1487
Reporting AI nonsense. A future news media, driven by virtual assistants 🤖
OpenAI o1 launches on API with structured outputs and vision tools
OpenAI introduces the o1 model with upgrades in reasoning, cost efficiency, and performance, alongside Realtime API updates featuring WebRTC support and cost reductions. A new Preference Fine-Tuning method and SDKs for Go and Java enhance integration and customization options.
🗞 #chatgpt
Google develops Gemini Creative Partner tools for image editing and animation
Google is advancing its Creative Partner Solution within Gemini, enabling users to edit images via the UI. Key tools include resizing, area selection, manual edits with a brush, and one-click animation, with potential integration into web and AI Studio platforms.
🗞 #gemini
OpenAI’s ChatGPT now offers web search for free worldwide
OpenAI has expanded ChatGPT’s web search to all logged-in users, including free-tier access, enabling global web searches on desktop and mobile. The update also adds maps integration, providing real-time location details via Foursquare's Places API.
🗞 #chatgpt
Veo 2 and Imagen 3.1: Google’s AI models transform video and image production
Google unveiled Veo 2 and Imagen 3 AI models, focusing on advanced video and image generation. Veo 2 supports 4K video creation, while Imagen 3 improves image detail and style versatility. Both are integrated into tools like VideoFX and ImageFX for creative use.
🗞 #aitestkitchen
Grok AI slashes API costs, boosts instruction following with grok-2-1212
xAI, led by Elon Musk, has released Grok-2-1212 with reduced API costs to boost adoption. The model features faster responses, real-time web searches, multilingual support, and citation generation, aligning with xAI’s focus on advancing AI-driven social media analysis.
🗞 #grok
Personalized search in ChatGPT: Memory integration under development
OpenAI plans to introduce "memories" in ChatGPT, enabling it to retain user preferences and contexts across sessions for personalized responses. Users can manage what is remembered, with potential implications for tailored search results and targeted advertising.
🗞 #chatgpt
NotebookLM gets interactive Audio Overviews and redesigned UI
Google's NotebookLM updates include a redesigned interface with Sources, Chat, and Studio panels, a beta Interactive Audio Overviews feature for real-time AI-host interaction, and NotebookLM Plus, offering advanced tools, team collaboration, and enterprise-grade security.
🗞 #notebooklm
OpenAI adds Santa voice along with video support to ChatGPT Voice Mode
OpenAI has introduced Santa Mode and video/screen-sharing features to ChatGPT. Santa Mode, active globally until Jan 2025, offers an AI Santa voice feature. Video/screen sharing is available for select users, with broader access expected in 2025, excluding certain regions.
🗞 #chatgpt
Scheduled prompts feature may soon roll out to Gemini users
Google is developing a prompts feature that allows users to schedule and manage reminders, with settings for time, frequency, and activation. Currently in testing, it may launch alongside other new features expected from Google in December.
🗞 #gemini
Google introduces Gemini Deep Research with chain-of-thought reasoning
Google's Gemini introduced "Deep Research," a feature using Gemini 1.5 Pro for detailed research with reasoning. It presents editable plans before execution, processes over 50 sources, and competes with tools like Perplexity, prioritizing thoroughness over speed.
🗞 #gemini
Gemini Flash 2.0 launches on AI Studio and Gemini on the web
Google launched Gemini Flash 2.0, now available on Gemini and AI Studio as an "experimental" tool. It supports image input, advanced tools, and a 1M token context window. Outperforming Gemini 1.5 Pro and Anthropic's Claude 3.5, it excels in speed, quality, and coding tasks.
🗞 #gemini
ChatGPT Canvas now supports Python code execution
OpenAI's Canvas, integrated into GPT-4o, offers a dedicated workspace for writing and coding, now available to all users. Features include Python code execution, integration with custom GPTs, advanced editing tools, and real-time feedback across web and desktop platforms.
🗞 #chatgpt
Premium users on X gain photorealistic image generation with xAI’s Aurora
Elon Musk's xAI integrated photorealistic image generation into its Grok assistant Aurora model. While praised for speed and realism, limitations in resolution, features, and ethical safeguards raise concerns compared to competitors like OpenAI's DALL-E.
🗞 #grok
Google advances image editing tools with Creative Partner for Gemini
Google appears to be refining its Creative Partner solution for Gemini, potentially introducing image editing with inpainting. Additionally, Gemini 2.0 Flash was briefly spotted, sparking speculation about its upcoming release following tests aligning it with a prior model.
🗞 #gemini
Google Illuminate expands AI podcasts with voice customization
Google's Illuminate update allows users to select hosts for podcasts, add any URL as a source, and play audio alongside transcripts. Features include Q&A chat and public sharing links, broadening its functionality for creating customizable podcasts efficiently.
🗞 #aitestkitchen
ChatGPT to soon support task scheduling with new beta feature
OpenAI's upcoming Tasks feature, codenamed Jawbone, will let users schedule prompts up to two years in advance with options for repetition. Integrated into ChatGPT, it offers a UI for managing tasks and push notifications, competing with Google's similar Scheduled Prompts.
🗞 #chatgpt
Batch Generation now available to Ideogram Pro subscribers
Ideogram has introduced a Batch Generation feature for Pro users, enabling bulk image creation through CSV or Excel uploads. With up to 12,000 image generations monthly, this tool targets professionals needing efficient, high-volume workflows.
🗞 #ai
Whisk AI tool by Google empowers creators to remix images using visual inputs
Google's new AI tool, Whisk, lets creative professionals generate and remix images using visual prompts. Powered by Gemini and Imagen 3 models, it focuses on rapid ideation, combining elements from user-supplied images for unique visuals via Google Labs in the U.S.
🗞 #aitestkitchen
Google developing canvas editing and Drive sync for Gemini Deep Research
Google is introducing a canvas editing feature in Gemini's deep search, allowing direct output modifications. A sync button hints at integration with Google Drive. Other updates include scheduled prompts and an image editor, potentially tied to Gemini 2.0's multi-model capabilities.
🗞 #gemini
Pika Labs launches Pika 2.0 featuring image to video Scene Ingredients
Pika Labs launched Pika 2.0, an AI video generation model featuring "Scene Ingredients," enabling users to upload images for customized, coherent scenes. Accessible on Pika.art, it offers free basic features and advanced options through paid plans.
🗞 #ai
Perplexity AI adds domain-specific search option to Spaces
Perplexity AI's new feature on Spaces lets users refine searches by selecting specific web domains or files, improving result precision. Accessible via a new menu, it supports personalized research needs by focusing on relevant sources.
🗞 #perplexity
OpenAI rolled out Projects feature in ChatGPT for simpler chat management
OpenAI's "Projects" for ChatGPT lets users organize chats, files, and instructions into workspaces for streamlined task management. Available for Plus, Pro, and Team users, it expands to Enterprise and free tiers by 2025, supporting efficient organization and reference use.
🗞 #chatgpt
Gemini 2.0 brings real-time voice and vision capabilities to AI Studio
Google AI Studio introduces Stream Real-Time, enabling voice-to-vision interaction with Gemini 2.0 Flash, plus diverse output voices. Updates also include Starter Apps like Map Explorer and Video Analyzer, blending developer tools with mobile-accessible consumer features.
🗞 #aistudio
Trusted testers invited for Astra, Jules, and Marine trials by Google
Google announced three projects for trusted testers: Project Astra (camera-powered Gemini Live integration), Jules (AI coding assistant with GitHub access), and Project Mariner (Gemini-controlled Chrome extension). Broader releases are anticipated in 2025.
🗞 #gemini
Gemini 2.0 brings real-time voice and vision capabilities to AI Studio
Google AI Studio introduced Stream Real-Time, enabling multi-modal inputs with voice and vision capabilities via Gemini 2.0. It also added Starter Apps like Map Explorer and Video Analyzer, showcasing practical uses of the Gemini API for developers and consumers alike.
🗞 #aistudio
DeepSeek unveils upgraded AI model and real-time search tool
DeepSeek introduced its updated model "DeepSeek-V2.5-1210" with improved performance and a real-time web search feature. This tool aggregates data from multiple sources, including Chinese content, though it lacks widgets like weather updates or stock charts for now.
🗞 #ai
Development underway for custom styles in Claude mobile apps
Anthropic’s app update introduces condensed UI elements, moving project selection to the bottom and adding custom writing styles. Style selection visibly applies to conversations. Claude supports tasks like brainstorming, coding, and more, compatible with multiple file formats.
🗞 #claude
OpenAI launches AI-powered video generator Sora for creators
OpenAI has launched Sora, an AI video generation model, accessible via ChatGPT Plus and Pro plans. The tool allows text-based video creation, remixing, and advanced editing, targeting creators and educators. Currently available in 137 countries outside the EU and UK.
🗞 #chatgpt
Aurora debuts in Grok, temporarily pulled after first rollout
xAI launched Grok 2 + Aurora, adding in-house image generation to its core. Initial features were briefly available before being suspended, likely for testing. Expanded access for free users began, with rumors of Grok 3 and feature updates expected by December’s end.
🗞 #grok
OpenAI hints Projects, a new way of managing chats on ChatGPT
OpenAI is developing a "Projects" feature to help users organize chats, files, and custom instructions in one place. Built on custom GPTs, this long-requested addition could launch during the 12 Days of OpenAI event, improving chat management functionality.
🗞 #chatgpt