Every week, someone somewhere slips upand threat actors slip in. A misconfigured setting, an overlooked vulnerability, or a too-convenient cloud tool becomes the perfect entry point. But what happens when the hunters become the hunted? Or when old malware resurfaces with new tricks?Step behind the curtain with us this week as we explore breaches born from routine oversightsand the unexpected cracks they reveal in systems we trust. Threat of the WeekGoogle Patches Actively Exploited Chrome 0-Day Google has addressed a high-severity security flaw in its Chrome browser for Windows that has been exploited by unknown actors as part of a sophisticated attack aimed at Russian entities. The flaw, CVE-2025-2783 (CVSS score: 8.3), is said to have been combined with another exploit to break out of the browser's sandbox and achieve remote code execution. The attacks involved distributing specially crafted links via phishing emails that, when clicked and launched using Chrome, triggered the exploit. A similar flaw has since been patched in Mozilla Firefox and Tor Browser (CVE-2025-2857), although there is no evidence that it has been exploited.Download Now Top NewsCritical Flaws Uncovered in Ingress NGINX Controller for Kubernetes A set of vulnerabilities, collectively named IngressNightmare, has been disclosed in the Ingress NGINX Controller for Kubernetes that could result in unauthenticated remote code execution. The most severe of the five flaws is CVE-2025-1974 (CVSS score: 9.8), which an unauthenticated attacker with access to the pod network could exploit to achieve arbitrary code execution in the context of the ingress-nginx controller under certain conditions. Following responsible disclosure, the vulnerabilities have been addressed in Ingress NGINX Controller versions 1.12.1, 1.11.5, and 1.10.7.BlackLock Data Leak Site Exposed Threat hunters have managed to infiltrate the data leak site associated with a ransomware group called BlackLock, uncovering crucial information about their modus operandi in the process. Thanks to a local file inclusion (LFI) vulnerability, cybersecurity company Resecurity said it was able to extract configuration files, credentials, as well as the history of commands executed on the server. The threat actors have been found using Rclone to exfiltrate data to the MEGA cloud storage service. As many as eight accounts have been created on MEGA to store and backup victim data. The development comes as KELA revealed the possible real-world identities of Rey and Pryx, the key players driving the Hellcat ransomware operations. Rey (aka Saif and Hikki-Chan) is likely of Palestinian and Jordanian origin, while Pryx (aka Adem) is said to be an Arabic speaker involved in carding since 2018. "Ironically, Rey and Pryx, who heavily relied on info stealer logs in their operations, fell victim to it themselves," KELA said.46 Flaws in Solar Inverters From Sungrow, Growatt, and SMA As many as 46 security bugs have discovered in products from three solar inverter vendors, Sungrow, Growatt, and SMA that, if successfully exploited, could permit attackers to seize control of devices and cause potential power blackouts. The vulnerabilities, collectively named SUN:DOWN, "can be exploited to execute arbitrary commands on devices or the vendor's cloud, take over accounts, gain a foothold in the vendor's infrastructure, or take control of inverter owners' devices."RedCurl Linked to First Case of Ransomware RedCurl, a threat actor known for its corporate espionage attacks since late 2018, has been observed delivering a custom ransomware family called QWCrypt via a sophisticated multi-stage infection chain. Bitdefender, which flagged the activity, said the "unusual deviation" in tactics raises more questions than answers about their motivations, raising the possibility that it may be either a cyber mercenary group or it's a discreet operation designed to generate consistent revenue.Hackers Using Atlantis AIO for Credential Stuffing and Brute-Force Attacks Threat actors are making use of an e-crime tool called Atlantis AIO Multi-Checker to automate credential stuffing attacks across more than 140 platforms, allowing them to test millions of stolen credentials in "rapid succession." The software also comes with capabilities to conduct brute-force attacks against email platforms and automate account recovery processes associated with eBay and Yahoo.Weaver Ant Goes Undetected for Over 4 Years A suspected Chinese state-backed hacking group called Weaver Ant managed to stay under the radar after it breached a major telecommunications company located in Asia. The attack involved the exploitation of a misconfiguration in a public-facing application to gain initial access and drop web shells for persistent remote access. The web shells were then used to drop additional payloads to facilitate lateral movement and carry out reconnaissance activities. Over the past year, Chinese hacking crews have also targeted a trade group in the United States and a research institute in Mexico to deliver ShadowPad and two new variants of a backdoor known as SparrowDoor. The activity has been attributed to a threat actor tracked as FamousSparrow.Morphing Meerkat Uses DNS MX and DoH to Distribute Spam A newly discovered phishing-as-a-service (PhaaS) operation called Morphing Meerkat has been leveraging the Domain Name System (DNS) mail exchange (MX) records to determine the victim's email service provider and dynamically serve fake login pages that impersonate about 114 brands. The platform also makes use of the DNS-over-HTTPS (DoH) protocol to evade detection when firing a DNS query to Google or Cloudflare to find the MX records of the victim's email domain. The credentials captured on the spoofed pages are then exfiltrated via Telegram or AJAX requests to external servers. Morphing Meerkat is known to have been active since at least 2020. It features a centralized SMTP infrastructure to distribute thousands of spam emails, with 50% of the traced emails originating from internet services provided by iomart and HostPapa. Trending CVEsAttackers love software vulnerabilitiesthey're easy doors into your systems. Every week brings fresh flaws, and waiting too long to patch can turn a minor oversight into a major breach. Below are this week's critical vulnerabilities you need to know about. Take a look, update your software promptly, and keep attackers locked out.This week's list includes CVE-2025-2783, CVE-2025-2476 (Google Chrome), CVE-2025-2857 (Mozilla Firefox, Tor Browser), CVE-2025-1974 (Kubernetes NGINX Ingress Controller), CVE-2025-26512 (NetApp SnapCenter), CVE-2025-22230 (VMware Tools for Windows), CVE-2025-2825 (CrushFTP), CVE-2025-20229 (Splunk), CVE-2025-30232 (Exim), CVE-2025-1716, CVE-2025-1889, CVE-2025-1944, CVE-2025-1945 (picklescan), and CVE-2025-2294 (Kubio AI Page Builder plugin). Around the Cyber World23andMe Files for Bankruptcy Genetic testing business 23andMe filed for Chapter 11 bankruptcy, amplifying concerns that the DNA records and personal information of its 15 million customers could soon be up for sale. "Any buyer will be required to comply with applicable law with respect to the treatment of customer data," the company said in an FAQ. The development has prompted California Attorney General Rob Bonta to issue a privacy consumer alert, detailing the steps users can take to delete their genetic data and destroy their samples. The U.K. Information Commissioner's Office said it's "monitoring the situation closely." While 23andMe notes that genetic data is anonymized and stored separately from personally identifiable information, its privacy policy states the company will retain users' genetic information, date of birth, and sex as required for compliance with applicable legal obligations. In October 2023, it suffered a major data breach, exposing the genetic information of more than six million people. Konni Uses AsyncRAT in New Campaign The North Korea-linked Konni threat actor has been observed using Windows shortcut (LNK) files that masquerade as PDF files to trigger a multi-stage infection sequence that involves using legitimate cloud services like Dropbox and Google Drive to host intermediate payloads that pave the way for the download and deployment of AsyncRAT. The hacking group gets its name from the use of an eponymous RAT called Konni RAT, which offers data exfiltration, command execution, and persistence capabilities. "The final execution of AsyncRAT has been changed to operate by receiving C&C server information as an execution argument," Enki said. "This is more flexible than the previous method of hard-coding C&C server information into malicious code, and anyone can take advantage of malicious code by building a separate server."FBI Warns of Fake File Converters Used to Push Malware Malware peddlers are targeting users who are searching for free file converter services and tools that give them access to the victims' machines. "These converters and downloading tools will do the task advertised, but the resulting file can contain hidden malware giving criminals access to the victim's computer," the U.S. Federal Bureau of Investigation (FBI) said. The tools can also scrape the submitted files for any sensitive information, including credentials and financial details.New SvcStealer Information Stealer Emerges in the Wild A new information stealer called SvcStealer, written in Microsoft Visual C++, has been detected in the wild spreading via phishing campaigns. This malware harvests sensitive data such as system metadata, files matching certain extensions, running processes, installed software, and user credentials, as well as information from cryptocurrency wallets, messaging applications, and web browsers.Meta Begins AI Rollout in Europe But With Limitations Meta has announced that its AI-powered virtual assistant, Meta AI, is finally launching across Facebook, Instagram, WhatsApp, and Messenger in the European Union and United Kingdom over the coming weeks. "It's taken longer than we would have liked to get our AI technology into the hands of people in Europe as we continue to navigate its complex regulatory system," the company said. The European launch follows regulatory and privacy pushback about tapping user data to train AI models. Meta's approach to seeking user consent has come under scrutiny by the Irish Data Protection Commission (DPC), the company's lead data protection regulator in the bloc, forcing the company to halt processing local users' information to train AI models. "The model powering these Meta AI features wasn't trained on first-party data from users in the E.U.," Meta told TechCrunch.INDOHAXSEC Linked to DDoS and Ransomware Attacks An Indonesian-based hacktivist collective dubbed INDOHAXSEC has been linked to a string of distributed denial-of-service (DDoS) and ransomware attacks against numerous entities and governmental bodies located in Australia, India, Israel, and Malaysia using a mix of custom and publicly available tools. The group, which maintains GitHub, Telegram, and social media accounts, emerged in October 2024. It has since announced partnerships with other hacktivist groups like NoName057(16). The ransomware attacks have been found to use a locker called ExorLock, which has been assessed to be written by an earlier iteration of the group when they were active under the name AnonBlackFlag.Orion Framework Paves the Way for Privacy-Preserving AI Models A group of academic researchers from New York University has detailed Orion, a framework that brings support for fully homomorphic encryption (FHE) to deep learning, thereby allowing AI models to practically and efficiently operate directly on encrypted data without needing to decrypt it first. Orion "converts deep learning models written in PyTorch into efficient FHE programs," the team said. "The framework also streamlines encryption-related processes, making it easier to manage accumulated noise and execute deep learning computations efficiently."U.S. Court Upholds Conviction of Joseph Sullivan The U.S. Court of Appeals for the Ninth Circuit unanimously upheld the conviction of former Uber Chief Security Officer Joseph Sullivan, who was previously held liable for failing to disclose a 2016 breach of customer and driver records to regulators and attempting to cover up the incident. The court said the verdict "underscores the importance of transparency even in failure situations especially when such failures are the subject of federal investigation."Russia Arrests 3 People Tied Mamont Malware Russian authorities have arrested three individuals suspected of developing an Android malware known as Mamont. The suspects, whose names were not disclosed, were apprehended from the Saratov region, The Record reported. Earlier this January, the Ministry of Internal Affairs of Russia revealed that the malware was being propagated in the form of APK files via Telegram with the ultimate aim of stealing sensitive personal and financial information from victims' devices. Russian cybersecurity company Kaspersky said it also discovered threat actors using novel social engineering tactics to distribute the banking trojan targeting Android devices in the country.2 Serbian Journalists Targeted by NSO Group's Pegasus Two investigative journalists in Serbia, who work for the Balkan Investigative Reporting Network (BIRN), were targeted with Pegasus, a commercial spyware developed by NSO Group. The two journalists received last month suspicious messages on the Viber messaging app from an unknown Serbian number linked to Telekom Srbija, the state-telecommunications operator, Amnesty International said. The messages contained a link that, if clicked, would have led to the deployment of the information-gathering tool via a decoy site. Both the journalists did not click on the link. The development marks the third time Pegasus has been used against civil society in Serbia in two years. Serbian authorities have also recently used Cellebrite software to secretly unlock civilians' phones so they could install another brand of homegrown spyware codenamed NoviSpy.IOCONTROL Found Listed for Sale The Iran-linked malware called IOCONTROL, which is explicitly designed to target industrial environments, has been listed for sale on Telegram and BreachForums, per Flashpoint. The malware is attributed to a hacking group called Cyber Av3ngers. Also called OrpaCrab, the sophisticated Linux-based backdoor is capable of surveillance, lateral movement, data exfiltration, system manipulation, and remote control.U.K. Issues Warning About Sadistic Online Harm Groups The U.K. National Crime Agency (NCA) has warned of a "deeply concerning" trend of online networks called The Com that have resorted to inflicting harm and committing various kinds of criminal acts. "These online forums or communities [...] see offenders collaborate or compete to cause harm across a broad spectrum of criminality both on and offline including cyber, fraud, extremism, serious violence, and child sexual abuse," the NCA said. Part of this cybercrime ecosystem is the infamous Scattered Spider group, which is known for its advanced social engineering techniques to conduct extortion and ransomware attacks. Last month, Richard Ehiemere, 21, an East London member of the network, was convicted on charges of fraud and making indecent images of children. Part of a group called CVLT, the accused and other members are said to target girls on social media platforms such as Discord and persuade them to send intimate photos of themselves. "Members threatened to 'dox' their victims, which involves revealing real-world identities and publishing other personal information online, in order to coerce them into complying with their demands," the NCA said. "Girls were forced to join group calls, where they would be instructed to carry out sexual acts and acts of self-harm for their audience. In severe cases, vulnerable victims were encouraged to kill themselves on camera." A month prior to that, 19-year-old Cameron Finnigan was jailed for encouraging suicide, possession of indecent images of children, and two counts of criminal damage.Unknown Threat Actor Registers Over 10k Domains for Smishing Scams Over 10,000 domains bearing the same domain pattern have been registered for conducting various kinds of SMS phishing scams. "The root domain names all begin with the string: com-," Palo Alto Networks Unit 42 said. "Since the root domain begins with "com-" next to a subdomain, the full domain might trick potential victims into doing a casual inspection." The campaigns are designed to trick users into revealing their personal information, including credit or debit card and account information.Exploiting Car Infotainment System to Plant Spyware NCC Group researchers Alex Plaskett and McCaulay Hudson have demonstrated a trio of zero-day exploits (CVE-2024-23928, CVE-2024-23929, and CVE-2024-23930) that could be weaponized to break into Pioneer DMH-WT7600NEX, gain shell access, and install malicious software on the in-vehicle infotainment (IVI) system. This could then be used to exfiltrate data from the infotainment system to track an individual's location, contacts, and call history. Previously, the duo revealed multiple vulnerabilities in Phoenix Contact CHARX SEC-3100, an electric vehicle (EV) charger controller, that could facilitate privilege escalation and remote code execution (CVE-2024-6788, CVE-2024-25994, CVE-2024-25995, and CVE-2024-25999). Expert WebinarIs ASPM the future of AppSecor just another trend? Join Amir Kaushansky from Palo Alto Networks to find out. In this free webinar, you'll learn how Application Security Posture Management (ASPM) helps teams fix security gaps by connecting code and runtime data. See how it brings all your AppSec tools into one place, so you can spot real risks faster, automate policies, and reduce the need for last-minute fixes. If you want to simplify security and stay ahead of threats, this session is for you. Save your seat now.AI Is Fueling AttacksLearn How to Shut Them Down AI isn't the future threatit's today's biggest challenge. From deepfake phishing to AI-powered reconnaissance, attackers are moving faster than legacy defenses can keep up. In this session, Zscaler's Diana Shtil shares practical ways to use Zero Trust to defend against AI-driven threatsbefore they reach your perimeter.AI Tools Are Bypassing Your ControlsHere's How to Find and Stop Them You can't protect what you can't see. Shadow AI tools are quietly spreading across SaaS environmentsoften unnoticed until it's too late. Join Reco's Dvir Sasson for a real-world look at hidden AI usage, stealthy attack paths, and how to get visibility before threats become incidents. Cybersecurity ToolsNetBird NetBird makes it easy to build secure private networks without complex setups. It connects your devices using WireGuard, with encrypted tunnels and no need to open ports or configure firewalls. Use it at home or work, in the cloud, or self-hosted. Manage access from one place with easy-to-use controls. Fast to install, simple to scale, and works anywhere.Dalfox It is a fast, flexible open-source tool built for modern XSS testing. Designed with automation at its core, it streamlines everything from parameter analysis to vulnerability verificationmaking it a favorite for security researchers and bug bounty hunters. With support for multiple scanning modes, advanced discovery techniques, and customizable payloads, Dalfox offers deep insights into reflected, stored, and DOM-based XSS vulnerabilitiesall while providing detailed, developer-friendly output. Tip of the WeekDisable Browser Autofill for Sensitive Fields Autofill might save time, but it can silently leak your data. Attackers can craft hidden form fields on malicious websites that your browser unknowingly fills with your email, phone number, or even credit card infowithout you ever clicking a thing. It's a quiet but real threat, especially in phishing attacks.To stay safer, disable autofill for personal and sensitive fields in your browser settings. In Chrome, go to Settings Autofill, and turn off Passwords, Payment methods, and Addresses. In Firefox, head to Settings Privacy & Security, and uncheck all Forms and Autofill options. For Edge, go to Profiles Personal Info & Payment Info, and switch off both. On Safari, navigate to Preferences AutoFill and deselect every category.For even more control, use a password manager like Bitwarden or KeePassXCthey only autofill when you explicitly approve it. Convenience is great, but not at the cost of silent data leaks.ConclusionWe often place trust in tools, platforms, and routinesuntil they become the very weapons used against us.This week's stories are a reminder that threat actors don't break the rulesthey bend the conveniences we rely on. It's not just about patching systems; it's about questioning assumptions.Found this article interesting? Follow us on Twitter and LinkedIn to read more exclusive content we post.

Mar 31, 2025The Hacker NewsIntrusion Detection / VulnerabilityIf you're using AWS, it's easy to assume your cloud security is handled - but that's a dangerous misconception. AWS secures its own infrastructure, but security within a cloud environment remains the customer's responsibility.Think of AWS security like protecting a building: AWS provides strong walls and a solid roof, but it's up to the customer to handle the locks, install the alarm systems, and ensure valuables aren't left exposed.In this blog, we'll clarify what AWS doesn't secure, highlight real-world vulnerabilities, and how cloud security scanners like Intruder can help.Understanding the AWS Shared Responsibility ModelAWS operates on a Shared Responsibility Model. In simple terms:AWS is responsible for securing the underlying infrastructure (e.g., hardware, networking, data centers) - the "walls and roof."The customer is responsible for securing their data, applications, and configurations within AWS - the "locks and alarms."Understanding this distinction is essential for maintaining a secure AWS environment.5 Real-World AWS Vulnerabilities You Need to AddressLet's look at some real-world vulnerabilities that fall under the customer's responsibility and what can be done to mitigate them.Server-Side Request Forgery (SSRF)Applications hosted in AWS are still vulnerable to attacks like SSRF, where attackers trick a server into making requests on their behalf. These attacks can result in unauthorized data access and further exploitation.To defend against SSRF:Regularly scan and fix vulnerabilities in applications.Enable AWS IMDSv2, which provides an additional security layer against SSRF attacks. AWS provides this safeguard, but configuration is the customer's responsibility.Access Control WeaknessesAWS Identify and Access Management (IAM) allows customers to manage who can access what resources - but it's only as strong as its implementation. Customers are responsible for ensuring users and systems only have access to the resources they truly need.Common missteps include:Overly permissive roles and accessMissing security controlsAccidentally public S3 bucketsData ExposuresAWS customers are responsible for the security of the data they store in the cloud - and for how their applications access that data.For example, if your application connects to an AWS Relational Database Service (RDS), the customer must ensure that the application doesn't expose sensitive data to attackers. A simple vulnerability like an Insecure Direct Object Reference (IDOR) is all it would take for an attacker with a user account to access data belonging to all other users.Patch ManagementIt almost goes without saying, but AWS does not patch servers! Customers who deploy EC2 instances are fully responsible for keeping the operating system (OS) and software up to date.Take Redis deployed on Ubuntu 24.04 as an example - the customer is responsible for patching vulnerabilities in both the software (Redis) and the OS (Ubuntu). AWS only manages underlying hardware vulnerabilities, like firmware issues.AWS services like Lambda reduce some patching responsibilities, but you're still responsible for using supported runtimes and keeping things up to date.Firewalls and Attack SurfaceAWS gives customers control over their attack surface, but isn't responsible for what they choose to expose.For instance, if a GitLab server is deployed on AWS, the customer is responsible for layering it behind a VPN, using a firewall, or placing it inside a Virtual Private Cloud (VPC) while ensuring their team has a secure way to access it. Otherwise, a zero-day vulnerability could leave your data compromised, and AWS won't be at fault.The Key TakeawayThese examples make one thing clear: cloud security doesn't come out of the box. While AWS secures the underlying infrastructure, everything built on top of it is the customer's responsibility. Overlooking that fact can expose an organization to serious risk - but with the right tools, staying secure is entirely within reach.Level Up Your Cloud Security With IntruderIntruder helps you stay ahead of all these vulnerabilities and more, by combining agentless cloud security scanning, vulnerability scanning, and attack surface management in one powerful, easy-to-use platform.Why it's a game changer:Find what others miss: Intruder combines external vulnerability scanning with information from AWS accounts to find risks that other solutions might miss.No false alarms: CSPM tools can overhype severity. Intruder prioritizes real risks so you can focus on what truly matters.Crystal clear fixes: Issues are explained in plain English with step-by-step remediation guidance.Continuous protection: Stay ahead with continuous monitoring and alerts when new risks emerge.Predictable pricing: Unlike other cloud security tools that can rack up unpredictable costs, there's no surprise charges with Intruder.Get set up in minutes and receive instant insights into your cloud security start your 14 day free trial today.Found this article interesting? This article is a contributed piece from one of our valued partners. Follow us on Twitter and LinkedIn to read more exclusive content we post.SHARE

With more AI models emerging, how do CIOs and CISOs at enterprises go about establishing security guardrails to reduce risks as the technology gets deployed?

John Edwards, Technology Journalist & AuthorMarch 31, 20255 Min ReadAndrey Suslov via Alamy Stock VectorAgentic AI uses sophisticated reasoning and iterative planning to autonomously solve complex, multi-step problems. By absorbing massive amounts of data from multiple sources, the technology can build strategies, analyze challenges, and execute tasks in an almost endless range of business and research sectors, including supply chains, cybersecurity, and healthcare.Traditional AI systems typically excel at narrowly defined tasks under tightly controlled conditions, says Michael Craig, staff scientist at AI drug discovery firm Valence Labs. Agentic AI systems aren't restricted to a single, narrow purpose. "They can identify which questions to explore, what experiments to run, and how to adjust a methodology as new data emerges," he notes via email.Agentic AI functions like a workflow compared to other AI applications, says Joe Fernandes, vice president and general manager at enterprise open-source software provider Red Hat's AI unit. "Rather than a typical generative AI model generating a single response to a question, an agentic AI system may execute several steps on its own to complete the task," he explains in an email interview. This could include analyzing the request, mapping out a strategy, and executing the task, which in itself could be calling out to additional models or external systems, such as a search engine or querying a database.Related:A Force MultiplierWhen fully realized, agentic AI can be a force multiplier to an extreme degree, Fernandes says. "Looking at it from the perspective of a traditional enterprise IT organization, it's like having an incredibly specialized individual -- or team of individuals -- that doesn't mind having the same task, every day, with no creativity or scope expansion."Given its powerful and wide-ranging abilities, agentic AI presents an opportunity to advance scientific research by analyzing petabytes of data, formulating hypotheses, and pinpointing salient patterns in an asynchronous manner. "This has the potential to accelerate advancement in data-heavy fields like biology, chemistry, and drug discovery," Craig says. "Furthermore, agentic AI can update plans based on intermediate findings without needing continuous human supervision, which can result in a broader exploration of possible solutions." Perhaps most importantly, by testing ideas in simulated environments, agentic AI can lower reliance on expensive wet lab experiments, improving the likelihood that subsequent experiments will drive insight.Agentic AI can also free IT team members from maintenance and other low-level tasks, Fernandes says. Instead, staff can work on integrating new systems or applications, engaging more closely with customers, and handling other important duties. "In this scenario, agentic AI takes on the unpleasant tasks of IT work and lets a technology organization drive incredible value for the broader business rather than being stuck in cycles of system maintenance."Related:Over time, agentic AI has the ability to improve its performance by learning from experience, becoming increasingly effective at achieving desired outcomes, says Marinela Profi, global AI market strategy lead at business analytics software and services provider SAS in an online interview. "For example, it might reschedule deliveries to avoid traffic or change a factorys production plan if demand rises."First AdoptersInitial agentic Ai adopters will likely be enterprises looking to maximize their AI investments, boost productivity, and tackle complex business challenges, predicts Lan Guan, chief AI officer at business advisory firm Accenture. "These organizations are particularly interested in solutions that can scale across multiple functions and operate with minimal human oversight," she notes via email.Related:Enterprises across a wide range of verticals are most likely to be the first to commit to agentic AI, eying the potential for reduced costs, Fernandes says. "Looking at agentic AI in the long term, its feasible that almost every organization in nearly every industry can benefit from adopting agentic AI agents in some fashion."First StepsThe best way to get started with agentic AI is by establishing a strong foundational infrastructure and resilient data management practices, Guan says. "Organizations are at varying stages of readiness, and those with a robust enterprise platform architecture are better positioned to ensure seamless accessibility to foundation models."An easier approach to agentic AI is simply experimenting with the technology. "The good news here is that much of the innovation surrounding agentic AI, and AI in general, is happening in open source," Fernandes observes. He points to several emerging agent tools/ frameworks, including CrewAI and LangChain, among many others.On the downside, agentic AI faces some of the same challenges as other generative AI use cases. "The underlying GenAI models need to be trained and tuned on your data and deployed for inference across a hybrid environment that may extend from public clouds, to private data centers and out to the edge," Fernandes explains. "This needs to be done in a cost-effective way to ensure a positive ROI, which is a challenge given that this generally requires accelerated compute hardware, namely GPUs."Adopting organizations also must possess the internal skills and resources needed to effectively train models on their data, as well as a clear deployment strategy. "AI agents offer a solid pathway to production AI, but the constantly evolving market, from new model introductions and technologies to training and RAG-type techniques, means that most IT organizations are currently being very deliberate in their pursuit," Fernandes says.About the AuthorJohn EdwardsTechnology Journalist & AuthorJohn Edwards is a veteran business technology journalist. His work has appeared in The New York Times, The Washington Post, and numerous business and technology publications, including Computerworld, CFO Magazine, IBM Data Management Magazine, RFID Journal, and Electronic Design. He has also written columns for The Economist's Business Intelligence Unit and PricewaterhouseCoopers' Communications Direct. John has authored several books on business technology topics. His work began appearing online as early as 1983. Throughout the 1980s and 90s, he wrote daily news and feature articles for both the CompuServe and Prodigy online services. His "Behind the Screens" commentaries made him the world's first known professional blogger.See more from John EdwardsWebinarsMore WebinarsReportsMore ReportsNever Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.SIGN-UPYou May Also Like

Veteran MP says there was never a sense that department had taken responsibilityClive Betts MPThe former chair of the housing select committee has criticised the staggering underperformance of the Ministry for Housing, Communities and Local Government (MHCLG) in its efforts to tackle the cladding remediation crisis.Clive Betts, MP for Sheffield South East, spoke to BDs sister title Building in the wake of a highly critical report by the public accounts committee of which he is now deputy chair.The report said the governments plan to remediate dangerous buildings across the UK was both insufficiently ambitious and at risk of not delivering what it promises, and found that as many as 7,000 unsafe buildings had yet to be identifiedEight years on from the Grenfell Tower fire, in which 72 people died, Betts said this high number of unidentified buildings demonstrated an incredible underachievement by successive governments and housing ministers.Seeking to explain the failure, he told Building that there was never a sense that the department took up responsibility.MHCLG has primarily been responsible for the governments approach to remediation, much of which was developed while Michael Gove was the secretary of state.>> Read more:MPs sceptical government can deliver on building safety remediation promisesGove was widely seen as taking a firm approach with developers in an effort to encourage them to fix legacy building safety issues themselves.Betts said the department had felt its job was to lay out legal frameworks, and then others, like leaseholders and building owners got on with it.Obviously thats not worked [and] I think the department has to take a much clearer central role, he said.The MP, who chaired the housing committee in Commons for 14 years, said that financial arrangements to support remediation had been disparate and said he was concerned that the deferment of the Building Safety Levy could push the Treasurys contribution higher than the stated 5bn limit.However he added that it was reassuring that there is a recognition now that this has to be accelerated and there has to be central oversight, after the new Labour government published its plan to address the issue late last year.

Scheme reflects wider move to diversify uses within the City of Londons commercial coreSource: Chris Dyson ArchitectsThe proposed new designChris Dyson Architects has submitted a planning application to the city of London corporation for the conversion and extension of a prominent postwar office building at 1 Farringdon Street into an 85-room hotel.The proposed 3,460m development, located by Ludgate Circus, is owned by Lipman Properties.The plans include the addition of a new mansard roof storey, set back from the existing parapet line.The application also includes proposals for a publicly accessible rooftop garden and bar. This top-level terrace would be served by new lift access housed within a brick-clad volume designed to resemble a chimney stack.At ground level, the proposals include new caf and dining spaces, with co-working areas facing the street.The proposed new designSource: Chris Dyson ArchitectsExisting buildingSource: Chris Dyson Architects1/2show captionChris Dyson said: The building has a fantastic location in terms of serving visitors to the City, whether for work or leisure. Its close to St Pauls Cathedral, the new Museum of London, the Elizabeth Line, and within the burgeoning Fleet Street Quarter, which he described as helping turn a too often overlooked part of London into an appealing, walkable destination.He said he hoped the scheme would be one of a wave of new creative reinventions of City buildings, contributing to lower carbon emissions by reusing existing architecture and bringing a greater diversity of activities to the square mile.The building, which is unlisted, retains the principal north and west facades of a four-storey commercial block dating from the early 1870s. While the London stock brick and stone-detailed elevations remain, the structure was substantially reconstructed and extended during the early 1980s.The scheme targets a BREEAM Excellent rating and is intended to support the Citys growing visitor economy.>> Also read:Axiom Architects submits plans to transform City tower into 420-bed Premier Inn>> Also read:Beyond work: how hotels are redefining the office conversion

Teams selected for the maximum four-year agreement will provide specialist development advice to the quango, local authorities and other public bodies across England to promote happier healthy communities and better places to live.The framework is divided into three lots covering strategic outcomes planning and leisure services delivery, assessments of need, and urban design services. Key aims include ensuring new and renewed places support people to become more physically active.According to the brief: We know bespoke procurement exercises are frequently conducted to appoint specialist support in strategic planning, leisure procurement, urban design and planning.AdvertisementThis framework will provide access to specialist consultants (small and medium sized enterprises, as well as larger organisations) with extensive knowledge and expertise, who will be selected for their ability to deliver high quality, consistent services for contracting authorities to support active environments and to create places and spaces for people to be more active and build healthier active communities.Sport England was created in 1996 with a remit to promote government and lottery investment in new facilities to increase the number of people participating in sport and activity.The organisation aims to promote both regular sport and general activity among a wide range of people including groups which statistically engage less in sports. The organisation has provided support and guidance on a range of high-profile projects including the Olympics, Commonwealth Games and Wembley Stadium development.Key recent research by the organisation has included the creation of affordable models for new sporting facilities including school sports halls, sports centres and adult community swimming pools.Recent Sport England-supported projects have included a new timber-framed grandstand by Hopkins Architects at Herne Hill Velodrome in south London.AdvertisementThe latest framework will be open to local authorities, town and parish councils, non-departmental public bodies and leisure services providers to help them procure specialist services that help promote active lifestyles.Individual projects let through the framework will be awarded either by mini competition, call-off or direct award. Fees are expected to be worth around 3 per cent-to-5 per cent of the total framework value over its lifetime.Bids for inclusion on the latest framework will be evaluated 70 per cent on quality, 20 per cent on cost and 10 per cent on social value.Competition detailsProject title SE1184 Active Environments FrameworkClient Sport EnglandContract value 5.4 millionFirst round deadline Midday, 18 July 2025Restrictions TbcMore information https://www.find-tender.service.gov.uk/Notice/011842-2025

Construction begets destruction. Architecture has long been admired for being a slow art, but it has been slow too in facing up to the deeply-rooted historical prejudices that continue to cast a long shadow over the profession and its role in what the United Nations describes as the triple planetary crisis of climate change, pollution, and biodiversity loss. In September, I will begin a three-year piece of original research into the extractive history of architecture as part of a Major Research Fellowship awarded by the Leverhulme Trust.Situated at the intersection of architecture, coloniality and the triple planetary crisis, this work will confront the disciplines role in heralding the Anthropocene a potentially new geological epoch defined by our species planetary impact. The research intends to investigate the discipline not as a chronology or assemblage of built objects, styles, personalities, or experiences predominantly from the West, but as an extractive, invariably inequitable and planetary process of world-making.This work builds on my long-standing interest in researching non-canonical histories that have been overlooked, marginalised, or ignored and, more recently, their intersection with the Anthropocene.AdvertisementWhile debates still rage around the definition of the Anthropocene, the terms effectiveness as an overarching cross-disciplinary framework for researchers remains vital.As Chris Thomas,head of the Leverhulme Centre for Anthropocene Biodiversity at the University of York (UK) recently wrote inNature, The concept of an era of human-driven change provides convenient common ground to collaborate with researchers from other disciplines.This is something that people in the arts and humanities and the social sciences have picked up as well [] It is a means of enabling communication about the extent to which we are living in a truly unprecedented and human-altered world.It is in this spirit that this project was conceived, providing a response to this existential challenge from the discipline of architecture, both as a major study of its past and as an advocation for a different and better future.As someone professionally invested in the built environment and its history for over a quarter of a century,I have long been uncomfortable with how the discipline of architecture privileges the experiences of a global minority over those of what theeducator and anti-racist activist Rosemary Campbell-Stephens has coined the global majority.As we increasingly confront the reality of our planetary impact and limitations, this partial view of the environments we have constructed most of which have been built comparatively recently is not merely disingenuous, but increasingly and dangerously irresponsible.Today, the buildings and construction sector accounts for around 37 per cent of total global CO2 emissions, much of which comes from the production and use of cement for concrete. According to the United States Geological Survey, in the three years from 2011-2013, China consumed more cement than the USA did throughout the entire 20th century. In 2007, as the global human population was nearing 7 billion (it now exceeds 8 billion), humans became an urbanised species for the first time, with more people living in cities than in rural areas. Since the early 20thcentury, the mass of material humans produce (a large proportion of which are building materials) has doubled every two decades, resulting in the total mass of human-made material now outweighing all biomass on earth.Such statistics reveal a bigger picture of extraction over longer timeframes, evidencing architectures planetary impact.These facts not only reveal our species suicidal appetite for building anew, they also mask, as Professor Jeremy Till has stated, architectures addiction to extraction. The roots of this addiction run deep into extractive histories of coloniality, wherein architecture was deployed as a means of projecting and sustaining power intellectually and geopolitically the canon and the cannon.AdvertisementFor more than half a millennium, architecture has been rooted in the extraction of natural and cultural resources through colonial domination and cultural exploitation, from coal to humans, from data to indigenous knowledge and from energy to intellectual property. When published, my research aims to join and support the growing voices arguing for architecture to become a restorative, regenerative and reparative practice, encouraging an emergent and urgent change in architectural education and practice that shifts a centuries-old focus on building anew to repairing and improving the already built.Edward Denison is Professor of Architecture and Global Modernities at the Bartlett School of Architecture 2025-03-31will hurstcomment and share

The iPhone company is reportedly moving away from smaller phones to even bigger handsets.

March 30, 2025Shark Sounds, Molecules on Mars and Continued Federal CutsCuts to federal health and science agencies continue. Plus, we discuss the sounds of sharks, the meaning of Martian molecules and one big dino claw. Anaissa Ruiz Tejada/Scientific AmericanSUBSCRIBE TO Science QuicklyRachel Feltman: Happy Monday, listeners! For Scientific Americans Science Quickly, Im Rachel Feltman. Lets kick off the week and wrap up the month with a quick roundup of the latest science news.[CLIP: RFK Jr. announces the planned cuts on Thursday in a HHS video: We're gonna eliminate an entire alphabet soup of departments and agencies.]Feltman: Last Thursday the U.S. Department of Health and Human Services announced plans to cut 10,000 full-time jobs across the department. Another 10,000 individuals have already accepted voluntary retirement and buyouts. The layoffs will hit the Food and Drug Administration, the Centers for Disease Control and Prevention, the National Institutes of Health and the Centers for Medicare and Medicaid Services.On supporting science journalismIf you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.[CLIP: RFK Jr.: Twenty-eight great divisions will become 15. The entire federal workforce is downsizing now, so this will be a painful period for HHS as we downsize from 82,000 full-time employees to around 62,000.]Feltman: HHS Secretary Robert F. Kennedy Jr. said in a statement that the aim of these cuts is to save money and boost efficiency.Meanwhile, last week the Trump administration also moved to cancel more than $12 billion in federal grant funding to state and local health departments. Axios reports that the main targets are grants for COVID testing, initiatives aimed at tackling health disparities, and vaccinations. As of last Thursday those cuts had reportedly already led to layoffs at the Virginia Department of Health.Well, of course, be watching these developments and keeping you posted. But for now, lets move on to some exciting news from Mars. According to a study published last week in the Proceedings of the National Academy of Sciences USA, NASAs Curiosity rover has found the biggest carbon-based molecules ever seen on the Red Planet. The long-chain alkanes are thought to have come from fatty acids, which are the building blocks of cell membranes in living organisms on Earth.Now, these long molecules arent necessarily a smoking gun for Martian life. We know that fatty acids can form by way of chemistry instead of biology. In fact, some scientists think we first got fatty acids on Earth thanks to the interaction of water and minerals in hydrothermal vents. So while fatty acids are necessary for life as we know it, its possible they formed on Mars without life ever finding a way. Still, this finding is another point for Mars in the quest to determine potential past habitability. Plus, since these compounds were found preserved in a 3.7-billion-year-old rock, the discovery gives scientists hope that if microbial life once existed on Mars, we might still be able to find signs of it.Speaking of size superlatives: paleontologists are showing off a really freaking big dinosaur claw in pristine condition. It belongs to a new species of therizinosaur, which was described in a study published in the journal iScience last Tuesday.Writing for National Geographic, Riley Black explained that therizinosaurs were, generally speaking, a weird bunch. The dinosaurs were descended from carnivores but had come to eat plants. They were kind of slothlike, apparently, down to their three giant clawsexcept that they were also giant and covered in feathers. But a specimen found in Mongolia's Gobi Desert back in 2012 has revealed a new species that stands out for having just two fingers instead of three.One of the fingers still has a sheath of keratin that would have protected the actual bone of the claw. This protective covering also added length, creating a talon nearly a foot long. Scientists think the new species likely lost its third digit as a result of evolution. While the creatures sharp claws look like something a raptor would use to tear at prey, these oddballs probably used them to hook branches while foragingwhich the authors of the new study think could have been done more efficiently with a two-fingered grasp than a three-fingered one.Well keep the animal theme rolling to wrap us up with a couple of new papers on animal behavior under the sea. First, a new study on sharks. The predators are known for their stealth, but research published Wednesday in the journal Royal Society Open Science is absolutely blowing up their spot. While the study authors note that sharks and other elasmobranchs, which is a group that also includes rays, are not historically viewed as active sound producers, the researchers managed to catch rig sharks making little clicking noises.The studys lead author reportedly heard some unusual sounds while working with sharks back in grad school but wasnt able to investigate further until recently. In the new study she and her colleagues observed 10 rig sharks in tanks tricked out with underwater microphones. They caught the sharks making extremely shortlike, shorter-than-a-human-blink short so literally blink and youll miss it stuff. And those noises reached a maximum of 156 decibels, on average. The sharks made a lot more noise when handlers first touched them, and the noises tended to subside as they got used to being held. That could mean these are deliberate sounds, like a whats the big idea or a guys, heads-up, these humans are pretty handsy. But well need a lot more research to be sure.And in case youre wondering those clicks sound like this:[CLIP: Rig sharks make clicklike sounds.]Feltman: Sharks lack the swim bladder that most fish use to make noises, but researchers suspect the rigs make these clicks through the forceful snapping of their teeth. As a habitual tooth grinder I can certainly relate. Since sharks are, generally speaking, a pretty toothy bunch, it stands to reason that other species could be producing sounds similar to these.And while sharks are potentially using sound to communicate, cuttlefish are apparently using visual tricks to mesmerize their prey. Cuttlefish are known for having specialized skin cells that allow them to rapidly change color and create patterns for camouflage. Last month a group of researchers published examples of different visual displays that one cuttlefish species might use to trick prey. The scientists recorded broadclub cuttlefish seemingly mimicking floating leaves and branching pieces of coral, as well as generating some pulsing patterns, an effect that makes it look like a dark stripe is moving down a cuttlefishs body. That's kind of a surprising tactic because to human eyes its like a flashing sign that says cuttlefish incoming. But in a new study published last Wednesday in Science Advances, the same researchers argue that this passing-stripe display helps a cuttlefish hunt by overwhelming a prey animals senses. From the perspective of a crab, for example, these fast-moving stripes could distract from the actual movements of the approaching cuttlefish. So its all very pay no attention to the cuttlefish behind the striped curtains!Thats all for this weeks news roundup. Well be back on Wednesday with special guest Wendy Zukerman from Science Vs to talk about the science behind a big debate surrounding a certain sexual phenomenon.Science Quickly is produced by me, Rachel Feltman, along with Fonda Mwangi, Kelso Harper, Naeem Amarsy and Jeff DelViscio. This episode was edited by Alex Sugiura. Shayna Posses and Aaron Shattuck fact-check our show. Our theme music was composed by Dominic Smith. Subscribe to Scientific American for more up-to-date and in-depth science news.For Scientific American, this is Rachel Feltman. Have a great week!