Daily Tech Digest - January 25, 2025


Quote for the day:

“You live longer once you realize that any time spent being unhappy is wasted.” -- Ruth E. Renkl


How to Prepare for Life After NB-IoT

Last November, the IoT world was caught off guard by AT&T’s announcement to discontinue its support for Narrowband IoT (NB-IoT) by Q1 2025. For many, this came as a big surprise. NB-IoT was considered the prodigy technology, promising low power, long-range, and low-cost connectivity. While NB-IoT never reached mass adoption in the US, this decision still struck as a blow for the ones who did invest in the technology, and raised concerns about its validity among people outside the US. ... Fortunately, most IoT modules supporting NB-IoT also support LTE-M. Modules typically select the optimal network and technology based on signal quality and internal radio settings. Devices with roaming enabled will automatically switch networks or technologies if the primary connection fails. Once AT&T shuts down its network, your devices will automatically switch to another technology or network if set up correctly. However, rather than waiting for the network to become unavailable, you may want to stay in control of transitioning to another technology. This also allows you to test the process with a subset of devices before rolling out updates to your entire fleet. Assuming your cellular modules support LTE-M, and you have remote access to update configurations, you can update the radio access technology (RAT) using a simple AT Command. 
Creating efficient, relevant, and lasting regulations requires several key factors. First and foremost, policymakers need a working definition of the object of their laws, which requires thorough work to capture the essence of what will be affected by their text. This is a challenging task in the case of AI because its definition remains in flux as the technology evolves. ... unprecedented surge in generative AI’s popularity created uncertainty for policymakers about how to navigate the new landscape. There was an urgent need for frameworks, definitions, and language to fully understand the impact of this technology and how to frame it. As the technology outpaced expectations, earlier regulatory efforts to address these tools quickly became inadequate and obsolete, leaving policymakers scrambling to catch up. This is precisely the situation Chinese regulators faced in their initial efforts to address the generative AI sector. The basic provisions outlined in the law were insufficient to address the profound societal impacts of generative AI’s widespread adoption. The attempt to establish China as an early player in AI regulation was overtaken by the pace of technological progress and private-sector innovation, rendering even the terminology obsolete.


How to Simplify Automated Security Testing in CI/CD Pipelines

Dependency management is where many teams stumble, and we’ve all seen the fallout from poorly managed libraries (hello, Log4Shell). Automating dependency checks is not an option, it’s a must. Tools like Dependabot, OWASP Dependency-Check, and Renovate take the grunt work out of monitoring for vulnerabilities, raising alerts, and even creating pull requests to fix issues. Imagine a Node.js team drowning in a sea of outdated packages. With Dependabot hooked into their GitHub workflow, every vulnerability gets an automatic pull request to update to a safe version. No manual labor, no guessing games—just a steady rhythm of secure, up-to-date code. Go deeper by leveraging Software Composition Analysis (SCA) tools that don’t just look at direct dependencies but dive into the murky waters of transitive dependencies too.  ... Instead of vague warnings like “Potential SQL Injection Found,” imagine getting, “SQL Injection vulnerability detected at line 45 in user_controller.js. Here’s how to fix it…” Tools like CodeQL and Semgrep do precisely this. They integrate directly into CI pipelines, flag issues, suggest fixes, and provide links to further reading, all without overwhelming the dev team.


Security automation and integration can smooth AppSec friction

Whether you perceive friction between development and security testing to be an impediment or not often depends on your role in the organization. Of the AppSec team members who responded to the survey used for “Global State of DevSecOps” report, 65% felt that that testing impeded pipelines “moderately” or “severely.” While the report didn’t survey why they feel this way, we can speculate that it’s due to their proximity to the testing process, or potentially because they’re feeling pressure to accelerate review processes. Since they are closest to the task, they face the highest scrutiny for its efficiency. Of the development and engineering team members who replied to the survey, 58% share the sentiment of their AppSec counterparts. It is, however, important to consider that an additional 12% of the surveyed developers and engineers report that they just don’t have enough visibility into security testing to know what’s going on. Were they to have greater visibility into security testing processes, it is quite possible that they, too, would perceive AppSec testing as an impediment to pipelines. And this lack of visibility makes concerted DevSecOps initiatives more difficult to implement since contributors are unable to close feedback loops or optimize development and testing efforts.


The Power of Many: Crowdsourcing as A Game-Changer for Modern Cyber Defense

Although shared expertise significantly boosts threat detection & hunting efficiency while simultaneously empowering cybersecurity education, there are several stumbling blocks to address on the way to building global crowdsourcing initiatives. While working towards a safer future, contributors to crowdsourced efforts often face issues related to intellectual property rights and the recognition of the significance of individual contributions within the professional network. Ensuring proper recognition for discoveries and contributions to global cyber defense at all levels, from the support of author attribution in the code of a detection rule to sharable digital credentials issued by organizations to recognize exceptional individual involvement and contributions to the crowdsourcing initiatives, is essential to maintaining motivation and fairness. Another challenge is adherence to privacy imperative and compliance with security regulations, including TLP protocol, while sharing information with a wide audience, since disclosure of sensitive information about vulnerabilities or cyber attacks can pose significant risks both to crowdsourcing program contributors and beneficiaries.


How to Harness the Power of Fear and Transform It Into a Leadership Strength

One of the most powerful ways to address fear is to reframe it as a perception rather than an absolute truth. Fear does not objectify threats; it is just one of the mental faculties. By reframing it as a perception, a leader can make proper decisions, attacking the instances of fear. Refocusing the process does not stop fear; it changes the process. Leaders are able to go from impulsive to composed behavior by understanding that fear is a conceptual state rather than an actual one. Calm neurotransmitters like serotonin and endorphins take the role of stress chemicals like cortisol and adrenaline, promoting emotional equilibrium and resilience. For leaders, that shift can be radical. By approaching challenges and approach with strength and rationality, not fear, they can spread the ripple effect into their companies. It's a way of creating an environment in which teams feel empowered, excited and pushed to grow and thrive. ... Recognizing fear as a perceived threat allows leaders to respond with reason and confidence. Mastering fear is a critical leadership skill, fostering innovation and collaboration. By transforming fear into a tool for growth, leaders unlock their full potential and inspire others, paving the way for sustained progress.


Nuclear-Powered Data Centers: When Will SMRs Finally Take Off?

Taking stock of the nuclear-powered data center market in 2024, Alan Howard, principal analyst of cloud and colocation services at Omdia,* said: “It’s nothing short of exciting that Amazon, Google, and Microsoft have all signed deals for nuclear power… and Meta is publicly on the hunt.” Still, these deals are relatively small by the standards of the data center industry, and Howard cautioned against impatience, citing the mid-2030s as the earliest we can expect to see broad commercial availability of nuclear energy in powering data centers. “The reality is that these [nuclear reactors under construction] are essentially test reactors which is part of the long regulatory road nuclear technology companies must follow.” ... One of the chief challenges facing data center companies is the five-to-seven-year permitting and construction timelines for nuclear facilities, according to Ryan Mallory, COO at data center firm Flexential. “Data center companies must begin securing permits, ground space, and operational expertise to prepare for SMRs to become scalable and repeatable by the 2030s,” Mallory said. There are also technological challenges, according to Steven Carlini, chief data center and AI advocate at Schneider Electric. “Integrating SMRs into the existing ecosystem will be complex,” he said.


13 Cybersecurity Predictions for 2025

AI capabilities are awesome, yet I’m finding that most of the AI capabilities being developed are focused on just getting them to work and into the marketplace as soon as possible. We need to do a much better job of incorporating cybersecurity best practices and secure-by-design principles into the creation, operation, and sustainment of AI systems. The AI Security and Incident Response Team (AISIRT)[ii] here at the Software Engineering Institute has discovered numerous material weaknesses and flaws in AI capabilities resulting in vulnerabilities that can be leveraged by hostile entities. AI vulnerabilities are cyber vulnerabilities, and the list of reported vulnerabilities continue to grow. Software engineers are trained to incorporate secure-by-design principles into their work. But neural-network models, including generative AI and LLMs, bring along a wide range of additional kinds of weaknesses and vulnerabilities, and for many of these it is a struggle to develop effective remediations. Until the AI community is able to develop AI-appropriate secure-by-design best practices to augment the secure-by-design practices already familiar to software engineers, I believe we’ll see preventable cyber incidents affecting AI capabilities in 2025. ... Ransomware criminal activity continues to feast on the cyber poor. Cyber criminals have been feasting on those who operate below the cyber poverty line.


Biometrics Institute identifies dire need for clear language in biometrics and AI

Most biometrics experts agree that no one is exactly sure what anyone is talking about. The Biometrics Institute is trying to help, via its Explanatory Dictionary, a resource that aims to capture the nuances in biometric terminology, “considering both formal definitions and how they are perceived by the public – for example, how someone might explain biometrics or AI to a friend.” Because, as of now, there isn’t a standard that is universally agreed-on, nor is there really a clear way to explain biometrics and AI to your neighbour Ted who works in marketing. “There are no universal definitions of biometrics or AI and those put forward by ISO and some governments are either too technical, obtuse or are not fully aligned with one another or are hidden behind paywalls and not accessible to the majority of the general public.” The paper drills down on the semantics of biometric grammar. What does it mean for a biometric application to “have AI”? Conflation of certain terms in both regulatory and public contexts exacerbates the problem. Media struggles to pick apart the web of language, and contributes its own strands in the process. Is a tool “AI-driven,” or “AI-equipped”? Where do algorithms fit in?


How AI Copilots Are Transforming Threat Detection and Response

The rise of AI copilots in cybersecurity is a transformative moment, but it requires a shift in mindset. Security teams should view these tools as partners, not replacements. AI copilots excel at processing vast datasets and identifying patterns, but humans are irreplaceable when it comes to judgment and understanding context. The future of cybersecurity lies in this hybrid approach, where AI enhances human capabilities rather than attempting to replicate them. Business leaders should focus on fostering this collaboration, equipping their teams with the skills and tools needed to work effectively with AI. Additionally, transparency is non-negotiable. Teams must understand how their AI copilots make decisions, ensuring accountability and reducing the risk of errors. This also involves rigorous testing and ongoing monitoring to detect and mitigate biases or vulnerabilities before they can be exploited. ... By empowering security teams with advanced capabilities, businesses can stay ahead of adversaries and secure a resilient future. Looking ahead, AI copilots are just the beginning. As these tools become more advanced, they will evolve beyond copilots into more autonomous AI agents—a shift often referred to as agentic AI. 


No comments:

Post a Comment