Daily Tech Digest - October 25, 2025

---

Quote for the day:

"The most powerful leadership tool you have is your own personal example." -- John Wooden

---

The day the cloud went dark

This week, the impossible happened—again. Amazon Web Services, the backbone of the digital economy and the world’s largest cloud provider, suffered a large-scale outage. If you work in IT or depend on cloud services, you didn’t need a news alert to know something was wrong. Productivity ground to a halt, websites failed to load, business systems stalled, and the hum of global commerce was silenced, if only for a few hours. The impact was immediate and severe, affecting everything from e-commerce giants to startups, including my own consulting business. ... Some businesses hoped for immediate remedies from AWS’s legendary service-level agreements. Here’s the reality: SLA credits are cold comfort when your revenue pipeline is in freefall. The truth that every CIO has faced at least once is that even industry-leading SLAs rarely compensate for the true cost of downtime. They don’t make up for lost opportunities, damaged reputations, or the stress on your teams. ... This outage is a wake-up call. Headlines will fade, and AWS (and its competitors) will keep promising ever-improving reliability. Just don’t forget the lesson: No matter how many “nines” your provider promises, true business resilience starts inside your own walls. Enterprises must take matters into their own hands to avoid existential risk the next time lightning strikes.

Application Modernization Pitfalls: Don't Let Your Transformation Fail

Modernizing legacy applications is no longer a luxury — it’s a strategic imperative. Whether driven by cloud adoption, agility goals, or technical debt, organizations are investing heavily in transformation. Yet, for all its potential, many modernization projects stall, exceed budgets, or fail to deliver the expected business value. Why? The transition from a monolithic legacy system to a flexible, cloud-native architecture is a complex undertaking that involves far more than just technology. It's a strategic, organizational, and cultural shift. And that’s where the pitfalls lie. ... Application modernization is not just a technical endeavor — it’s a strategic transformation that touches every layer of the organization. From legacy code to customer experience, from cloud architecture to compliance posture, the ripple effects are profound. Yet, the most overlooked ingredient in successful modernization isn’t technology — it’s leadership: Leadership that frames modernization as a business enabler, not a cost center; Leadership that navigates complexity with clarity, acknowledging legacy constraints while championing innovation; Leadership that communicates with empathy, recognizing that change is hard and adoption is earned, not assumed. Modernization efforts fail not because teams lack skill, but because they lack alignment. 

CIOs will be on the hook for business-led AI failures

While some business-led AI projects include CIO input, AI experts have seen many organizations launch AI projects without significant CIO or IT team support. When other departments launch AI projects without heavy IT involvement, they may underestimate the technical work needed to make the projects successful, says Alek Liskov, chief AI officer at data refinery platform provider Datalinx AI. ... “Start with the tech folks in the room first, before you get much farther,” he says. “I still see many organizations where there’s either a disconnect between business and IT, or there’s lack of speed on the IT side, or perhaps it’s just a lack of trust.” Despite the doubts, IT leaders need to be involved from the beginning of all AI projects, adds Bill Finner, CIO at large law firm Jackson Walker. “AI is just another technology to add to the stack,” he says. “Better to embrace it and help the business succeed then to sit back and watch from the bench.” ... “It’s a great opportunity for CIOs to work closely with all the practice areas both on the legal and business professional side to ensure we’re educating everyone on the capabilities of the applications and how they can enhance their day-to-day workflows by streamlining processes,” Finner says. “CIOs love to help the business succeed, and this is just another area where they can show their value.”

Three Questions That Help You Build a Better Software Architecture

You don’t want to create an architecture for a product that no one needs. And in validating the business ideas, you will test assumptions that drive quality attributes like scalability and performance needs. To do this, the MVP has to be more than a Proof of Concept - it needs to be able to scale well enough and perform well enough to validate the business case, but it does not need to answer all questions about scalability and performance ... yet. ... Achieving good performance while scaling can also mean reworking parts of the solution that you’ve already built; solutions that perform well with a few users may break down as load is increased. On the other hand, you may never need to scale to the loads that cause those failures, so overinvesting too early can simply be wasted effort. Many scaling issues also stem from a critical bottleneck, usually related to accessing a shared resource. Spotting these early can inform the team about when, and under what conditions, they might need to change their approach. ... One of the most important architectural decisions that teams must make is to decide how they will know that technical debt has risen too far for the system to be supportable and maintainable in the future. The first thing they need to know is how much technical debt they are actually incurring. One way they can do this is by recording decisions that incur technical debt in their Architectural Decision Record (ADR).

Ransomware recovery perils: 40% of paying victims still lose their data

Decryptors are frequently slow and unreliable, John adds. “Large-scale decryption across enterprise environments can take weeks and often fails on corrupted files or complex database systems,” he explains. “Cases exist where the decryption process itself causes additional data corruption.” Even when decryptor tools are supplied, they may contain bugs, or leave files corrupted or inaccessible. Many organizations also rely on untested — and vulnerable — backups. Making matters still worse, many ransomware victims discover that their backups were also encrypted as part of the attack. “Criminals often use flawed or incompatible encryption tools, and many businesses lack the infrastructure to restore data cleanly, especially if backups are patchy or systems are still compromised,” says Daryl Flack, partner at UK-based managed security provider Avella Security and cybersecurity advisor to the UK Government. ... “Setting aside funds to pay a ransom is increasingly viewed as problematic,” Tsang says. “While payment isn’t illegal in itself, it may breach sanctions, it can fuel further criminal activity, and there is no guarantee of a positive outcome.” A more secure legal and strategic position comes from investing in resilience through strong security measures, well-tested recovery plans, clear reporting protocols, and cyber insurance, Tsang advises.

In IoT Security, AI Can Make or Break

Ironically, the same techniques that help defenders also help attackers. Criminals are automating reconnaissance, targeting exposed protocols common in IoT, and accelerating exploitation cycles. Fortinet recently highlighted a surge in AI-driven automated scanning (tens of thousands of scans per second), where IoT and Session Initiation Protocol (SIP) endpoints are probed earlier in the kill chain. That scale turns "long-tail" misconfigurations into early footholds. Worse, AI itself is susceptible to attack. Adversarial ML (machine learning) can blind or mislead detection models, while prompt injection and data poisoning can repurpose AI assistants connected to physical systems. ... Move response left. Anomaly detection without orchestration just creates work. It's important to pre-stage responses such as quarantine VLANs, Access Control List (ACL) updates, Network Access Control (NAC) policies, and maintenance window tickets. This way, high-confidence detections contain first and ask questions second. Finally, run purple-team exercises that assume AI is the target and the tool. This includes simulating prompt injection against your assistants and dashboards; simulating adversarial noise against your IoT Intrusion Detection System (IDS); and testing whether analysts can distinguish "model weirdness" from real incidents under time pressure.

Cyber attack on Jaguar Land Rover estimated to cost UK economy £1.9 billion

Most of the estimated losses stem from halted vehicle production and reduced manufacturing output. JLR’s production reportedly dropped by around 5,000 vehicles per week during the shutdown, translating to weekly losses of approximately £108 million. The shock has cascaded across hundreds of suppliers and service providers. Many firms have faced cash-flow pressures, with some taking out emergency loans. To mitigate the fallout, JLR has reportedly cleared overdue invoices and issued advance payments to critical suppliers. ... The CMC’s Technical Committee urged businesses and policymakers to prioritise resilience against operational disruptions, which now pose the greatest financial risk from cyberattacks. The committee recommended identifying critical digital assets, strengthening segmentation between IT and operational systems, and ensuring robust recovery plans. It also called on manufacturers to review supply-chain dependencies and maintain liquidity buffers to withstand prolonged shutdowns. Additionally, it advised insurers to expand cyber coverage to include large-scale supply chain disruption, and urged the government to clarify criteria for financial support in future systemic cyber incidents.

Thinking Machines challenges OpenAI's AI scaling strategy: 'First superintelligence will be a superhuman learner'

To illustrate the problem with current AI systems, Rafailov offered a scenario familiar to anyone who has worked with today's most advanced coding assistants. "If you use a coding agent, ask it to do something really difficult — to implement a feature, go read your code, try to understand your code, reason about your code, implement something, iterate — it might be successful," he explained. "And then come back the next day and ask it to implement the next feature, and it will do the same thing." The issue, he argued, is that these systems don't internalize what they learn. "In a sense, for the models we have today, every day is their first day of the job," Rafailov said. ... "Think about how we train our current generation of reasoning models," he said. "We take a particular math problem, make it very hard, and try to solve it, rewarding the model for solving it. And that's it. Once that experience is done, the model submits a solution. Anything it discovers—any abstractions it learned, any theorems—we discard, and then we ask it to solve a new problem, and it has to come up with the same abstractions all over again." That approach misunderstands how knowledge accumulates. "This is not how science or mathematics works," he said. ... The objective would fundamentally change: "Instead of rewarding their success — how many problems they solved — we need to reward their progress, their ability to learn, and their ability to improve."

Demystifying Data Observability: 5 Steps to AI-Ready Data

Data observability ensures data pipelines capture representative data, both the expected and the messy. By continuously measuring drift, outliers, and unexpected changes, observability creates the feedback loop that allows AI/ML models to learn responsibly. In short, observability is not an add-on; it is a foundational practice for AI-ready data. ... Rather than relying on manual checks after the fact, observability should be continuous and automated. This turns observability from a reactive safety net into a proactive accelerator for trusted data delivery. As a result, every new dataset or transformation can generate metadata about quality, lineage, and performance, while pipelines can include regression tests and alerting as standard practice. ... The key is automation. Rather than policies that sit in binders, observability enables policies as code. In this way, data contracts and schema checks that are embedded in pipelines can validate that inputs remain fit for purpose. Drift detection routines, too, can automatically flag when training data diverges from operational realities while governance rules, from PII handling to lineage, are continuously enforced, not applied retroactively. ... It’s tempting to measure observability in purely technical terms such as the number of alerts generated, data quality scores, or percentage of tables monitored. But the real measure of success is its business impact. Rather than numbers, organizations should ask if it resulted in fewer failed AI deployments. 

AI heavyweights call for end to ‘superintelligence’ research

Superintelligence isn’t just hype. It’s a strategic goal determined by a privileged few, and backed by hundreds of billions of dollars in investment, business incentives, frontier AI technology, and some of the world’s best researchers. ... Human intelligence has reshaped the planet in profound ways. We have rerouted rivers to generate electricity and irrigate farmland, transforming entire ecosystems. We have webbed the globe with financial markets, supply chains, air traffic systems: enormous feats of coordination that depend on our ability to reason, predict, plan, innovate and build technology. Superintelligence could extend this trajectory, but with a crucial difference. People will no longer be in control. The danger is not so much a machine that wants to destroy us, but one that pursues its goals with superhuman competence and indifference to our needs. Imagine a superintelligent agent tasked with ending climate change. It might logically decide to eliminate the species that’s producing greenhouse gases. ... For years, efforts to manage AI have focused on risks such as algorithmic bias, data privacy, and the impact of automation on jobs. These are important issues. But they fail to address the systemic risks of creating superintelligent autonomous agents. The focus has been on applications, not the ultimate stated goal of AI companies to create superintelligence.