Daily Tech Digest by Kannan Subbiah: robots

Showing posts with label robots. Show all posts

Daily Tech Digest - December 29, 2025

Quote for the day:

"What great leaders have in common is that each truly knows his or her strengths - and can call on the right strength at the right time." -- Tom Rath

Beyond automation: Physical AI ushers in a new era of smart machines

“Physical AI has reached a critical inflection point where technical readiness aligns with market demand,” said James Davidson, chief artificial intelligence officer at Teradyne Robotics, a leader in advanced robotics solutions. “The market dynamics have shifted from skepticism to proof. Early adopters are reporting tangible efficiency and revenue gains, and we’ve entered what I’d characterize as the early-majority phase of adoption, where investment scales dramatically.” ... To train and prepare these models, a new specialized class of AI model emerged: World Foundation Models. WFMs serve two primary functions for robotics AI: They enable engineers to develop vast synthetic datasets rapidly to train robots on unseen actions, and they test these robots in virtual environments before real-world deployment. WFMs allow developers to create virtual training grounds that mimic reality through “digital twins” of environments. Within these simulated scenes, robots learn to navigate real-world challenges safely and at a pace far exceeding what physical presence would permit. ... Despite grabbing a lot of headlines, humanoid robots only represent a small fraction of AI robotics deployments. For now, it’s collaborative robots, robotic arms and autonomous mobile robots that are transforming warehouse and factory settings. The forefront example is Amazon.com Inc., which uses intelligent robots across its warehouses.

When Digital Excellence Turns Into Strategic Technical Debt

Asian Paints' digital architecture was built for a world that valued scale, predictability and discipline. Its systems continuously optimize for efficiency, minimize variability and ensure consistency across thousands of dealers and SKUs. For nearly 20 years, these capabilities have directly contributed to better margins, improved service levels and increased shareholder confidence. But today's market is different. New entrants, backed by capital and "largely free from legacy" process constraints, are willing to accept inefficiencies to gain market share quickly. ... The result is a market that is more volatile, more tactical, and less patient. Additionally, new technology plays a vital role in creating a competitive edge. This is where the strategic technical debt surfaces. Unlike traditional technical debt, this isn't about outdated systems or underinvestment. ... The difference lies in architecture and intent. Newer players are born cloud-native, with a more modular approach, better governance and greater tolerance for experimentation. They use analytics and AI proactively to adjust incentives quickly, test local pricing strategies and pivot dealer engagement models in response to demand. Speed and flexibility matter more than optimization. ... Strategic technical debt accumulates because CIOs are rewarded for stability, uptime and optimization. Optionality, speed and the ability to unlearn don't appear on scorecards. Over time, this imbalance becomes part of the architecture and results in digital stress.

The Evolution of North Korea – And What To Expect In 2026

What has changed most notably through 2024 and 2025 is the shift away from “purely external intrusion” towards “abuse of legitimate access,” says Pontiroli. “Rather than breaking in, North Korean operators increasingly aim to be hired as remote IT workers inside real companies, gaining steady income, trusted network access, and the option to pivot into espionage, data theft, or follow on attacks.” ... The workers claim to be US based with IT experience, “but in reality, they are North Korean or proxied by North Korean networks,” he explains. Over time, the threat actors have developed deep expertise in software engineering, mobile applications, blockchain infrastructure, and cryptocurrency ecosystems says Tom Hegel, distinguished threat researcher, SentinelLABS. ... In parallel, cybersecurity researchers have observed related campaigns with distinct names and tradecraft. A malicious campaign dubbed Contagious Interview involves threat actors masquerading as recruiters or employers to lure job seekers, particularly in tech and cryptocurrency sectors, into fake interviews that deliver malware such as BeaverTail, InvisibleFerret, and variants such as OtterCookie, says Pontiroli. ... Today, fake worker schemes remain an “active and growing threat,” says Jack. KnowBe4 offers training to customers to combat this and strengthen their security culture, he says. Security leaders must assume that the hiring pipeline itself is part of the attack surface, says Hegel.

Five Attack-Surface Management Trends to Watch in 2026

In 2026, regulators will anchor security and risk leaders’ approaches to exposure strategy. This will mean not only demonstrating due diligence during annual audits, but also demonstrating proof of resilience every day. Exposure management platforms that can map external assets against regulatory expectations; provide real-time compliance dashboards and metrics; and quantify benefits and exposures to boardrooms will become table stakes. ... Attackers see the enterprise as a single, unified attack surface, with each constituent part informing the next priority: cloud workloads, SaaS, subsidiaries, shadow IT, and third-party dependencies. In 2026, savvy security leaders will be adopting that same perspective. Point-in-time, penetration-test-style engagements and bug-bounty programs will give way to organizations that expect full-scope, attacker-centric discovery of digital asset footprints, as well as automated prioritization to cut through the noise. ... In 2026, successful vendor choices will be those that strike a balance between consolidation and integration. Enterprises will demand more flexible integration into existing workflows, including third-party APIs and visibility into SIEM, SOAR, and GRC tools, as well as the ability to support hybrid and multi-cloud environments without friction. Transparency and visibility into roadmap, enterprise-readiness proofs, and customer success will become significant differentiators in a category that has been defined by mergers and acquisitions.

Daon outlines five digital identity shifts for 2026

Daon said non-human identities, including agentic AI systems, are expanding quickly across enterprise networks. It cited independent 2025 studies reporting roughly 44% year-on-year growth in non-human identities and a rise in machine-to-human ratios from around 80:1 to 144:1 in some environments. The prediction for 2026 is that enterprises will treat autonomous and agentic systems as full participants in the identity lifecycle. These systems would be registered, authenticated, authorised and monitored under formal policies, with containment processes defined in case of compromise or misbehaviour. ... Daon said progress in techniques such as zero-knowledge proofs, federated learning and sensor attestation now enables biometric checks on personal devices while reducing movement of raw biometric data. On-device processing can bind verification to a specific capture environment and lower the risk of replay or injection. Local storage of biometric templates supports data-minimisation approaches. The company expects these on-device checks to align with proof-of-possession flows and hardware-backed sensor attestations. It said federated learning and zero-knowledge techniques allow systems to validate claims without sharing underlying biometric templates with servers. ... Daon expects continued pressure on pre-hire verification because of deepfake applicants and impersonation. It said the more significant change in 2026 will come after hiring as employers adopt continuous workforce assurance.

Quantum computing made measurable progress toward real-world use in 2025

Fully functional quantum computers remain out of reach, but optimism across the field is rising. At the Q2B Silicon Valley conference in December, researchers and executives pointed to a year marked by tangible progress – particularly in hardware performance and scaling – and a growing belief that quantum advantage for real-world problems may be achievable sooner than expected. "More people are getting access to quantum computers than ever before, and I have a suspicion that they'll do things with them that we could never even think of," said Jamie Garcia at IBM. ... Aaronson, long known for his critical analysis of claims in quantum computing, described the progress in qubit fidelity and control systems as "spectacular." However, he cautioned that new algorithms remain essential for converting that hardware performance into practical value. While technical strides have been impressive, translating those advances into applications remains difficult. Ryan Babbush of Google Quantum AI said hardware continues to outpace software in usefulness. ... Dutch startup QuantWare introduced an architecture aimed at solving one of the industry's most significant hardware limitations: scaling up without losing reliability. The company's superconducting quantum processor design targets 10,000 qubits, roughly 100 times more than today's leading devices. QuantWare's Matt Rijlaarsdam said the first systems of this size could be operational within 2.5 years.

Ship Reliable AI: 7 Painfully Practical DevOps Moves

In AI land, “what changed” is anything that teaches or nudges the model: training data slices, prompt templates, system instructions, retrieval schemas, embeddings pipelines, tokenizer versions, and the model binary itself. We treat each as code. Prompts live next to code with unit tests. We commit small evaluation sets in-repo for quick signals, and keep larger benchmarks in object storage with content hashes and a manifest. ... Shiny demos hide flaky edges. We force those edges to show up in CI, where they’re cheap. Our pipeline runs fast unit tests, a tiny evaluation suite, and a couple of safety checks against handcrafted adversarial prompts. The goal isn’t to solve safety in CI; it’s to block footguns. We test the glue code around the model, we lint prompts for hard-to-diff formatting changes, and we run a 50-example eval that catches obvious regressions in latency, grounding, and accuracy. ... For AI pods, that starts with resource quotas and limits. GPU nodes are expensive; “just one more experiment” can melt the budget by lunch. We set namespace-level quotas for GPU and memory, and we stop requests that try to sneak past. For egress, we deny everything and allow only the API endpoints our apps need. When someone tries to point a staging pod at a random external endpoint “just to test it,” the policy does the talking.

What support is available for implementing Agentic AI systems

The adoption of Agentic AI systems is reshaping the way organizations implement security measures, particularly for NHIs. Agentic AI—capable of self-directed learning and decision-making—proves advantageous in deploying security protocols that adapt in real-time to evolving threats. By utilizing such technology, organizations can leverage data-driven insights to enhance their NHI management strategies. ... Given the critical role of NHIs in maintaining robust cloud security, organizations need to adopt advanced methodologies that integrate seamlessly with their existing security frameworks. ... Effective NHI management relies heavily on leveraging insights that stem from analyzing large data sets. Organizations that prioritize the use of data analytics in their cybersecurity strategies can efficiently discover, classify, and monitor machine identities and their associated secrets. Advanced analytical tools can help security teams identify patterns and anomalies in system activities, providing early indicators of potential security threats. These insights make it possible to implement more effective security protocols and prevent unauthorized access before it happens. ... The security of an organization is not solely the responsibility of the IT department; it is a shared responsibility across all stakeholders. Building a culture of security awareness is crucial in ensuring that every member of an organization understands the role that NHIs play in cybersecurity.

Godspeed curtain twitchers: DPDP and its peers just got ruthless

Organisations will have to work on privacy very seriously- in everyday business operations and in every area, Bhambry cautions. They will have to make sure it pervades product development, processes (From the onset), internal audit, regular training and the very culture of that company and its employees. Enterprises will have to focus on individual rights, consent protocols and data governance.” There is no doubt that data privacy is going to get stronger, transparent, and comprehensive, affirms Advocate Dr. Bhavna Sharma, Delhi High Court. Cybercrime Expert and Legal Consultant, Delhi Police and a techno-legal policy professional. But it is also going to get complex in 2026 as it shifts from abstract legal principles to a tangible operational mandate with the notification of the DPDPA Rules, 2025, adds Dr. Sharma ... “India’s DPDPA and MeitY’s localisation mandates echo a growing consensus that data sovereignty equals digital sovereignty. Governments are recognising that control over citizen data is foundational to national security and economic resilience.” Cheema explains. In an era marked by competition among nations with their own data systems, state leaders are taking control, Yadav observes. “They are not willing to allow strategic assets to slip through their fingers. And as a result, the government calls for ‘localisation’ to trap extra-territorial storage simply because it has yet to be regulated by authorities in those countries.

Tech innovations fuelling Indian GCCs as BFSI powerhouses

Responsible AI governance, model explainability, and auditability remain difficult across regulated domains worldwide. Institutions everywhere also face constraints around scalable compute, high-quality data flows, and real-time analytics. As AI systems process more sensitive financial data, cybersecurity risks are rising across the industry, prompting greater investment in zero-trust architectures, model-security testing, and stronger third-party controls. ... GCCs in India have been instrumental in orchestrating cloud migrations for complex banking systems, allowing banks and insurers to transition from monolithic legacy systems toward microservices and API-led platforms. This modular architecture has enabled financial institutions to launch products rapidly and build disaster resilience. Additionally, regulatory complexity and rising compliance costs have created a fertile ground for RegTech innovation. Indian GCCs are helping global enterprises build AI-powered KYC and Anti-Money Laundering (AML) solutions, compliance dashboards, and automated regulatory reporting pipelines that reduce manual work and false positives and make audits more efficient. ... Security, observability, and governance have also become board-level priorities. According to industry insights, as GCCs ingest more sensitive financial data and run mission-critical AI models, investments in cyber-resilience, third-party access monitoring, and federated data controls have surged.

Daily Tech Digest - October 31, 2025

Quote for the day:

“The more you loose yourself in something bigger than yourself, the more energy you will have.” -- Norman Vincent Peale

Breaking the humanoid robot delusion

The robot is called . The company says NEO is the world’s first consumer-ready humanoid robot for the home. It is designed to automate routine chores and offer personal help so you can spend time on other things. ... Full autonomy in perceiving, planning, and manipulating like a human is a massive technology challenge. Robots have to be meticulously and painstakingly trained on every single movement, learn to recognize every object, and “understand” — for lack of a better word — how things move, how easily they break, what goes where, and what constitute appropriate actions. One major way humanoid robots are trained is with . A person wearing special equipment remotely controls prototype robots, training them for many hours on how to, say, fold a shirt. Many hours more are required to train the robot how to fold a smaller child’s shirt. Every variable, from the height of the folding table to the flexibility of the fabrics has to be trained separately. ... The temptation to use impressive videos of remotely controlled robots where you can’t see the person controlling them to raise investment money, inspire stock purchases and outright sell robot products, appears to be too strong to resist. Realistically, the technology for a home robot that operates autonomously the way the NEO appears to do in the videos in arbitrary homes under real-world conditions is many years in the future, possibly decades.

Your vendor’s AI is your risk: 4 clauses that could save you from hidden liability

The frontier of exposure now extends to your partners’ and vendors’ use. The main question being: Are they embedding AI into their operations in ways you don’t see until something goes wrong? ... Require vendors to formally disclose where and how AI is used in their delivery of services. That includes the obvious tools and embedded functions in productivity suites, automated analytics and third-party plug-ins. ... Include explicit language that your data may not be used to train external models, incorporated into vendor offerings or shared with other clients. Require that all data handling comply with the strictest applicable privacy laws and specify that these obligations survive the termination of the contract. ... Human oversight ensures that automated outputs are interpreted in context, reviewed for bias and corrected when the system goes astray. Without it, organizations risk over-relying on AI’s efficiency while overlooking its blind spots. Regulatory frameworks are moving in the same direction: for example, high-risk AI systems must have documented human oversight mechanisms under the . ... Negotiate liability provisions that explicitly cover AI-driven issues, including discriminatory outputs, regulatory violations and errors in financial or operational recommendations. Avoid generic indemnity language. Instead, AI-specific liability should be made its own section in the contract, with remedies that scale to the potential impact.

AI chatbots are sliding toward a privacy crisis

The problem reaches beyond internal company systems. Research shows that some of the most used AI platforms collect sensitive user data and share it with third parties. Users have little visibility into how their information is stored or reused, leaving them with limited control over its life cycle. This leads to an important question about what happens to the information people share with chatbots. ... One of the more worrying trends in business is the growing use of , where employees turn to unapproved tools to complete tasks faster. These systems often operate without company supervision, allowing sensitive data to slip into public platforms unnoticed. Most employees admit to sharing information through these tools without approval, even as IT leaders point to data leaks as the biggest risk. While security teams see shadow AI as a serious problem, employees often view it as low risk or a price worth paying for convenience. “We’re seeing an even riskier form of shadow AI,” says Tim Morris, “where departments, unhappy with existing GenAI tools, start building their own solutions using open-source models like .” ... Companies need to do a better job of helping employees understand how to use AI tools safely. This matters most for teams handling sensitive information, whether it’s medical data or intellectual property. Any data leak can cause serious harm, from damaging a company’s reputation to leading to costly fines.

The true cost of a cloud outage

The top 2000 companies in the world pay approximately $400 billion for downtime each year. A simple calculation reveals that these organizations, including the Dutch companies , , , , and , lose around $200 million from their annual accounts due to unplanned downtime. Incidentally, what the Splunk study really revealed were the hidden costs of financial damage caused by problems with security tools, infrastructure, and applications. These can wipe billions off market values. ... A more conservative estimate of downtime costs can be found at Information Technology Intelligence Consulting, which conducted research on behalf of . The majority of the parties surveyed had more than 200 employees, but the combination was more diverse than the top 2000 companies worldwide. The costs of downtime were substantial: at least $300,000 per hour for 90 percent of the companies in question. Forty-one percent stated that IT outages cost between $1 million and $5 million. ... In theory, the largest companies can rely on a . In addition, hyperscalers absorb many local outages by routing traffic to other regions. However, multicloud is not something that you can just set up as a start-up SME. In addition, you usually do not build your applications in a fully redundant form in different clouds. Furthermore, it is quite possible that you can continue to work yourself, but that your product is inaccessible.

5 Reasons Why You’re Not Landing Leadership Roles

Is your posture confident? Do you maintain steady eye contact? Is the cadence, pace and volume of your voice engaging, assertive and compelling? Recruiters assess numerous factors on the . ... Are you showing a grasp of the prospective employer’s pain points and demonstrating an original point of view for how you will approach these problems? Treat senior level interviews like – you are an expert on their business, uncovering potential opportunities with insightful questions, and sharing enough of your expertise that you’re perceived as the solution. ... Title bumps are rare, so you need to give the impression that you are already operating at the in order to be hired as such. Your interview examples should include stories about how you initiated new ideas or processes, as well as measurable results that impact the bottom line. Your examples should specify how many people and dollars you have managed. Ideally, you have stories that show you can get results in up and down markets. ... The hiring process extends over multiple rounds, especially for leadership roles. Keep track of everyone that you have met, as well as what you have specifically discussed with each of them. Send personalized follow-up emails that engage each interviewer uniquely based on what you discussed. This differentiates you as someone who listens and cares about them specifically.

Why understanding your cyber exposure is your first line of defence

Thanks to AI, attacks are faster, more targeted and increasingly sophisticated. As the lines between the physical and digital blur, the threat is no longer isolated to governments or critical national infrastructure. Every organisation is now at risk. Understanding your cyber exposure is the key to staying ahead. This isn’t just a buzzword either; it’s about knowing where you stand and what’s at risk. Knowing every asset, every connection, every potential weakness across your digital ecosystem is now the first step in building a defence that can keep pace with modern threats. But before you can manage your exposure, you need to understand what’s driving it – and why the modern attack surface is so difficult to defend. ... By consolidating data from across the environment and layering it with contextual intelligence, cyber exposure management allows security teams to move beyond passive monitoring. It’s not just about seeing more, it’s about knowing what matters and acting on it. That means identifying risks earlier, prioritising them more effectively and taking action before they escalate. ... Effective and modern cybersecurity is shifting to shaping the battlefield before threats even arrive. That’s down to the value of understanding your cyber exposure. After all, it’s not just about knowing what’s in your environment, it’s about knowing how it all fits together – what’s exposed, what’s critical and where the next threat is likely to emerge.

Applications and the afterlife: how businesses can manage software end of life

Both enterprise software and personal applications have a lifecycle, set by the vendor’s support and maintenance. Once an application or operating system goes out of support, it will continue to run. But there will be no further feature updates and vitally, often no security patches. ... When software end of life is unexpected, it can cause serious disruption to business processes. In the very worst-case scenarios, enterprises will only know there is a problem when a key application no longer functions, or if a malicious actor exploits a vulnerability. The problem for CIOs and CISOs is keeping track of the end of life dates for applications across their entire stack, and understanding and mapping dependencies between applications. This applies equally to in-house applications, off the shelf software and open source. “End of life software is not necessarily bad,” says Matt Middleton-Leal, general manager for EMEA at Qualys. “It’s just not updated any more, and that can lead to vulnerabilities. According to our research, nearly half of the issues on the CISA Known Exploited Vulnerabilities (KEV) list are found in outdated and unsupported software.” As CISA points out, attackers are most likely to exploit older vulnerabilities, and to target unpatched systems. Risks come from old, and known vulnerabilities, which IT teams should have patched.

Tips for CISOs switching between industries

Building a transferable skill set is essential for those looking to switch industries. For Dell’s first-ever CISO, Tim Youngblood, adaptability was never a luxury but a requirement. His early years as a consultant at gave him a front-row seat to the challenges of multiple industries before he ever moved into cybersecurity. Those early years also taught Youngblood that while every industry has its own nuances, the core security principles remain constant. ... Making the jump into a new industry isn’t about matching past job titles but about proving you can create impact in a new context. DiFranco says the key is to demonstrate relevance early. “When I pitch a candidate, I explain what they did, how they did it, and what their impact was to their organization in their specific industry,” he says. “If what they did and how they did it, and what their impact was on the organization resonates where that company wants to go, they’re a lot more likely to say, ‘I don’t really care where this person comes from because they did exactly what I want done in this organization’. ... The biggest career risk for many CISOs isn’t burnout or data breach, it’s being seen as a one-industry operator. Ashworth’s advice is to focus on demonstrating transferable skills. “It’s a matter of getting whatever job you’re applying for, to realise that those principles are the same, no matter what industry you’re in. Whether it’s aerospace, healthcare, or finance, the principles are the same. Show that, and you’ll avoid being pigeonholed.”

Awareness Is the New Armor: Why Humans Matter Most in Cyber Defense

People remain the most unpredictable yet powerful variable in cybersecurity. Lapses like permission misconfiguration, accidental credential exposure, or careless data sharing continue to cause most incidents. Yet when equipped with the right tools and timely information, individuals can become the strongest line of defense. The challenge often stems from behavior rather than intent. Employees frequently bypass security controls or use unapproved tools in pursuit of productivity, unintentionally creating invisible vulnerabilities that go unnoticed within traditional defences. Addressing this requires more than restrictive policies. Security must be built into everyday workflows so that safe practices become second nature. ... Since technology alone cannot secure an organization, a culture of security-first thinking is essential. Leaders must embed security into everyday workflows, promote upskilling, and focus on reinforcement rather than punishment. This creates a workforce that takes ownership of cybersecurity, checking email sources, verifying requests, and maintaining vigilance in every interaction. is both a reminder and a rallying cry. India’s digital economy presents immense opportunity, but its threat surface expands just as fast.

Creepy AI Crawlers Are Turning the Internet into a Haunted House

The degradation of the internet and market displacement caused by commercial AI crawlers directly undermines people’s ability to access information online. This happens in various ways. First, the put significant technical strain on the internet, making it more difficult and expensive to access for human users, as their activity increases the time needed to access websites. Second, the trained on this scraped content now provide answers directly to user queries, reducing the need to visit the original sources and cutting off the traffic that once sustained content creators, including media outlets. ... AI crawlers represent a fundamentally different economic and technical proposition––a vampiric relationship rather than a symbiotic one. They harvest content, news articles, blog posts, and open-source code without providing the semi-reciprocal benefits that made traditional crawling sustainable. Little traffic flows back to sources, especially when search engines like Google start to provide AI generated summaries rather than sending traffic on to the websites their summaries are based on. ... What makes this worse is that these actors aren’t requesting books to read individual stories or conduct genuine research, they’re extracting the entire collection to feed massive language model systems. The library’s resources are being drained not to serve readers, but to build commercial AI products that will never send anyone back to the library itself.

Daily Tech Digest - April 15, 2025

Quote for the day:

“Become the kind of leader that people would follow voluntarily, even if you had no title or position.” -- Brian Tracy

Critical Thinking In The Age Of AI-Generated Code

Besides understanding our code, code reviewing AI-generated code is an invaluable skill nowadays. Tools like GitHub's Copilot and DeepCode can code-review better than a junior software developer. Depending on the complexity of the codebase, they can save us time in code reviewing and pinpoint cases that we may have missed, but, after all, they are not flawless. We still need to verify that the AI assistant's code review did not provide any false positives or false negatives. We need to verify that the code review did not miss anything important and that the AI assistant got the context correctly. The hybrid approach seems to be the most effective one: let AI handle the grunt work and rely on developers for the critical analysis. ... After all, code reviewing AI-generated code is an excellent opportunity to educate ourselves while improving our code-reviewing skills. Keep in mind that, to date, AI-generated code optimizes for patterns in its training data. This may not be aligned with coding first principles. AI-generated code may follow templated solutions rather than custom designs. It may include unnecessary defensive code or overly generic implementations. We need to check that it has chosen the most appropriate solution for each code block generated. Another common problem is that LLMs may hallucinate.

DeepCoder: Revolutionizing Software Development with Open-Source AI

One of the DeepCoder project’s most significant contributions is the introduction of verl-pipeline, an optimized extension of the very open-source RLHF library. The team identified sampling, the generation of long token sequences as the primary bottleneck in training and developed “one-off pipelining” to address this challenge. This technique overlaps sampling, reward calculation and training, reducing end-to-end training times by up to 2.5x. This optimization is game-changing for coding tasks requiring thousands of unit tests per reinforcement learning iteration, making previously prohibitive training runs accessible to smaller research teams and independent developers. For DevOps professionals, DeepCoder represents an opportunity to integrate advanced code generation directly into CI/CD pipelines without dependency on API-gated services. Teams can fine-tune the model on their codebase, creating customized assistants that understand their specific architecture and coding patterns. ... DeepCoder’s open-source nature aligns with the DevOps collaboration and shared improvement philosophy. As more organizations adopt and contribute to the model, we can expect to see specialized versions emerge for different programming languages and problem domains.

Transforming Software Development

AI assistants are getting smarter, moving beyond prompt-based interactions to anticipate developers’ needs and proactively offer suggestions. This evolution is driven by the rise of AI agents, which can independently execute tasks, learn from their experiences and even collaborate with other agents. Next year, these agents will serve as a central hub for code assistance, streamlining the entire software development lifecycle. AI agents will autonomously write unit tests, refactor code for efficiency and even suggest architectural improvements. Developers’ roles will need to evolve alongside these advancements. AI will not replace them. Far from it; proactive AI assistants and their underlying agents will help developers build new skills and free up their time to focus on higher-value, more strategic tasks. ... AI models are more powerful when trained on internal company data, which allows them to generate insights specific to an organization’s unique operations and objectives. However, this often requires running models on premises for security and compliance reasons. With open source models rapidly closing the performance gap with commercial offerings, more businesses will deploy models on premises in 2025. This will allow organizations to fine-tune models with their own data and deploy AI applications at a fraction of the cost.

Cybercriminal groups embrace corporate structures to scale, sustain operations

We have seen cross collaboration between groups that specialize in specific activities. For example, one group specializes in social engineering, while another focuses on scaling malware and botnets to uncover open servers that yield database breaches. They, in turn, can sell access to those who focus on ransomware attacks. Recently, we have seen collaboration between AL/ML developers who scrape public records to build Org Charts, as well as lists of real estate holdings. This data is then used en masse with situational and location data to populate PDF attachments in emails that look like real invoices, with executives’ names in fake prior email responses, as part of the thread. ... the recent development in hackers organizing into larger groups has allowed the stakes to get even higher. Look at the Lazarus Group, who pulled off one of the largest heists ever by targeting Bybit and stealing $1.5 billion in Ethereum, as well as subsequently converting $300 million in unrecoverable funds. This group is likely state-sponsored and funding North Korean military programs. Therefore, understanding North Korean national interests will hint at future targets. The increasing scale of their attacks likely reflects greater resources allocated by North Korea, more sophisticated tooling and capabilities, lessons learned from previous operations, and a growing number of personnel trained in cyber operations.

Agentic AI might soon get into cryptocurrency trading — what could possibly go wrong?

Not everyone is bullish on the intersection of Web3, agentic AI and blockchain. Forrester Research vice president and principal analyst Martha Bennett is among those who are skeptical. In 2023, she co-authored an online post critical of Worldcoin, now the World project, and her opinion hasn’t changed in several regards. World project still faces major challenges, including privacy issues and concerns about its iris biometric technology, she said. And Agentic AI is still in its early stages and not yet capable of supporting Web3 transactions. Most current generative AI (genAI) tools, including LLMs, lack the autonomy defined as “agentic AI.” “There’s no AI technology today that would be able automate Web3 transactions in a reliable and secure manner,” she said. Given the risks and the potential for exploitation, it’s too soon to rely on AI systems with high autonomy for Web3 transactions. She did note, however, that Web3 already uses automation through smart contracts — self-executing electronic contracts with the terms of the agreement directly written into code. “Will Web3 go mainstream in 2025? My overall answer is no, but there are nuances,” she said. “If mainstream means mass consumer adoption, it’s a definite no. There’s simply not enough utility there for consumers.” Web3, Bennett said, is largely a self-contained financial ecosystem, and efforts to boost adoption through Decentralized Physical Infrastructure Networks (DePIN), such as Tools for Humanity’s, haven’t led to major breakthroughs.

Artificial Intelligence fuels rise of hard-to-detect bots

“The surge in AI-driven bot creation has serious implications for businesses worldwide,” said Tim Chang, General Manager of Application Security at Thales. “As automated traffic accounts for more than half of all web activity, organisations face heightened risks from bad bots, which are becoming more prolific every day.” ... “This year’s report sheds light on the evolving tactics and techniques utilised by bot attackers. What were once deemed advanced evasion methods have now become standard practice for many malicious bots,” Chang said. “In this rapidly changing environment, businesses must evolve their strategies. It’s crucial to adopt an adaptive and proactive approach, leveraging sophisticated bot detection tools and comprehensive cybersecurity management solutions to build a resilient defense against the ever-shifting landscape of bot-related threats.” ... Analysis in the report reveals a deliberate strategy by cyber attackers to exploit API endpoints that manage sensitive and high-value data. Implications of this trend are especially impactful for industries that rely on APIs for their critical operations and transactions. Financial services, healthcare, and e-commerce sectors are bearing the brunt of these sophisticated bot attacks, making them prime targets for malicious actors seeking to breach sensitive information.

Humans at the helm of an AI-driven grid

A growing number of utilities are turning to AI-based tools to process vast data streams and streamline tasks once managed by manual calculation. For instance, algorithms can analyse weather patterns, historical consumption, and real-time sensor readings to make more accurate power demand and renewable energy generation forecasts. This supports more efficient balancing of supply and demand, reducing the likelihood of overloaded transformers or unexpected brownouts. Some utilities are also exploring AI-driven alarm management, which can filter the flood of alerts triggered by a network issue. Instead of operators sifting through hundreds of notifications, AI tools can be used to identify and highlight the most critical issues in real time. Another AI application is with congestion management, detecting trouble spots on the grid where demand might exceed capacity and even propose rerouting strategies to keep electricity flowing reliably. While still in their early stages, AI tools hold promise for driving operational efficiency in many daily scenarios. ... Even the smartest algorithm, however, lacks the broader perspective and accountability that people bring to grid management. Power and Utility companies are tasked with a public service mandate: they must ensure safety, affordability, and equitable access to electricity.

CISO Conversations: Maarten Van Horenbeeck, SVP & CSO at Adobe

The digital divide is simple to understand but complex to solve. Fundamentally, it separates those who have access to cyber and cyber knowledge from those who do not. There are areas of the world and socio-economic groups or demographics who have little or very limited access to the internet, and consequently very little awareness of cybersecurity. But cyber and cyber threats are worldwide; and technology is increasingly integrated and interconnected globally. “Cyber issues emanating from the digital divide don’t just play out far away from our homes – they play out very close to our homes as well,” warns Van Horenbeeck. “There’s a huge divide between people who know, for example, not to reuse passwords, to use multi factor authentication, and those individuals that have none of that experience at all.” In effect the digital divide creates a largely invisible and unseen threat surface for the long-connected world. He believes that technology companies can play a part in solving this problem by making cybersecurity features easy to understand and use. and cites two examples of the Adobe approach. “We invested, for example, in support for passkeys because we feel it’s a more effective and easier method of authentication that is also more secure.”

How AI, Robotics and Automation Transform Supply Chains

Enterprises designing robots to augment the human workforce need to take design thinking and ergonomic approaches into consideration. Designers must think about how robots comprehend and understand their physical surroundings without tripping over cables or objects on the floor, obstructing movement or causing human injuries. These robots are created with the aim to collaborate with humans for repetitive tasks and lift heavy loads. Last year, OT.today featured stories on how humanoid robots augmented the human workforce at Amazon, Mercedes, NASA and the Piaggio Group. In 2017, Alibaba invested in AI labs and the DAMO Academy. At its flagship Computing Conference in 2018, held in Hangzhou, China, Alibaba showcased a range of robots designed for warehouses, autonomous deliveries and other sectors, including hospitality and pharmaceuticals. More recently, Alibaba invested in LimX Dynamics, a company specializing in humanoid and robotic technology. Japanese automobile manufacturers have been using industrial robots since the early 1980s. Chip manufacturing companies in Taiwan and other countries also use them. Robots assist in surgeries in the healthcare sector. But none of those early manufacturing robots resembled humanoids or even had advanced AI seen in today's robots.

CIOs are overspending on the cloud — but still think it’s worth it

CIOs should also embrace DevOps practices tied to cost reduction when consuming cloud resources, Sellers says. One pitfall that doesn’t get enough attention: Many organizations don’t educate developers on the cost of cloud services, despite the glut of developer services large cloud providers make trivial to call. “I’ve lost track of how many services Amazon provides that developers can just use, and some of those can be quite expensive, but a developer doesn’t really know that,” Sellers says. “They’re like, ‘Instead of writing my own solution to this, I can just call this service that Amazon already provides, and boom, my job is done.’” The disconnect between developers and financial factors in the cloud is a real problem that leads to increased cloud costs, adds Nick Durkin, field CTO at Harness, provider of an AI-driven software development platform. Without knowing the costs of accessing a cloud-based GPU or CPU, for example, a developer is like a home builder who doesn’t know the cost of wood or brick, Durkin says. “If you’re not giving your smartest engineers access to the information about services that they can optimize on, how would you expect them to do it?” he says. “Then, finance comes back a month later with a beating stick.”

Daily Tech Digest - February 28, 2025

Quote for the day:

“Success is most often achieved by those who don't know that failure is inevitable.” -- Coco Chanel

Microservice Integration Testing a Pain? Try Shadow Testing

Shadow testing is especially useful for microservices with frequent deployments, helping services evolve without breaking dependencies. It validates schema and API changes early, reducing risk before consumer impact. It also assesses performance under real conditions and ensures proper compatibility with third-party services. ... Shadow testing doesn’t replace traditional testing but rather complements it by reducing reliance on fragile integration tests. While unit tests remain essential for validating logic and end-to-end tests catch high-level failures, shadow testing fills the gap of real-world validation without disrupting users. Shadow testing follows a common pattern regardless of environment and has been implemented by tools like Diffy from Twitter/X, which introduced automated-response comparisons to detect discrepancies effectively. ... The environment where shadow testing is performed may vary, providing different benefits. More realistic environments are obviously better:Staging shadow testing — Easier to set up, avoids compliance and data isolation issues, and can use synthetic or anonymized production traffic to validate changes safely. Production shadow testing — Provides the most accurate validation using live traffic but requires safeguards for data handling, compliance and test workload isolation.

The rising threat of shadow AI

Creating an Office of Responsible AI can play a vital role in a governance model. This office should include representatives from IT, security, legal, compliance, and human resources to ensure that all facets of the organization have input in decision-making regarding AI tools. This collaborative approach can help mitigate the risks associated with shadow AI applications. You want to ensure that employees have secure and sanctioned tools. Don’t forbid AI—teach people how to use it safely. Indeed, the “ban all tools” approach never works; it lowers morale, causes turnover, and may even create legal or HR issues. The call to action is clear: Cloud security administrators must proactively address the shadow AI challenge. This involves auditing current AI usage within the organization and continuously monitoring network traffic and data flows for any signs of unauthorized tool deployment. Yes, we’re creating AI cops. However, don’t think they get to run around and point fingers at people or let your cloud providers point fingers at you. This is one of those problems that can only be solved with a proactive education program aimed at making employees more productive and not afraid of getting fired. Shadow AI is yet another buzzword to track, but also it’s undeniably a growing problem for cloud computing security administrators.

Can AI live up to its promise?

The debate about truly transformative AI may not be about whether it can think or be conscious like a human, but rather about its ability to perform complex tasks across different domains autonomously and effectively. It is important to recognize that the value and usefulness of machines does not depend on their ability to exactly replicate human thought and cognitive abilities, but rather on their ability to achieve similar or better results through different methods. Although the human brain has inspired much of the development of contemporary AI, it need not be the definitive model for the design of superior AI. Perhaps by freeing the development of AI from strict neural emulation, researchers can explore novel architectures and approaches that optimize different objectives, constraints, and capabilities, potentially overcoming the limitations of human cognition in certain contexts. ... Some human factors that could be stumbling blocks on the road to transformative AI include: the information overload we receive, the possible misalignment with our human values, the possible negative perception we may be acquiring, the view of AI as our competitor, the excessive dependence on human experience, the possible perception of futility of ethics in AI, the loss of trust, overregulation, diluted efforts in research and application, the idea of human obsolescence, or the possibility of an “AI-cracy”, for example.

The end of net neutrality: A wake-up call for a decentralized internet

We live in a time when the true ideals of a free and open internet are under attack. The most recent repeal of net neutrality regulations is taking us toward a more centralized, controlled version of the internet. In this scenario, a decentralized, permissionless internet offers a powerful alternative to today’s reality. Decentralized systems can address the threat of censorship by distributing content across a network of nodes, ensuring that no single entity can block or suppress information. Decentralized physical infrastructure networks (DePIN) demonstrate how decentralized storage can keep data accessible even when network parts are disrupted or taken offline. This censorship resistance is crucial in regions where governments or corporations try to limit free expression online. Decentralization can also cultivate economic democracy by eliminating intermediaries like ISPs and related fees. Blockchain-based platforms allow smaller, newer players to compete with incumbent services and content companies on a level playing field. The Helium network, for example, uses a decentralized model to challenge traditional telecom monopolies with community-driven wireless infrastructure. In a decentralized system, developers don’t need approval from ISPs to launch new services.

Steering by insights: A C-Suite guide to make data work for everyone

With massive volumes of data to make sense of, having reliable and scalable modern data architectures that can organise and store data in a structured, secure, and governed manner while ensuring data reliability and integrity is critical. This is especially true in the hybrid, multi-cloud environment in which companies operate today. Furthermore, as we face a new “AI summer”, executives are experiencing increased pressure to respond to the tsunami of hype around AI and its promise to enhance efficiency and competitive differentiation. This means companies will need to rely on high-quality, verifiable data to implement AI-powered technologies Generative AI and Large Language Models (LLMs) at an enterprise scale. ... Beyond infrastructure, companies in India need to look at ways to create a culture of data. In today’s digital-first organisations, many businesses require real-time analytics to operate efficiently. To enable this, organisations need to create data platforms that are easy to use and equipped with the latest tools and controls so that employees at every level can get their hands on the right data to unlock productivity, saving them valuable time for other strategic priorities. Building a data culture also needs to come from the top; it is imperative to ensure that data is valued and used strategically and consistently to drive decision-making.

The Hidden Cost of Compliance: When Regulations Weaken Security

What might be a bit surprising, however, is one particular pain point that customers in this vertical bring up repeatedly. What is this mysterious pain point? I’m not sure if it has an official name or not, but many people I meet with share with me that they are spending so much time responding to regulatory findings that they hardly have time for anything else. This is troubling to say the least. It may be an uncomfortable discussion to have, but I’d argue that it is long since past the time we as a security community have this discussion. ... The threats enterprises face change and evolve quickly – even rapidly I might say. Regulations often have trouble keeping up with the pace of that change. This means that enterprises are often forced to solve last year’s or even last decade’s problems, rather than the problems that might actually pose a far greater threat to the enterprise. In my opinion, regulatory agencies need to move more quickly to keep pace with the changing threat landscape. ... Regulations are often produced by large, bureaucratic bodies that do not move particularly quickly. This means that if some part of the regulation is ineffective, overly burdensome, impractical, or otherwise needs adjusting, it may take some time before this change happens. In the interim, enterprises have no choice but to comply with something that the regulatory body has already acknowledged needs adjusting.

Why the future of privileged access must include IoT – securing the unseen

The application of PAM to IoT devices brings unique complexities. The vast variety of IoT devices, many of which have been operational for years, often lack built-in security, user interfaces, or associated users. Unlike traditional identity management, which revolves around human credentials, IoT devices rely on keys and certificates, with each device undergoing a complex identity lifecycle over its operational lifespan. Managing these identities across thousands of devices is a resource-intensive task, exacerbated by constrained IT budgets and staff shortages. ... Implementing a PAM solution for IoT involves several steps. Before anything else, organisations need to achieve visibility of their network. Many currently lack this crucial insight, making it difficult to identify vulnerabilities or manage device access effectively. Once this visibility is achieved, organisations must then identify and secure high-risk privileged accounts to prevent them from becoming entry points for attackers. Automated credential management is essential to replace manual password processes, ensuring consistency and reducing oversight. Policies must be enforced to authorise access based on pre-defined rules, guaranteeing secure connections from the outset. Default credentials – a common exploit for attackers – should be updated regularly, and automation can handle this efficiently.

Understanding the AI Act and its compliance challenges

There is a clear tension between the transparency obligations imposed on providers of certain AI systems under the AI Act and some of their rights and business interests, such as the protection of trade secrets and intellectual property. The EU legislator has expressly recognized this tension, as multiple provisions of the AI Act state that transparency obligations are without prejudice to intellectual property rights. For example, Article 53 of the AI Act, which requires providers of general-purpose AI models to provide certain information to organizations that wish to integrate the model downstream, explicitly calls out the need to observe and protect intellectual property rights and confidential business information or trade secrets. In practice, a good faith effort from all parties will be required to find the appropriate balance between the need for transparency to ensure safe, reliable and trustworthy AI, while protecting the interests of providers that invest significant resources in AI development. ... The AI Act imposes a number of obligations on AI system vendors that will help in-house lawyers in carrying out this diligence. Under Article 13 of the AI Act, vendors of high-risk AI systems are, for example, required to provide sufficient information to (business) deployers to allow them to understand the high-risk AI system’s operation and interpret its output.

Why fast-learning robots are wearing Meta glasses

The technology acts as a sophisticated translator between human and robotic movement. Using mathematical techniques called Gaussian normalization, the system maps the rotations of a human wrist to the precise joint angles of a robot arm, ensuring natural motions get converted into mechanical actions without dangerous exaggerations. This movement translation works alongside a shared visual understanding — both the human demonstrator’s smartglasses and the robot’s cameras feed into the same artificial intelligence program, creating common ground for interpreting objects and environments. ... The EgoMimic researchers didn’t invent the concept of using consumer electronics to train robots. One pioneer in the field, a former healthcare-robot researcher named Dr. Sarah Zhang, has demonstrated 40% improvements in the speed of training healthcare robots using smartphones and digital cameras; they enable nurses to teach robots through gestures, voice commands, and real-time demonstrations instead of complicated programming. This improved robot training is made possible by AI that can learn from fewer examples. A nurse might show a robot how to deliver medications twice, and the robot generalizes the task to handle variations like avoiding obstacles or adjusting schedules.

Targeted by Ransomware, Middle East Banks Shore Up Security

The financial services industry in UAE — and the Middle East at large — sees cyber wargaming as an important way to identify weaknesses and develop defenses to the latest threats, Jamal Saleh, director general of the UAE Banks Federation, said in a statement announcing the completion of the event. "The rapid adoption and deployment of advanced technologies in the banking and financial sector have increased risks related to transaction security and digital infrastructure," he said in the statement, adding that the sector is increasingly aware "of the importance of such initiatives to enhance cybersecurity systems and ensure a secure and advanced environment for customers, especially with the rapid developments in modern technology and the rise of cybersecurity threats using advanced artificial intelligence (AI) techniques." ... Ransomware remains a major threat to the financial industry, but attackers have shifted from distributed denial-of-service (DDoS) attacks to phishing, data breaches, and identity-focused attacks, according to Shilpi Handa, associate research director for the Middle East, Turkey, and Africa at business intelligence firm IDC. "We see trends such as increased investment in identity and data security, the adoption of integrated security platforms, and a focus on operational technology security in the finance sector," she says.

Daily Tech Digest - February 16, 2025

Quote for the day:

"Leaders should influence others in such a way that it builds people up, encourages and edifies them so they can duplicate this attitude in others." -- Bob Goshen

A look under the hood of transfomers, the engine driving AI model evolution

Depending on the application, a transformer model follows an encoder-decoder architecture. The encoder component learns a vector representation of data that can then be used for downstream tasks like classification and sentiment analysis. The decoder component takes a vector or latent representation of the text or image and uses it to generate new text, making it useful for tasks like sentence completion and summarization. For this reason, many familiar state-of-the-art models, such the GPT family, are decoder only. Encoder-decoder models combine both components, making them useful for translation and other sequence-to-sequence tasks. For both encoder and decoder architectures, the core component is the attention layer, as this is what allows a model to retain context from words that appear much earlier in the text. ... Currently, transformers are the dominant architecture for many use cases that require LLMs and benefit from the most research and development. Although this does not seem likely to change anytime soon, one different class of model that has gained interest recently is state-space models (SSMs) such as Mamba. This highly efficient algorithm can handle very long sequences of data, whereas transformers are limited by a context window.

McKinsey On Return To Office: Leaders Are Focused On The Wrong Thing

Unsurprisingly, older employees report higher satisfaction with on-site work than their younger colleagues. Nevertheless, employees across all work models report similar satisfaction levels, which debunks the belief that bringing people back in person automatically enhances engagement or retention. Worse still, leaders consistently overestimate their organizations’ maturity regarding the very factors used to justify returning to the office. ... The balance of power may have shifted back to bosses, but, as Voltaire said first and Spider-Man famously learns from Uncle Ben, “with great power comes great responsibility.” No matter what workplace model a given employee finds themselves in today, the past few years likely opened their eyes to the power of choice and flexibility and the chasm between modern hospitality and retail-oriented experiences and the vibrancy and community in a traditional office. ... So employees believe they are doing the work, and they may accept that flexibility is a reward for objectively high performance. If executives believe the purpose of the office is to accelerate innovation, connectivity, and mentoring, they are on the hook to ensure it does. Leaders must model new behaviors, invest in workplace experience, and learn to measure outcomes without a bias for presence. Employees may quit as soon as the power pendulum swings back.

8 tips for being a more decisive leader

“Clarity is what is expected from a leader,” says Malhotra. “Clarity of vision, clarity in strategy, clarity of plan, clarity in the process, and clarity in how to measure success.” Showing up with an answer is not as important to the decision as bringing clarity to the process. “As a leader, you’re the force multiplier for your organization,” he says. “Force multiplying is a vector quantity, not a scalar quantity. It’s a vector quantity because the direction is very important. It’s not just the magnitude. It’s the direction, too. So being a force multiplier requires that you are clear when it comes to the end state you are trying to achieve.” ... “There are two things you have to consider: the urgency and the importance of the decision,” says Efrain Ruh, field CTO for Continental Europe at Digitate. If something is complex and important, take your time and gather as much information as possible. But if it is a decision that is easy to come back from, he says, “I try not to go too deep.” “There are ‘single-door decisions’ and ‘double-door decisions,’” agrees Malhotra. When it’s a single-door decision, you can never come back through that door after you have walked through it. ... When you step into a leadership role, you begin to see everything from a high-level strategy point of view. But your decisions will often affect people with their boots on the ground.

Can English Dethrone Python as Top Programming Language?

IDC predicts that by 2028, natural language will become the most widely used programming language, with developers using it to create 70% of net-new digital solutions. (Source: IDC FutureScape: Worldwide Developer and DevOps 2025 Predictions) “I actually think that the best phrasing of this prediction would be to replace ‘natural language’ with ‘English’ because of the dominance of English as a spoken and written language worldwide,” Dayaratna said. Moreover, he said he believes that in four to five years, developers will increasingly go to a chatbot-like interface and use natural language to produce digital solutions. Meanwhile, code will be used to innovate on the technology substrate that enables this kind of technology. “In other words, we’re not far from a world that witnesses the demise of commercial off-the-shelf software simply because it will be so easy to create such software, in a custom way, for an organization’s business processes,” Dayaratna said. Hence, he explained that we are seeing the emergence of what Amjad Masad, CEO of Replit, called the era of “personal software.” “Just as the Mac inaugurated personal computing in 1984, generative AI has initiated the era of ‘personal software’ that recognizes the specificity of individual and organizational preferences,” Dayaratna said.

What is anomaly detection? Behavior-based analysis for cyber threats

“Anomaly detection is the holy grail of cyber detection where, if you do it right, you don’t need to know a priori the bad thing that you’re looking for,” Bruce Potter, CEO and founder of Turngate, tells CSO. “It’ll just show up because it doesn’t look like anything else or doesn’t look like it’s supposed to. People have been tilting at that windmill for a long time, since the 1980s, trying to figure out what normal is so they can look for deviations from it to find all the bad things happening in their enterprises.” ... Although predicated on advanced math concepts, anomaly detection, or as the NIST Cybersecurity Framework 2.0 calls it, “adverse event analysis,” has over the past two decades been incorporated into a wide range of cybersecurity tools, including endpoint detection and response (EDR), firewall, and security information and event management (SIEM) tools. “In general, you can split the detection universe into two halves,” Potter says. “One is finding known bads, and then one is finding things that might be bad. Known bads are typically like a signature base where I know very specifically if I see this file or this exact thing happened on the system, it’s bad.” Known bads are typically flagged by fundamental cybersecurity tools.

Open Source AI Models: Perfect Storm for Malicious Code, Vulnerabilities

Executable data files are not the only threats, however. Licensing is another issue: While pretrained AI models are frequently called "open source AI," they generally do not provide all the information needed to reproduce the AI model, such as code and training data. Instead, they provide the weights generated by the training and are covered by licenses that are not always open source compatible. Creating commercial products or services from such models can potentially result in violating the licenses, says Andrew Stiefel, a senior product manager at Endor Labs. "There's a lot of complexity in the licenses for models," he says. "You have the actual model binary itself, the weights, the training data, all of those could have different licenses, and you need to understand what that means for your business." Model alignment — how well its output aligns with the developers' and users' values — is the final wildcard. DeepSeek, for example, allows users to create malware and viruses, researchers found. Other models — such as OpenAI's o3-mini model, which boasts more stringent alignment — has already been jail broken by researchers. These problems are unique to AI systems and the boundaries of how to test for such weaknesses remains a fertile field for researchers, says ReversingLabs' Pericin.

Risk Matters: Cyber Risk and AI – The Changing Landscape

Although AI assists organizations defend against cyber-attacks, it is a double-edged sword. More to the point, AI is also providing cyber attackers with an array of cost-efficient techniques that facilitate their cyber-attacks. Sophisticated AI-generated phishing attacks, social engineering attacks, and ransomware attacks are just a few of the ways AI has made the cyber-attack landscape more lethal. AI-generated models used by cyber attackers and cyber defenders have been evolving at a rapid pace. As a result, the strategic interactions between cyber attackers and cyber defenders have become more automated, more dynamic, more adaptive, and more complex. These developments have increased, and substantially changed, the game-theoretic aspects associated with cyber risk. ... Besides considering the total amount to spend on cybersecurity-related activities, a subsidiary question for organizations to answer is: How much of our organization’s cybersecurity-related budget should be devoted to developing and implementing AI models designed to reduce the likelihood of a cyber incident? In answering this subsidiary question, organizations need to consider the costs associated with the AI models.

Juniper CEO: ‘I am disappointed and somewhat puzzled’ by DOJ merger rejection

“They’re taking such a narrow view of the total transaction, which is the wireless line segment, a relatively small part of Juniper’s business, a small part of HPE’s business. And even if you do take a look at the wireless segment, you know we’re talking about a very competitive area with eight or nine different competitors. It’s unfortunate that we’re in the situation that we’re in, but that said, that’s okay. We’re prepared to take it to court and to prove our case and ultimately, hopefully, prevail,” Rahim said. HPE and Juniper met with the DOJ several times to go over the purchase, but the companies had no inclination the DOJ would go the direction it did—certainly with regards to its focus on the wireless market, Rahim said. The DOJ issued a Complaint “that ignores the reality that HPE and Juniper are two of at least ten competitors with comparable offerings and capabilities fighting to win customers every day,” the companies wrote. “A Complaint whose description of competitive dynamics in the wireless local area networking (WLAN) space is divorced from reality; and a Complaint that contradicts the conclusions reached by antitrust regulators around the world that have unconditionally cleared the transaction.”

The Benefits of the M&A Frenzy in Fraud Solutions

With businesses looking to reduce the number of vendors they work with to lower integration costs, David Mattei, strategic advisor at Datos Insights, expects "a higher momentum of M&A activities in 2025 as vendors race to grow." "Single-solution vendors have a harder time competing in today's world," and small to medium-sized single solution vendors "are likely to be acquired," Mattei said. LexisNexis' acquisition of IDVerse in December 2024 is an example of this this trend. ... Fraud executives agree that the most pragmatic approach today is proactive communication and awareness campaigns, and the data supports their effectiveness. However, the most anticipated and potentially effective solution is consortia-based fraud detection, combining risk signals from both sending and receiving financial institutions, Fooshee told Information Security Media Group. The challenge lies in overcoming resistance to information sharing - from fraud teams, compliance, legal and regulators - because of concerns over data integrity, integration complexities and privacy restrictions. Interestingly, markets most affected by scams and with simpler regulatory landscapes are finding ways to navigate these barriers more effectively.

Apple’s emotional lamp and the future of robots

It’s clear that Apple’s lamp is programmed to move in a way that deludes users into believing that the it has internal states that it doesn’t actually have. ... Apple’s lamp research definitely sheds light on where our interaction with robots may be heading—a new category of appliance that might well be called the “emotional robot.” A key component of the research was a user study comparing how people perceived a robot using functional and expressive movements versus one that uses only functional movements. ... The biggest takeaway from Apple’s ELEGNT research is likely that neither a human-like voice nor a human-like body, head, or face is required for a robot to successfully trick a human into relating to it as a sentient being with internal thoughts, feelings, and emotions. ELEGNT is not a prototype product; it is instead a lab and social experiment. But that doesn’t mean a product based on this research will not soon be available on a desktop near you. ... Apple is developing a desktop robot project, codenamed J595, and is targeting a launch within two years. According to reports based on leaks, the robot might look a little like Apple’s iMac G4, which was a lamp-like form factor featuring a screen at the end of a moveable “arm.”