Digital Phablet

Tag: assistant

  • Kagi’s AI Search Assistant Accesses Major Models in One Place

    Kagi’s AI Search Assistant Accesses Major Models in One Place

    Kagi has made its “Assistant” feature available to all subscription tiers, including those on a free trial. This feature allows users to access various language learning models for conversational interactions and web searches.

    If you’re unfamiliar with Kagi, it’s a paid search engine that takes its name from the Japanese word for “key.” The premise is straightforward: with Google, users trade their data and endure advertisements for a free service, whereas Kagi offers a private, ad-free experience in exchange for a subscription fee.

    While several open-source alternatives focus on data privacy, they often rely on ads and struggle to maintain consistent funding for improvements and development.

    Being a user-funded search engine, Kagi eliminates advertisements, safeguards your personal information, and continuously introduces new features that enhance your search experience.

    The homepage of Kagi resembles a typical search engine, featuring a prominent search bar. However, switching to Assistant mode transforms it into a chat interface powered by a language model. Users can select the AI model they prefer from a dropdown menu, along with a toggle to specify whether the model should access the web.

    When web access is disabled, users can engage with the chatbot in a conventional manner. If web access is turned on, the AI will leverage Kagi Search to provide answers, complete with hyperlinked references for additional context.

    Kagi search Assistant feature.
    Image credit: Kagi

    Kagi operates on a subscription model, where the benefits increase with each tier. The Ultimate plan, priced at $25 per month, offers access to the most advanced AI models along with limitless interactions and searches, targeting professionals and AI enthusiasts.

    On the other hand, most people might find Kagi’s $5 Starter plan sufficient, which allows for 300 searches and AI interactions each month. This plan includes access to several standard AI models, including:

    • GPT 4o mini
    • GPT 4.1 mini
    • Gemini 2.5 Flash
    • Mistral Pixtral
    • Llama 4 Scout
    • Nova Lite
    • DeepSeek Chat V3

    If you’re looking for unlimited searches and interactions but aren’t concerned about having the very best AI models, the Professional plan is available for $10 per month.

    The trial tier offers 100 searches and interactions through the Assistant feature, allowing users to sample the service. Kagi also follows a unique “Fair Pricing” model: if you do not use the service in a month, you won’t be charged, and your payment will be credited to the next month.

    While subscribing to a search engine might seem unusual at first, it’s essential to remember that these services have never really been free; they’ve been sustained by ad revenues. As the landscape evolves, tracking and data collection practices have become more prevalent. Personally, I’m curious about this service, and I plan to subscribe and explore what it has to offer.

  • AI Coding Assistant Comparison

    AI Coding Assistant Comparison

    In today’s rapidly evolving tech landscape, developers are continually seeking tools that can streamline their coding processes. AI coding assistants have emerged as crucial partners for programmers, offering functionalities like code generation, debugging, and optimization. With numerous options available, it’s essential to understand how they stack up against each other. This article provides a detailed comparison of some of the leading AI coding assistants.

    Key Features of AI Coding Assistants

    Understanding the core features of AI coding assistants helps developers choose the tool that best fits their needs. Here are the primary functionalities these tools typically offer:

    Code Generation

    • Automated Suggestions: Most AI coding assistants provide real-time suggestions as you type, helping to generate code snippets quickly.

    • Template Creation: Some tools can create complete templates based on user specifications, saving significant time during the initial setup of projects.

    Debugging Assistance

    • Error Detection: Many AI tools can automatically identify bugs and common issues in real-time.

    • Fix Recommendations: Beyond just identifying problems, some assistants offer contextual solutions or alternatives for the errors encountered.

    Language Support

    • Multiple Programming Languages: Leading AI coding assistants support a variety of programming languages such as Python, Java, JavaScript, C++, and more.

    • Framework Compatibility: Some tools are optimized for specific frameworks, enhancing their utility within those ecosystems.

    Popular AI Coding Assistants

    Here are some of the most recognized AI coding assistants currently available, along with their standout features:

    GitHub Copilot

    • Integration: Seamlessly integrates with popular editors like VS Code.

    • Contextual Awareness: Offers suggestions based on the code already written and can understand comments to create relevant code.

    Tabnine

    • Custom Models: Allows users to train custom AI models based on their own codebases for improved personalization.

    • Team Collaboration: Facilitates team settings, providing suggestions that align with shared coding styles.

    Kite

    • Documentation Assistance: Provides inline documentation that helps users understand code features and functions without leaving the editor.

    • Frequent Updates: Continuously updated to support new frameworks and libraries, keeping the tool relevant.

    Codeium

    • Free Access: Offers a robust free tier, making it accessible for individual developers and small teams.

    • Autocomplete Functionality: Excels in intelligent autocomplete features, enhancing coding speed drastically.

    User Experience and Usability

    When choosing an AI coding assistant, user experience plays a vital role. Here are some factors to consider:

    Ease of Use

    • Intuitive Interfaces: Most tools have user-friendly interfaces, allowing developers to easily navigate through options and suggestions.

    • Learning Curve: The best assistants minimize the learning curve, enabling users to become proficient quickly.

    Performance

    • Speed and Responsiveness: Assess how quickly the AI generates code suggestions or detects errors.

    • Accuracy: The effectiveness of auto-suggestions can significantly impact productivity; choosing a tool with high accuracy is crucial.

    Community and Support

    • Documentation and Tutorials: Comprehensive guides and tutorials enhance the onboarding experience for new users.

    • Active Community: A supportive user community can provide additional resources, tips, and troubleshooting assistance.

    Pricing Model

    Understanding the pricing of AI coding assistants is vital for budget consideration. Here’s a brief overview of how they typically structure their costs:

    • Free Versions: Many tools offer free tiers with limited features.

    • Subscription Plans: Most prominent services have monthly or annual subscriptions, often tiered based on the functionality offered.

    • Enterprise Licenses: Larger organizations may opt for enterprise solutions that offer advanced features and support for teams.

    Final Thoughts on Choosing the Right AI Coding Assistant

    Selecting the right AI coding assistant depends on various factors, including the specific needs of a developer or team. It’s important to evaluate the features, ease of use, performance, and pricing models of each tool against your unique coding requirements. By carefully considering these aspects, developers can find the ideal coding companion that will enhance their workflow and productivity.

  • Apple Confirms Long Delay For AI-Enhanced Siri Assistant

    Apple Confirms Long Delay For AI-Enhanced Siri Assistant

    While Apple continues to enhance its ecosystem with cutting-edge AI functionalities, it faces significant challenges, especially when compared to Google’s Gemini. A notable underperformer in this area has been the Siri virtual assistant, which has yet to incorporate the new features Apple announced at last year’s developer conference.

    Recently, the company confirmed that a revamped Siri, capable of accessing user data stored on devices and interacting with applications, has been postponed until next year. This next-generation version, referred to internally as “LLM Siri,” might not make its debut until 2024, with the possibility of further delays extending into 2027.

    “We’ve also been working on a more personalized Siri that possesses greater awareness of your personal context and can perform actions across your apps. Delivering these features is taking longer than we anticipated, but we expect to roll them out over the next year,” an Apple spokesperson told Daring Fireball.

    Activating Siri on an iPhone.
    Nadeem Sarwar / Digital Trends

    Initially, Apple had hoped to launch the updated Siri with the iOS 18.4 rollout in April. However, concerns about the implementation of the new features were raised by senior executives, including software engineering chief Craig Federighi, during internal evaluations, as reported by Bloomberg.

    Some employees speculate that the entire upgrade plan may be scrapped, requiring Apple to start anew. As a temporary workaround, the company has aligned with OpenAI to integrate ChatGPT functionalities on iPhones and Macs, which will work alongside Siri as part of Apple’s AI framework.

    While ChatGPT offers advanced language processing and multimodal features, such as image creation and file analysis, Siri still lacks essential capabilities that could significantly improve user experience. For instance, it currently cannot analyze user data to provide contextualized responses, nor can it engage deeply with third-party applications.

    Integration of ChatGPT and Siri on iPhone.
    Siri is currently leveraging ChatGPT capabilities. Nadeem Sarwar / Digital Trends

    In contrast, Google’s Gemini is now integrated into high-end Android devices and fully utilizes the data available across its Workspace tools, including Gmail, Docs, Maps, and YouTube. Additionally, it can translate voice commands into actionable tasks across various applications, even those developed by third parties like Spotify.

    Apple seems to be lagging behind in this dynamic landscape, and recent delays raise questions about the company’s innovation trajectory. Nevertheless, it is making strides in other areas, recently partnering with Alibaba and Baidu to accelerate the development of its AI systems for the Chinese market.

  • Pixel 10 Launches With Google’s Smartest Assistant

    Pixel 10 Launches With Google’s Smartest Assistant

    In 2023, whispers about a new Google AI assistant, codenamed “Pixie,” began circulating. This idea of having a smart assistant that operates directly on devices was attractive due to its potential benefits, such as enhanced privacy and efficiency, but it did not launch as scheduled. Google remained silent on the matter—until now. Pixie is returning, rebranded as Pixel Sense, and is expected to debut with the Google Pixel 10.

    According to Google, this new application aims to deliver a highly personalized experience based on user activity on their Pixel devices. Pixel Sense is designed to gather and organize information from various applications, as well as from text messages, images, and media files. It can even utilize screenshots similarly to the existing Pixel Screenshot feature.

    Despite the integration of Gemini in many Google applications, it currently can’t leverage the information from these platforms to provide contextual answers. However, with Pixel Sense, users can expect the assistant to respond to queries involving personal data. Importantly, this processing happens on the device itself, ensuring that sensitive information never reaches Google’s servers.

    Now Playing feature on Pixel device.
    Bryan M. Wolfe / Digital Trends

    So, what makes this possible now when it wasn’t before? The answer lies in Gemini Nano, a more streamlined version of the Gemini system. Picture it like this: a small bottle with a powerful genie. Coupled with the new Tensor chips in the Pixel 10, the device possesses the necessary capabilities to execute functions that previous models were unable to deliver.

    The Google Pixel 10 is expected to launch around August 2025. At this time, the smartphone appears to be on schedule for that release date; unless any significant setbacks occur, you could soon experience Google’s most advanced AI technology yet.

  • Google’s New Agent To Power Mercedes’ Next Car AI Assistant

    Google’s New Agent To Power Mercedes’ Next Car AI Assistant

    The Mercedes MBUX
    Mercedes-Benz

    On Monday, Google Cloud unveiled its latest Automotive AI Agent, designating the Mercedes-Benz CLA as the pioneer model to be equipped with this technology later this year. This new Agent will enhance the functionalities of Mercedes’ MBUX Virtual Assistant, allowing for a broader range of conversational interactions with passengers.

    Last year at CES 2024, we were introduced to the next-generation assistant from Mercedes, although details about the underlying large language model were not disclosed at that time. This upcoming assistant offers more advanced capabilities compared to the existing MBUX system, which can currently manage around two dozen commands and gather information via sources such as ChatGPT and Bing. Presently, the MBUX is triggered with the phrase “Hey, Mercedes,” but it operates more like Siri or Google Assistant rather than ChatGPT’s Advanced Voice Mode, delivering static responses rather than engaging in two-way conversation.

    The new Google Agent is powered by the Gemini large language model and utilizes Vertex AI, specifically designed to enable automakers in creating highly personalized and intuitive in-car assistants that transcend the limits of conventional vehicle voice commands. Notably, the Agent supports multimodal and multilingual input and can answer follow-up questions. For instance, it can inform drivers about nearby Italian restaurants, provide reviews of those establishments, and even highlight the most popular dishes available. Its robust design facilitates multi-turn dialogues, allowing it to recall details from previous interactions.

    Moreover, the enhanced MBUX assistant will source real-time information from Google Maps, delivering “fresh and factual data” about over 250 million points of interest worldwide, as well as current traffic conditions. “At Mercedes-Benz, we aim to deliver an outstanding digital experience to our customers,” stated Ola Källenius, CEO of Mercedes-Benz Group AG, in a press release. “Our collaboration with Google Cloud will significantly enhance in-car navigation by merging sophisticated location data with generative AI.”

    In the future, Mercedes plans to expand the rollout of the new MBUX assistant to other models, although they have yet to announce specific vehicles that will receive this upgrade.

  • Delta Unveils AI Assistant And 4K QLED Displays At CES 2025

    Delta Unveils AI Assistant And 4K QLED Displays At CES 2025

    CES 2025
    Explore our full coverage of CES here
    Updated less than 5 minutes ago

    Delta Air Lines has commenced its centennial celebrations with several exciting announcements at CES 2025.

    During a presentation in Las Vegas, CEO Ed Bastian unveiled initiatives aimed at enhancing customer experiences through premium offerings and personalized services.

    One highlight is the introduction of Delta Concierge, an AI-powered assistant integrated into the Fly Delta app. This feature is designed to empower travelers with greater control and confidence throughout their journeys.

    Similar to other AI-driven chatbots, Delta Concierge allows users to communicate using natural language text or voice commands. Travelers will soon be able to ask the assistant to book flights using just their voice. Additionally, the tool aims to anticipate customer needs, provide context-aware guidance, and take specific actions on their behalf. For instance, it could alert you if your passport requires renewal or inform you about entry requirements for international destinations. Customers will also receive packing suggestions based on local weather conditions.

    Furthermore, Delta has announced a collaboration with Uber to improve its services, ensuring that an Uber vehicle is waiting for passengers as they exit the airport terminal. Customers can also earn extra SkyMiles through their Uber rides.

    Additionally, Delta is set to launch a pioneering cloud-based in-flight entertainment system. YouTube will feature as part of the Delta Sync experience, offering passengers access to fresh, ad-free content.

    Starting with selected new aircraft in 2026, travelers will enjoy the updated Delta Sync experience characterized by 4K HDR QLED seatback displays. This development is welcomed, especially considering the issues many passengers encounter with standard displays during flights. The upgraded screens are supported by a 96-terabyte storage system, significantly exceeding the typical capacity of Delta’s in-flight entertainment offerings.

    Bluetooth capabilities will also be introduced across all cabin classes, enabling seamless pairing with wireless devices.

    Bastian updated attendees on Delta’s partnership with Joby, a leader in electric vertical takeoff and landing (eVTOL) aircraft. The Delta Concierge will link with Joby’s upcoming air taxi service, expected to connect Delta’s hubs in New York (LGA and JFK) and Los Angeles (LAX), pending approval from federal and local authorities.