and document collection. The eCrime actor tracked as PUNK SPIDER used AI-generated scripts to speed up credential dumping and erase forensic evidence.

Shadow mode, drift alerts and audit logs: Inside the modern audit loop

When systems moved at the speed of people, it made sense to do compliance checks every so often. But AI doesn't wait for the next review meeting. The change to an inline audit loop means audits will no longer occur just once in a while; they happen all the time. Compliance and risk management should be "baked in" to the AI lifecycle from development to production, rather than just post-deployment. This means establishing live metrics and guardrails that monitor AI behavior as it occurs and raise red flags as soon as something seems off. ... Cultural shift is equally important: Compliance teams must act less like after-the-fact auditors and more like AI co-pilots. In practice, this might mean compliance and AI engineers working together to define policy guardrails and continuously monitor key indicators. With the right tools and mindset, real-time AI governance can “nudge” and intervene early, helping teams course-correct without slowing down innovation. In fact, when done well, continuous governance builds trust rather than friction, providing shared visibility into AI operations for both builders and regulators, instead of unpleasant surprises after deployment. ... Shadow mode is a way to check compliance in real time: It ensures that the model handles inputs correctly and meets policy standards before it is fully released. One AI security framework showed how this method worked: Teams first ran AI in shadow mode, then compared AI and human inputs to determine trust.

Making AI Compliance Practical: A Guide for Data Teams Navigating Risk, Regulation, and Reality

As AI tools become more embedded in enterprise workflows, data teams are encountering a growing reality: compliance isn’t only a legal concern but also a design constraint, a quality signal, and, often, a competitive differentiator. But navigating compliance can feel complex, especially for teams focused on building and shipping. What is the good news? It doesn’t have to be. When approached intentionally, compliance becomes a pathway to better decisions, not a barrier. ... Automation can help with regulations, but only if it's used correctly. I've looked at a tool before that used algorithms to find private information. It worked well with English, but when tested with material in more than one language, it missed a few personal identifiers. The group thought it was "smart enough." It wasn't. We kept the automation, but we added human review for rare cases, confidence levels to make checks happen, and alerts for input formats that aren't common. The automation stayed the same, but there were built-in checks and balances. ... The biggest compliance failures don’t come from bad people. They come from good teams moving fast, skipping hard questions, and assuming nothing will go wrong. But compliance isn’t a blocker. It’s a product quality signal. People will trust you more if they are aware that your team has carefully considered the details.

Tata Communications’ Andrew Winney on why SASE is now non-negotiable

is often discussed as a product decision, but in reality it is a journey. Many enterprises start with a few use cases, such as securing internet access or enabling remote access to private applications. But they do not always extend those principles across contractors, third-party users, software-as-a-service applications and hybrid environments. Practical Zero Trust requires enterprises to rethink access fundamentally. Every request must be evaluated based on who the user is, the context from which they are accessing, the device they are using and the resource they are requesting. Access must then be granted only to that specific resource. ... represents a structural convergence of networking and security rather than a simple technology swap. What are the most critical architectural and change-management considerations enterprises must address during this transition? SASE is not a one-time technology change. It represents the convergence of networking and security under unified orchestration and policy management. That transition takes time and must be managed carefully. We typically work with enterprises through phased transition plans. If an organisation’s immediate priority is securing internet access or private application access for remote users, we begin there and expand to additional use cases over time. Integration is critical. Enterprises have existing investments in cloud platforms, local area networks and security tools.

Daily Tech Digest - February 07, 2026

Quote for the day:

"Success in almost any field depends more on energy and drive than it does on intelligence. This explains why we have so many stupid leaders." -- Sloan Wilson

Tiny AI: The new oxymoron in town? Not really!

Could SLMs and minituarised models be the drink that would make today’s AI small enough to walk through these future doors without AI bumping into carbon-footprint issues? Would model compression tools like pruning, quantisation, and knowledge distillation help to lift some weight off the shoulders of heavy AI backyards? Lightweight models, edge devices that save compute resources, smaller algorithms that do not put huge stress on AI infrastructures, and AI that is thin on computational complexity- Tiny AI- as an AI creation and adoption approach- sounds unusual and promising at the onset. ... hardware innovations and new approaches to modelling that enable Tiny AI can significantly ease the compute and environmental burdens of large-scale AI infrastructures, avers Biswajeet Mahapatra, principal analyst at Forrester. “Specialised hardware like AI accelerators, neuromorphic chips, and edge-optimised processors reduces energy consumption by performing inference locally rather than relying on massive cloud-based models. At the same time, techniques such as model pruning, quantisation, knowledge distillation, and efficient architectures like transformers-lite allow smaller models to deliver high accuracy with far fewer parameters.” ... Tiny AI models run directly on edge devices, enabling fast, local decision-making by operating on narrowly optimised datasets and sending only relevant, aggregated insights upstream, Acharya spells out.

Kali Linux vs. Parrot OS: Which security-forward distro is right for you?

The first thing you should know is that Kali Linux is based on Debian, which means it has access to the standard Debian repositories, which include a wealth of installable applications. ... There are also the 600+ preinstalled applications, most of which are geared toward information gathering, vulnerability analysis, wireless attacks, web application testing, and more. Many of those applications include industry-specific modifications, such as those for computer forensics, reverse engineering, and vulnerability detection. And then there are the two modes: Forensics Mode for investigation and "Kali Undercover," which blends the OS with Windows. ... Parrot OS (aka Parrot Security or just Parrot) is another popular pentesting Linux distribution that operates in a similar fashion. Parrot OS is also based on Debian and is designed for security experts, developers, and users who prioritize privacy. It's that last bit you should pay attention to. Yes, Parrot OS includes a similar collection of tools as does Kali Linux, but it also offers apps to protect your online privacy. To that end, Parrot is available in two editions: Security and Home. ... What I like about Parrot OS is that you have options. If you want to run tests on your network and/or systems, you can do that. If you want to learn more about cybersecurity, you can do that. If you want to use a general-purpose operating system that has added privacy features, you can do that.

Bridging the AI Readiness Gap: Practical Steps to Move from Exploration to Production

To bridge the gap between AI readiness and implementation, organizations can adopt the following practical framework, which draws from both enterprise experience and my ongoing doctoral research. The framework centers on four critical pillars: leadership alignment, data maturity, innovation culture, and change management. When addressed together, these pillars provide a strong foundation for sustainable and scalable AI adoption. ... This begins with a comprehensive, cross-functional assessment across the four pillars of readiness: leadership alignment, data maturity, innovation culture, and change management. The goal of this assessment is to identify internal gaps that may hinder scale and long-term impact. From there, companies should prioritize a small set of use cases that align with clearly defined business objectives and deliver measurable value. These early efforts should serve as structured pilots to test viability, refine processes, and build stakeholder confidence before scaling. Once priorities are established, organizations must develop an implementation road map that achieves the right balance of people, processes, and technology. This road map should define ownership, timelines, and integration strategies that embed AI into business workflows rather than treating it as a separate initiative. Technology alone will not deliver results; success depends on aligning AI with decision-making processes and ensuring that employees understand its value.

Proxmox's best feature isn't virtualization; it's the backup system

Because backups are integrated into Proxmox instead of being bolted on as some third-party add-on, setting up and using backups is entirely seamless. Agents don't need to be configured per instance. No extra management is required, and no scripts need to be created to handle the running of snapshots and recovery. The best part about this approach is that it ensures everything will continue working with each OS update. Backups can be spotted per instance, too, so it's easy to check how far you can go back and how many copies are available. The entire backup strategy within Proxmox is snapshot-based, leveraging localised storage when available. This allows Proxmox to create snapshots of not only running Linux containers, but also complex virtual machines. They're reliable, fast, and don't cause unnecessary downtime. But while they're powerful additions to a hypervised configuration, the backups aren't difficult to use. This is key since it would render the backups less functional if it proved troublesome to use them when it mattered most. These backups don't have to use local storage either. NFS, CIFS, and iSCSI can all be targeted as backup locations. ... It can also be a mixture of local storage and cloud services, something we recommend and push for with a 3-2-1 backup strategy. But there's one thing of using Proxmox's snapshots and built-in tools and a whole different ball game with Proxmox Backup Server. With PBS, we've got duplication, incremental backups, compression, encryption, and verification.

The Fintech Infrastructure Enabling AI-Powered Financial Services

AI is reshaping financial services faster than most realize. Machine learning models power credit decisions. Natural language processing handles customer service. Computer vision processes documents. But there’s a critical infrastructure layer that determines whether AI-powered financial platforms actually work for end users: payment infrastructure. The disconnect is striking. Fintech companies invest millions in AI capabilities, recommendation engines, fraud detection, personalization algorithms. ... From a technical standpoint, the integration happens via API. The platform exposes user balances and transaction authorization through standard REST endpoints. The card provider handles everything downstream: card issuance logistics, real-time currency conversion, payment network settlement, fraud detection at the transaction level, dispute resolution workflows. This architectural pattern enables fintech platforms to add payment functionality in 8-12 weeks rather than the 18-24 months required to build from scratch. ... The compliance layer operates transparently to end users while protecting platforms from liability. KYC verification happens at multiple checkpoints. AML monitoring runs continuously across transaction patterns. Reporting systems generate required documentation automatically. The platform gets payment functionality without becoming responsible for navigating payment regulations across dozens of jurisdictions.

Context Engineering for Coding Agents

Context engineering is relevant for all types of agents and LLM usage of course. My colleague Bharani Subramaniam’s simple definition is: “Context engineering is curating what the model sees so that you get a better result.” For coding agents, there is an emerging set of context engineering approaches and terms. The foundation of it are the configuration features offered by the tools, and then the nitty gritty of part is how we conceptually use those features. ... One of the goals of context engineering is to balance the amount of context given - not too little, not too much. Even though context windows have technically gotten really big, that doesn’t mean that it’s a good idea to indiscriminately dump information in there. An agent’s effectiveness goes down when it gets too much context, and too much context is a cost factor as well of course. Some of this size management is up to the developer: How much context configuration we create, and how much text we put in there. My recommendation would be to build context like rules files up gradually, and not pump too much stuff in there right from the start. ... As I said in the beginning, these features are just the foundation for humans to do the actual work and filling these with reasonable context. It takes quite a bit of time to build up a good setup, because you have to use a configuration for a while to be able to say if it’s working well or not - there are no unit tests for context engineering. Therefore, people are keen to share good setups with each other.

Reimagining The Way Organizations Hire Cyber Talent

The way we hire cybersecurity professionals is fundamentally flawed. Employers post unicorn job descriptions that combine three roles’ worth of responsibilities into one. Qualified candidates are filtered out by automated scans or rejected because their resumes don’t match unrealistic expectations. Interviews are rushed, mismatched, or even faked—literally, in some cases. On the other side, skilled professionals—many of whom are eager to work—find themselves lost in a sea of noise, unable to connect with the opportunities that align with their capabilities and career goals. Add in economic uncertainty, AI disruption and changing work preferences, and it’s clear the traditional hiring playbook simply isn’t working anymore. ... Part of fixing this broken system means rethinking what we expect from roles in the first place. Jones believes that instead of packing every security function into a single job description and hoping for a miracle, organizations should modularize their needs. Need a penetration tester for one month? A compliance SME for two weeks? A security architect to review your Zero Trust strategy? You shouldn’t have to hire full-time just to get those tasks done. ... Solving the cybersecurity workforce challenge won’t come from doubling down on job boards or resume filters. But organizations may be able to shift things in the right direction by reimagining the way they connect people to the work that matters—with clarity, flexibility and mutual trust.

News sites are locking out the Internet Archive to stop AI crawling. Is the ‘open web’ closing?

Publishers claim technology companies have accessed a lot of this content for free and without the consent of copyright owners. Some began taking tech companies to court, claiming they had stolen their intellectual property. High-profile examples include The New York Times’ case against ChatGPT’s parent company OpenAI and News Corp’s lawsuit against Perplexity AI. ... Publishers are also using technology to stop unwanted AI bots accessing their content, including the crawlers used by the Internet Archive to record internet history. News publishers have referred to the Internet Archive as a “back door” to their catalogues, allowing unscrupulous tech companies to continue scraping their content. ... The opposite approach – placing all commercial news behind paywalls – has its own problems. As news publishers move to subscription-only models, people have to juggle multiple expensive subscriptions or limit their news appetite. Otherwise, they’re left with whatever news remains online for free or is served up by social media algorithms. The result is a more closed, commercial internet. This isn’t the first time that the Internet Archive has been in the crosshairs of publishers, as the organisation was previously sued and found to be in breach of copyright through its Open Library project. ... Today’s websites become tomorrow’s historical records. Without the preservation efforts of not-for-profit organisations like The Internet Archive, we risk losing vital records.

Who will be the first CIO fired for AI agent havoc?

As CIOs deploy teams of agents that work together across the enterprise, there’s a risk that one agent’s error compounds itself as other agents act on the bad result, he says. “You have an endless loop they can get out of,” he adds. Many organizations have rushed to deploy AI agents because of the fear of missing out, or FOMO, Nadkarni says. But good governance of agents takes a thoughtful approach, he adds, and CIOs must consider all the risks as they assign agents to automate tasks previously done by human employees. ... Lawsuits and fines seem likely, and plaintiffs will not need new AI laws to file claims, says Robert Feldman, chief legal officer at database services provider EnterpriseDB. “If an AI agent causes financial loss or consumer harm, existing legal theories already apply,” he says. “Regulators are also in a similar position. They can act as soon as AI drives decisions past the line of any form of compliance and safety threshold.” ... CIOs will play a big role in figuring out the guardrails, he adds. “Once the legal action reaches the public domain, boards want answers to what happened and why,” Feldman says. ... CIOs should be proactive about agent governance, Osler recommends. They should require proof for sensitive actions and make every action traceable. They can also put humans in the loop for sensitive agent tasks, design agents to hand off action when the situation is ambiguous or risky, and they can add friction to high-stakes agent actions and make it more difficult to trigger irreversible steps, he says.

Measuring What Matters: Balancing Data, Trust and Alignment for Developer Productivity

Organizations need to take steps over and above these frameworks. It's important to integrate those insights with qualitative feedback. With the right balance of quantitative and qualitative data insights, companies can improve DevEx, increase employee engagement, and drive overall growth. Productivity metrics can only be a game-changer if used carefully and in conjunction with a consultative human-based approach to improvement. They should be used to inform management decisions, not replace them. Metrics can paint a clear picture of efficiency, but only become truly useful once you combine them with a nuanced view of the subjective developer experience. ... People who feel safe at work are more productive and creative, so taking DevEx into account when optimizing processes and designing productivity frameworks includes establishing an environment where developers can flag unrealistic deadlines and identify and solve problems together, faster. Tools, including integrated development environments (IDEs), source code repositories and collaboration platforms, all help to identify the systemic bottlenecks that are disrupting teams' workflows and enable proactive action to reduce friction. Ultimately, this will help you build a better picture of how your team is performing against your KPIs, without resorting to micromanagement. Additionally, when company priorities are misaligned, confusion and complexity follow, which is exhausting for developers, who are forced to waste their energy on bridging the gaps, rather than delivering value.