Sign in

Author: DEROBS

  • TiDB breach highlights risks in cloud security assumptions

    TiDB breach highlights risks in cloud security assumptions

    A security breach involving TiDB has surfaced concerns regarding underlying assumptions in cloud security. The incident underscores a critical vulnerability: cloud providers may unknowingly host sensitive customer data, leaving both the provider and the customer exposed if a security failure occurs.

    When a cloud provider experiences a breach, the presence of these unidentified data stores can create significant risks for businesses and users alike. The situation serves as a reminder that data security requires clear visibility into what information is being stored and the security protocols applied to that data across cloud environments.

    Source: Dark Reading
    Original article: https://www.darkreading.com/cloud-security/tidb-breach-flawed-cloud-security

  • Boston to host TechCrunch Founder Summit on November 4

    Boston to host TechCrunch Founder Summit on November 4

    The upcoming TechCrunch Founder Summit is set to take place in Boston on November 4, 2026. The one-day event is designed to bring together founders from various development stages to facilitate networking and professional growth.

    The summit aims to connect startup leaders with experienced entrepreneurs and venture capitalists. Attendees will have access to sessions focused on providing tactical insights into the challenges of building and scaling businesses within the current digital economy.

    Source: TechCrunch
    Original article: https://techcrunch.com/2026/10/21/techcrunch-founder-summit-2026-everything-you-need-to-know/

  • Adaptive agentic intelligence framework optimizes AI request routing

    Adaptive agentic intelligence framework optimizes AI request routing

    Organizations developing agentic AI systems often rely on a one-size-fits-all approach, routing every request to a large, cloud-hosted model. This practice frequently results in unnecessary API expenses, increased latency, and a dependence on consistent network connectivity. To address these inefficiencies, the Adaptive Agentic Intelligence Framework (AAIF)—built using Strands Agents—proposes a tiered processing architecture that optimizes model selection based on task complexity.

    A Tiered Architecture for AI Orchestration

    The AAIF architecture organizes processing into three distinct layers to ensure that computational resources are matched appropriately to the task at hand:

    • Smart Orchestration: A lightweight small language model (SLM), specifically llama3.2:1b, acts as the decision-making layer, classifying incoming requests and routing them to the correct processing path without requiring hardcoded rules.
    • Local Processing: Routine tasks such as device telemetry retrieval and basic image analysis are handled locally on edge hardware via Ollama. This approach provides low latency and eliminates the need for cloud round-trips for simpler queries.
    • Cloud Processing: Complex analytical requests, which necessitate multi-step reasoning or broader context, are escalated to frontier models accessed through Amazon Bedrock.

    This design is highly flexible, supporting various deployment environments ranging from edge devices like a Raspberry Pi to full-scale cloud implementations. Organizations can also configure hybrid setups that capitalize on local speed for common tasks while reserving cloud capacity for more demanding analytical requirements.

    Operational Efficiency and Future Scaling

    The framework draws on research suggesting that SLMs are often sufficient and more economical for many agentic system invocations. By separating routing logic from processing tools, developers can independently reconfigure or swap models based on hardware constraints or data sovereignty needs. For instance, in air-gapped environments, the system can be configured to use larger local models, ensuring sensitive data remains on-premises.

    Looking ahead, the development team highlighted the potential for scaling this architecture using the Model Context Protocol (MCP). By exposing tools as MCP servers, organizations can centralize tool management, allowing multiple agents across distributed services to share capabilities transparently. To maintain security, the framework supports the integration of Amazon Bedrock Guardrails, which apply content filtering and grounding checks to cloud-bound requests, ensuring consistent safety standards across both local and cloud-based tiers.

    Source: aws.amazon.com
    Original article: https://aws.amazon.com/blogs/machine-learning/adaptive-agentic-intelligence-framework-using-strands-agents/

  • Onamae Desktop Cloud reports security breach affecting up to 27,000 servers

    Onamae Desktop Cloud reports security breach affecting up to 27,000 servers

    Onamae Desktop Cloud, a prominent provider of virtual private server (VPS) services in Japan, has suffered a significant security compromise. According to reports, unauthorized actors gained access to the company’s infrastructure and successfully deployed malware onto the servers of approximately 26,739 customers.

    The incident highlights the risks associated with centralized cloud management infrastructure. By compromising the provider’s systems, attackers were able to distribute malicious software across a large footprint of customer environments. At this time, the scope of the potential data exposure or the specific nature of the deployed malware remains under investigation.

    As the situation develops, users of the platform are advised to monitor their systems for unauthorized activity and follow official guidance provided by the company regarding remediation and security hardening steps.

    Source: BleepingComputer
    Original article: https://www.bleepingcomputer.com/news/security/japans-onamae-desktop-cloud-hacked-up-to-27-000-servers-affected/

  • Google integrates Gemini AI into Chrome for enhanced scam detection

    Google integrates Gemini AI into Chrome for enhanced scam detection

    Google has announced the integration of its Gemini AI technology into the Chrome browser to bolster its defense against increasingly sophisticated online threats. The new security capabilities are specifically designed to identify deceptive websites and warn users when they encounter threats associated with ‘ClickFix’ attacks.

    ClickFix is a form of social engineering where attackers manipulate users into executing malicious commands, often by tricking them into copying and pasting instructions into their computer’s terminal or other sensitive interfaces. By utilizing Gemini’s analytical capabilities, the updated Chrome security features aim to detect these deceptive tactics more effectively, providing proactive warnings to users before they interact with potentially harmful content.

    This development represents an expansion of Google’s browser-based security efforts, moving beyond traditional signature-based detection to leverage generative AI models for real-time threat identification. As browser-based attacks continue to evolve, the integration of AI models serves as a defensive layer to mitigate risks associated with social engineering and malicious scripts.

    Source: BleepingComputer
    Original article: https://www.bleepingcomputer.com/news/google/google-chrome-adds-gemini-powered-scam-detection-warns-users-of-clickfix-attacks/

  • Blackstone’s Jas Khaira addresses concerns regarding AI infrastructure debt

    Blackstone’s Jas Khaira addresses concerns regarding AI infrastructure debt

    Jas Khaira, head of AI at Blackstone, has publicly dismissed growing market apprehensions concerning the significant debt loads associated with the ongoing global expansion of artificial intelligence infrastructure. As the private equity firm continues to evaluate and deploy capital within the technology sector, Khaira’s comments suggest a strategic confidence in the long-term viability of these massive infrastructure projects despite prevailing industry jitters.

    The debate surrounding AI-related debt has intensified as companies funnel billions of dollars into data centers, specialized hardware, and power-intensive computing facilities. Critics frequently point to the potential risks posed by the high cost of this capital-intensive “AI buildout.” However, Khaira’s stance indicates that Blackstone maintains a more optimistic outlook on the economic fundamentals supporting these investments.

    The commentary from the firm’s AI leadership comes amid broader market scrutiny regarding the return on investment for generative AI initiatives and the financial stability of the firms driving the underlying infrastructure. By characterizing these concerns as overblown, Blackstone aligns itself with a perspective that prioritizes the essential, long-term nature of AI-ready digital infrastructure as a foundational component of the future digital economy.

    Source: TechCrunch
    Original article: https://techcrunch.com/2026/10/21/blackstones-jas-khaira-says-fears-over-ai-buildout-debt-are-overblown/

  • AWS explores production-grade multi-agent system architecture with Deep Insight

    AWS explores production-grade multi-agent system architecture with Deep Insight

    Building multi-agent systems often proves straightforward during pilot phases, yet moving these implementations into production environments reveals significant complexities. To address these difficulties, the AWS Solutions Architecture team developed Deep Insight, a production-grade multi-agent system designed to automate complex data analysis tasks while ensuring operational reliability.

    Addressing Production Challenges

    According to the AWS team, developers typically encounter five core challenges when scaling agentic systems: controlling execution flow, selecting appropriate large language models (LLMs) for specific tasks, ensuring reliable operations, securing the execution of AI-generated code, and maintaining system visibility.

    Deep Insight addresses these areas by integrating several technologies. Execution flow is managed via the Strands Agents SDK, utilizing a Graph pattern to define order and a tool-based architecture that allows a supervisor agent to invoke specialized sub-agents—such as a coder, validator, or reporter—on demand. This structure enables a human-in-the-loop (HITL) approach, allowing users to review and revise analysis plans before execution.

    Infrastructure and Security

    For production deployments, the system leverages Amazon Bedrock and Amazon Bedrock AgentCore. To mitigate security risks associated with executing LLM-generated Python code, the architecture employs ephemeral, per-session containers running on AWS Fargate. These containers operate within a private VPC with restricted access, preventing the agents from accessing sensitive system files or other users’ data.

    Operational visibility is achieved through integration with Amazon DynamoDB, Amazon Simple Notification Service (SNS), and Amazon Cognito. These components provide logging, alerting for system failures, and a secure administration dashboard for monitoring task metrics and performance.

    Real-World Applications

    The efficacy of this architectural design has been demonstrated through practical use cases. LG Electronics implemented a similar system for its marketing division, reducing the time required for data analysis from three days to 30 minutes. Additionally, the ‘TechRecon’ project presented at AWS re:Invent 2025 applied these orchestration principles to technology strategy analysis, further validating the approach as a general-purpose pattern for multi-agent workflows.

    Deep Insight is available as an open-source project on GitHub, providing developers with options ranging from local CLI-based development to fully managed production environments using AgentCore.

    Source: aws.amazon.com
    Original article: https://aws.amazon.com/blogs/machine-learning/5-production-challenges-in-multi-agent-systems-and-how-deep-insight-solved-them/

  • AWS details private network connectivity architecture for Amazon Quick

    AWS details private network connectivity architecture for Amazon Quick

    Amazon Web Services (AWS) has published detailed guidance on establishing secure, private network connectivity for Amazon Quick, the platform’s AI-powered workspace for analytics and automation. By utilizing a combination of VPC endpoints and VPC connections, organizations can ensure that sensitive enterprise data remains within controlled network boundaries, addressing common compliance requirements such as HIPAA, FedRAMP, and PCI-DSS.

    Dual-path connectivity architecture

    The solution relies on two distinct traffic management mechanisms. First, AWS PrivateLink powers VPC endpoints to handle user-facing traffic, including console access, API calls, and administrative functions. This prevents user sessions from traversing the public internet. Second, VPC connections enable Amazon Quick to access internal data sources—such as private databases and internal APIs—by placing elastic network interfaces (ENIs) directly into the user’s virtual private cloud (VPC).

    Implementation and configuration

    To implement this architecture, enterprises must already have an AWS Direct Connect or Site-to-Site VPN link to their VPC. The configuration process involves four primary building blocks: configuring VPC endpoints for user and API access, establishing VPC connections for data fetching, setting up endpoints for AWS data services, and managing outbound access to essential external domains. Security is further reinforced through the use of VPC endpoint policies and specific security group rules that restrict access to authorized corporate network ranges.

    Identity and access control

    Beyond network-level security, the framework integrates with AWS IAM Identity Center for workforce identity management. Role-based access control allows administrators to assign permissions based on user groups synced from an existing corporate identity provider. Furthermore, Amazon Quick supports row-level and column-level security for data, ensuring that access to sensitive information is governed by both the user’s identity and established organizational policies.

    Security and compliance implications

    This architecture is designed to minimize the attack surface by eliminating public internet exposure for internal data traffic. Organizations can also enforce strict access controls by implementing VPC endpoint restrictions, which ensure that only traffic originating from authorized endpoints or specified IP ranges can reach the Amazon Quick account. AWS advises that this layered approach—combining network isolation with granular identity controls—provides a robust foundation for maintaining compliance and securing sensitive AI-driven workflows.

    Source: aws.amazon.com
    Original article: https://aws.amazon.com/blogs/machine-learning/securing-quick-access-from-corporate-networks-with-aws-privatelink/

  • Manfred J. Schindler elected as 2027 IEEE president-elect

    Manfred J. Schindler elected as 2027 IEEE president-elect

    IEEE Life Fellow Manfred “Fred” J. Schindler has been elected as the 2027 IEEE president-elect. Following his term as president-elect, he is set to begin serving as the organization’s president on January 1, 2028. Schindler secured the position with 16,421 votes, surpassing the 13,874 votes cast for Senior Member David Alan Koehler.

    Strategic priorities

    Schindler has outlined three primary objectives for his tenure as president-elect. He intends to strengthen the organization’s connection with industry by coordinating cross-geographic and cross-disciplinary events aimed at revitalizing professional activities, with a specific focus on young professionals. Furthermore, he plans to champion the modernization of IEEE’s digital infrastructure to replace manual processes with updated interfaces. Finally, he aims to prioritize technological advancement to support future generations by bolstering existing programs and initiatives.

    As of the time of the announcement, the election results remain unofficial pending the formal acceptance of the IEEE Teller’s Committee report by the IEEE Board of Directors, which is scheduled for November.

    Background and experience

    A specialist in microwave semiconductor technology, Schindler is currently an independent consultant providing technical expertise and project management services. His career includes leadership roles as the CTO of Anlotek and as a director at Qorvo and the Boston design center of RFMD. Additionally, he held positions as an applications manager at IBM, an engineering manager at ATN Microwave, and a lab manager at Raytheon. Throughout his career, he has focused on the development of microwave integrated-circuit technology, contributing numerous publications and holding 11 patents.

    Schindler is a veteran IEEE volunteer with over 30 years of service. His leadership experience within the organization includes serving as the 2024 vice president of IEEE Technical Activities, 2021–2022 Division IV director, and president of the IEEE Microwave Theory and Technology Society (MTTS) in 2003. He has also been recognized with multiple awards, including the MTTS 2018 Distinguished Service Award and the 2015 N. Walter Cox Award.

    Source: IEEE Spectrum
    Original article: https://spectrum.ieee.org/schindler-2027-ieee-president-elect

  • Robbin launches credit intelligence integration for ChatGPT and Claude

    Robbin launches credit intelligence integration for ChatGPT and Claude

    Robbin has announced the launch of what it describes as Latin America’s first credit Model Context Protocol (MCP). This integration allows finance teams to access the company’s proprietary credit intelligence directly within widely used AI assistants, specifically OpenAI’s ChatGPT and Anthropic’s Claude.

    By embedding these capabilities into existing AI workflows, the technology enables teams to streamline credit-related tasks through natural language interaction. According to the company, users can perform real-time risk assessments across their buyer base and generate credit limit recommendations for specific customers.

    The integration also includes automated monitoring features, such as flagging account limits that require manual review. This development aims to provide finance professionals with a more direct way to leverage data-driven insights during their daily operations by reducing the need to switch between disparate financial software and generative AI platforms.

    Source: Finextra
    Original article: https://www.finextra.com/pressarticle/111392/robbin-launches-credit-integration-with-chatgpt-and-claude?utm_medium=rssfinextra&utm_source=finextrafeed