Newsfeed
title: “Daily Tech News Digest - October 04, 2025: OpenAI’s Sora 2, DoorDash’s Robot, and IBM’s Granite 4.0” description: “Stay ahead with our October 4, 2025 tech news digest. Discover the latest on OpenAI’s Sora 2, DoorDash’s delivery robot, IBM’s Granite 4.0, and key updates in AI, robotics, cloud, and quantum computing.” date: “2025-10-04T00:00:00Z” draft: false comments: true tags: [“Artificial Intelligence”, “Machine Learning”, “Autonomous Vehicles”, “Cloud Infrastructure”, “DevOps”, “Generative AI”, “Kubernetes”, “Open Source”, “Quantum Computing”, “Robotics”, “Software Engineering”, “Tech News”, “OpenAI”, “IBM”, “Salesforce”, “Microsoft”, “DoorDash”] categories: [“Artificial Intelligence & Machine Learning”, “Cloud Infrastructure”, “DevOps”, “Generative AI”, “Kubernetes”, “Open-Source”, “Quantum Computing”, “Robotics & Autonomous Vehicles”, “Software Engineering”]
AI & ML Updates: Google’s Cloud Restructure, Financial Access, and Enterprise Automation
This edition of the news roundup covers the latest developments in Artificial Intelligence and Machine Learning. Key stories include new research on how machine learning is expanding financial access while reducing risk, Google’s strategic layoffs in its Cloud division to double down on AI, and the launch of new AI-powered enterprise automation services.
Sources:
Ford Dealership Deploys Autonomous Parts Delivery Robot in Canada
A Ford Lincoln dealership in Canada, Yonge-Steeles Ford Lincoln, has become the first in the country to use a robot for delivering parts to service technicians. The initiative is a collaboration aimed at improving productivity and efficiency by reducing the time technicians spend retrieving parts. The robot, which resembles a service cart with a computer screen, navigates the dealership autonomously using sensors. By having the robot deliver parts, technicians can remain in their work bays, potentially cutting the total time for a parts order and retrieval in half. An app was specifically developed for this application, and it is estimated that the dealership could save over $500,000 annually in lost revenue from repair order downtime.
Sources:
DoorDash Unveils ‘Dot,’ Its First Commercial Autonomous Delivery Robot
DoorDash has introduced its first commercial autonomous delivery robot, named ‘Dot.’ Designed for quick local deliveries, Dot is one-tenth the size of a car and can travel on bike lanes, roads, and sidewalks. The electric robot has a top speed of 20 mph, making it suitable for neighborhood trips to support local businesses. DoorDash states that Dot is small enough to help reduce traffic congestion and agile enough for smooth handoffs to customers.
Sources:
OpenAI Launches Sora 2 Generative AI Video App
OpenAI has released Sora 2, the next version of its generative AI video model, which can create realistic short videos with corresponding audio from text prompts. The new model is available through a new standalone iOS app, currently on an invite-only basis. The app functions as a social network where users can create and publish their AI-generated videos to a feed, similar in style to TikTok. A feature called ‘Cameos’ allows users to insert their own image into the AI-generated clips. Sora 2 is positioned as a direct competitor to Google’s Veo 3, with capabilities to produce videos in various styles that do not immediately appear to be AI-generated.
Salesforce Launches Agentforce Vibes for Natural Language “Vibe Coding”
Salesforce has launched Agentforce Vibes, an enterprise-level platform for “vibe coding” that evolved from Salesforce Code Builder. This new tool integrates natural language to code capabilities directly into every paid Salesforce organization, aiming to make this style of development a mainstream enterprise feature with proper governance and DevOps alignment. Agentforce Vibes supports the entire application lifecycle management (ALM) workflow, from planning and building to testing and deployment, using natural language. It includes over forty MCP tools for interacting with metadata and managing DevOps tasks. A key feature is the Lightning Web Components MCP, which can automatically generate Jest tests for a component’s behavior.
Sources:
Black Box and Wind River Partner to Deliver Global Edge and Cloud Solutions
Essar’s digital unit, Black Box, has partnered with Wind River, an intelligent edge software company, to provide cloud infrastructure solutions to various industries worldwide. The collaboration will focus on delivering integrated edge and cloud-native infrastructure, secure private cloud deployments, and support for AI workloads. This partnership is expected to generate approximately Rs 1,350 crore in global revenue over the next five years. As part of the agreement, Black Box will have preferred status for delivering Wind River solutions in India and the Middle East and will also manage global customer engagements for Wind River.
Sources:
Critical Security Flaw (CVE-2025-10725) Found in Red Hat OpenShift AI
A critical security vulnerability, identified as CVE-2025-10725 with a CVSS score of 9.9, has been found in the Red Hat OpenShift AI service. This flaw could permit attackers with low-level authenticated access, such as a data scientist using a Jupyter notebook, to escalate their privileges to that of a full cluster administrator. A successful exploit could lead to a complete compromise of the cluster’s confidentiality, integrity, and availability, allowing for the theft of sensitive data and control over the underlying infrastructure. Red Hat has issued an advisory and recommends mitigating the risk by avoiding broad permissions for system-level groups and adhering to the principle of least privilege.
Sources:
ETSI Establishes New Committee for Quantum Technology Standardization
The European Telecommunications Standards Institute (ETSI) has established a new committee focused on quantum technologies. The creation of this committee signals the maturation of the quantum industry and the growing need for standardized frameworks. This initiative will address the evolving landscape of quantum security, computation, and networking. The committee’s work will be crucial in fostering interoperability and ensuring the secure and reliable deployment of quantum technologies in the future.
Microsoft Releases Open-Source AI Agent Framework
Microsoft has introduced the Microsoft Agent Framework, an open-source software development kit aimed at simplifying the creation of AI agents. This new framework unifies two of Microsoft’s existing projects: Semantic Kernel, known for its enterprise-level stability, and AutoGen, a research project focused on multi-agent workflows. The goal is to provide developers with a single platform that combines the innovative aspects of AutoGen with the reliability of Semantic Kernel. The framework is designed for both Python and .NET developers and supports various orchestration patterns, including sequential, concurrent, and group chat workflows. It is built on open standards, supporting OpenAPI and allowing for integration with external tools and collaboration between different agents. The public preview of the Agent Framework is now available, enabling developers to build and deploy AI agents with greater ease and efficiency.
Sources:
IBM Launches Granite 4.0, an Open-Source Generative AI Model with Hybrid Architecture
IBM has released Granite 4.0, a new family of open-source generative AI models. A key feature of Granite 4.0 is its hybrid Mamba/transformer architecture, which is designed to be more efficient and require less powerful GPUs compared to traditional transformer models. The initial release includes four small models, with more variations, including ’thinking’, medium, and nano sizes, planned for later in the year. According to IBM, the Granite 4.0 models are the first to be ISO 42001 certified and are cryptographically signed, ensuring a higher level of transparency and accountability. The models are available under the Apache 2.0 license and can be accessed through IBM’s watsonx.ai, as well as other platforms like HuggingFace and Kaggle. IBM has also established a bug bounty program for Granite, with rewards for identifying vulnerabilities.
Sources: