Google’s AI Mode has introduced interactive stock and mutual fund charts, greatly simplifying financial data analysis. Instead of providing just links, the AI generates custom, tappable charts tailored to specific queries, making complex information easier to understand. Google’s Gemini AI streamlines the data retrieval process, pulling real-time and historical data to present it visually. Users can ask for comparisons, like top CPG stock performance in a given year, and receive clear, interactive charts that highlight key data points for deeper insights. This feature promotes a conversational experience in Google Search, allowing users to ask follow-up questions seamlessly. While still in the experimental phase and available to Labs users in the U.S., this advancement represents a significant step in visualizing data and aligns with Google’s vision of integrating smart AI into various aspects of Search. Users are advised, however, to verify financial information before acting on it.

Source link

Google has unveiled a preview of its upgraded AI model, Gemini 2.5 Pro, branding it as the company’s “most intelligent” version to date. This enhanced model outperforms its predecessor from May’s I/O developer conference, achieving significant benchmark improvements. Notably, it scored 1470 on LMArena and 1443 on WebDevArena, both placing it at the top of these tests. Gemini 2.5 Pro excels in coding proficiency, topping the Aider Polyglot test and performing well in challenging assessments like GPQA and Humanity’s Last Exam, which evaluate reasoning and knowledge across various domains. Developers can access the model via Google AI Studio, Vertex AI, and the Gemini app, with full availability expected soon. This release reflects Google’s commitment to enhancing its AI capabilities in response to developer feedback and competitive pressures, particularly from contenders like OpenAI’s ChatGPT. Gemini signifies Google’s strategic focus on advanced reasoning and coding skills within its AI initiatives.

Source link

Langfuse, a platform for monitoring and analytics of applications using Large Language Models (LLMs), has announced that all features from its “Pro” version are now open source. This announcement, made on June 4, 2025, aims to empower developers and organizations with more flexibility for tracking and optimizing AI solutions. The core product is available on GitHub under the MIT License, excluding a few proprietary enterprise features. The open-source version includes key functionalities such as LLM-as-a-Judge evaluations and prompt experiments, previously only in the paid tier. Langfuse intends to enhance transparency, gain community feedback, and accelerate the advancement of LLM observability by removing commercial barriers. The open-source initiative allows organizations to self-host the platform, ensuring full control over data and privacy. Developers can anticipate faster updates and customization opportunities. More information and resources can be found in the GitHub repository and official documentation.

Source link

Phishing emails remain a critical cybersecurity threat, with attackers using increasingly sophisticated tactics that outpace traditional detection methods. Researchers at the University of Auckland have developed “MultiPhishGuard,” a multi-agent AI system designed to enhance phishing email detection. This innovative framework utilizes three specialized agents: a Text Analysis Agent, URL Analysis Agent, and Metadata Analysis Agent, each focusing on different components of emails. By employing reinforcement learning, MultiPhishGuard dynamically adjusts the importance of each agent’s input based on the email’s characteristics, significantly reducing false positives and increases accuracy. Additionally, an adversarial training module continuously tests and improves the system’s defenses against novel phishing tactics. The system also prioritizes transparency by providing clear, user-friendly explanations for its decisions, aiding user understanding and trust. Tested against various datasets, MultiPhishGuard achieved a high accuracy of 97.89%, showcasing its potential as a robust defense against phishing attacks in the evolving cybersecurity landscape.

Source link

Turning off Google’s Gemini AI is straightforward, allowing users to regain control over their Gmail, Docs, and Photos without tech complications. Many users appreciate the AI features for their convenience, while others prefer a more traditional interface. To disable Gemini in Gmail, simply navigate to the settings, locate “Google Workspace smart features,” and toggle off the primary switch. This action removes AI-generated extras like summaries and suggestions from your tools. For Google Photos, a separate toggle exists within its app settings to turn off Gemini. Users may choose to opt out due to privacy concerns or simply a preference for classic functionality. While disabling Gemini leads to the loss of some smart features, many find the trade-off worthwhile. It’s advisable to revisit settings occasionally, as Google may update menu locations. For ongoing tech tips, stay connected to Euro Weekly News.

Source link

Google has introduced a scheduling feature for its AI, Gemini, enhancing its capabilities to compete with similar offerings from rival platforms like ChatGPT. This new tool enables users to assign time-based tasks to Gemini, allowing it to automatically execute prompts at designated times. Available to subscribers of Google’s premium AI tiers and Workspace accounts, users can issue commands via text or voice to queue tasks. Similar to ChatGPT’s earlier feature, Gemini supports regular recurring tasks, thus facilitating automated reminders and updates. While ChatGPT’s scheduling is limited to a certain number of tasks and specific models, Google’s usage limits for Gemini have yet to be disclosed. This development indicates a competitive landscape where AI assistants are evolving from reactive tools into proactive participants in daily scheduling, blurring the lines between task management and automation. As these platforms improve their features, users may increasingly rely on them to enhance productivity and streamline personal management.

Source link

Apple will not announce any integration with Google’s Gemini AI platform at the upcoming Worldwide Developers Conference (WWDC) on June 9. Despite ongoing discussions, no agreement has been reached in time for the event. Instead, Apple is expected to focus on its anticipated Apple Intelligence platform, which will feature system-level generative AI across its devices, primarily utilizing OpenAI’s ChatGPT as its main AI service.

Reports suggest Apple’s reluctance to integrate Gemini may be influenced by strategic and regulatory concerns, particularly given the ongoing antitrust scrutiny faced by Google. While both companies’ CEOs have discussed a potential partnership, no formal deal is established. Apple’s current focus will be on its own AI technologies and privacy framework at WWDC, highlighting on-device processing and new tools for developers. The final keynote agenda will be released on June 10. Future collaborations with additional AI partners may hinge on evolving legal and market conditions.

Source link