1487
Reporting AI nonsense. A future news media, driven by virtual assistants 🤖
OpenAI’s Operator set to bring autonomous AI tasks to ChatGPT macOS app
OpenAI's upcoming "Operator" feature for the ChatGPT macOS app aims to function as an autonomous AI agent, automating tasks like coding and web browsing. Launching as a research preview in Jan 2025, it aligns with the industry's shift toward agent-based AI systems.
🗞 #chatgpt
Kling AI’s “Elements” in early testing to combine images into cohesive videos
Kling AI's "Elements" feature, in early access, enables creators to merge up to four images for coherent video creation with consistent characters and settings. Initially exclusive to select partners, it aligns with the platform’s goal of advancing professional content tools.
🗞 #ai
Le Chat users to gain access to AFP’s multilingual news archive in new partnership
Mistral AI partners with AFP to integrate multilingual news content into its Le Chat AI assistant, enhancing factual reliability. This marks Mistral's first media deal, positioning it as a competitor in generative AI while diversifying AFP’s revenue streams.
🗞 #mistral
Google’s NotebookLM experiments with AI-powered interactive mind maps
Google is developing an AI-driven mind map for NotebookLM, enabling users to visualize sources, click for related discussions, and export data. Currently in testing, it aligns with Google's focus on integrating AI into productivity and knowledge management tools.
🗞 #notebooklm
OpenAI updates Custom Instructions with new UI and personalization features
OpenAI updated Custom Instructions in ChatGPT with a redesigned interface and expanded options for personalization, allowing users to specify preferences like name, profession, or desired traits. The rollout started on January 17, 2025, excluding some regions initially.
🗞 #chatgpt
Year in review: key ChatGPT features released in 2024
The blog reflects on ChatGPT's milestones in 2023, from the introduction of GPT Store and advanced voice features to expanded vision and memory capabilities. Key highlights include updates during OpenAI’s “12 Days of OpenAI” and the Pro Plan's debut in December.
🗞 #chatgpt
Testing Cohere’s AI: Custom zero-shot tools and exclusive features
Cohere, an AI platform for enterprise, offers tools like customizable zero-shot capabilities, embedding jobs via API, and a no-code chatbot dashboard. It includes a Chat UI with model selection options, though some features remain restricted or under development.
🗞 #ai
Grok AI on web launches in Australia with web search and more
Elon Musk's xAI expanded Grok with a web app in Australia following its iOS release. Features include image creation, web/X post search, and a PDF viewer. Modes and privacy options are offered, with issues addressed via a Discord channel as it contends with leading AI chatbots.
🗞 #grok
Perplexity to add support for Gemini 2 and O1 models
Perplexity's updates include potential pro user access to OpenAI's O1 model with customizable reasoning, testing of Gemini model integration, a recency filter for space searches, an updated spaces tab in iOS/Android, a language selector for dictation, and a free pro trial offer.
🗞 #perplexity
Upcoming PRO plan hints at new features for Mistral AI
Mistral AI is updating its UI by relocating features and introducing tools like processing time display. A PRO plan with extended benefits is in progress, accompanied by usage limit trackers. New features include Notion integration, code refactoring support, and navigation updates.
🗞 #mistral
o3 model excels in math, science, and ethical AI reasoning
OpenAI introduced the o3 and o3-mini models, advancing AI reasoning with capabilities in complex tasks like coding and science. The o3 features simulated reasoning, while o3-mini offers faster, cost-efficient processing. Improved safety via deliberative alignment is included.
🗞 #chatgpt
Google unveils Gemini 2.0 Flash Thinking model with transparent reasoning
Google's Gemini 2.0 Flash Thinking is an experimental AI model focusing on transparent reasoning. It showcases its thought process step-by-step, handles multimodal inputs, integrates tools, and delivers fast, high-quality outputs. Available in AI Studio and via API.
🗞 #aistudio
Gemini rolled out model selector and new features on iOS
The Gemini iOS update introduces a model selector, access to Gemini Flash 2.0, code folder upload for all users, and gradual rollout of features like Google Home integration. Future updates, including reasoning mode and Gemini Live, are anticipated.
🗞 #gemini
ChatGPT for macOS adds Projects, Map Options, and teases vision-based Voice Mode
ChatGPT for macOS now supports Projects but only allows browsing, creating new chats, viewing settings, and accessing files. A new map provider option is introduced, and advanced voice mode with vision is under testing but remains slow and incomplete.
🗞 #chatgpt
X to roll out AI-powered image generation in post composer
X has integrated AI features into its platform via Grok, including image generation in the post composer using the Aurora model and a "More about this account" tool on hovercards, offering visual content creation and account context directly within the app.
🗞 #grok
Google may launch Gemini 2.0 Flash-Thinking-Exp-0123 on Jan 23
Google may introduce the Gemini 2.0 "Flash Thinking Exp-0123" AI model on January 23, 2025. Observed during a hackathon livestream, it likely advances decision-making efficiency, aligning with Google's AI Studio strategy for broader AI model applications.
🗞 #aistudio
Microsoft debuts AI-powered search for Windows 11 on Copilot+ PCs
Microsoft introduces AI-driven natural language search in Windows 11 for Snapdragon-powered Copilot+ PCs. The feature processes queries locally via NPUs, supports limited file types, and will expand to cloud storage. Rollout excludes non-Copilot+ hardware users.
🗞 #ai
Mistral AI reportedly developing memory feature for Le Chat assistant
Mistral AI plans to introduce a "Memories" feature in its chatbot, Le Chat, allowing it to retain past conversations. Likely tied to a Pro subscription, it addresses long-term recall limitations in AI but raises privacy questions. No release date is confirmed yet.
🗞 #mistral
o3-mini updates from OpenAI: Release date, features, and subscriber access
OpenAI is set to launch o3-mini in two weeks. The model emphasizes speed over performance compared to o1 pro, with simultaneous API and ChatGPT availability. Targeted at broad applicability, it will be initially accessible to OpenAI Plus subscribers.
🗞 #chatgpt
Minimax unveils T2A-01-HD with voice cloning and emotional intelligence features
MiniMax introduces the T2A-01-HD under Hailuo Audio HD, offering voice cloning, a 300+ voice library, emotional intelligence, and multilingual support in 17 languages. Available for trials and API integration, it targets diverse applications across industries.
🗞 #ai
Grok by xAI unveils updates for personalized AI experience on X
xAI’s Grok is gaining updates to boost personalization, including a “Grok Personalization” setting for user data preferences, follow-up suggestions on posts, and text file uploads. These changes enhance Grok’s adaptability and integration on X’s platform.
🗞 #grok
DeepSeek is preparing Deep Roles and released top rank V3 model
DeepSeek v3, with 671B parameters, triples the speed of its predecessor, outperforms competitors in math and coding tasks, and remains open source. Available on Hugging Face, it introduces "Deep Roles," a feature for creating and sharing customizable prompts.
🗞 #deepseek
Character.AI tests gaming features and Group Chats on web
Character AI is testing games like "Speakeasy" and "Say the Same Word," adding a game button and scores feature. It also plans to expand group chat to the web, currently a mobile-exclusive feature, highlighting efforts to broaden platform functionality and utility.
🗞 #characterai
Hume AI launches Octave, bringing dynamic voices to developers
Hume AI unveiled Octave, a speech-language model enabling real-time creation of customizable, emotionally expressive voices. Designed for applications like healthcare and education, it supports tone, pitch, and rhythm adjustments, integrating with empathic AI tools.
🗞 #ai
iOS Grok app launches in Australia amid web app speculation
The anticipated Grok web app at grok.com appears imminent, with a "Coming Soon" placeholder and references to "Grok 2.5" in the code. Meanwhile, a standalone iOS app is now available in Australia, offering image creation, uploads, and web search features.
🗞 #grok
OpenAI boosts ChatGPT with app integration in voice mode
OpenAI updates include voice command support for app tasks, conversation search, integration with note-taking and coding tools, customizable toolbars, and a sticky code copy button. Features aim to improve workflows and usability, with regional restrictions on voice features.
🗞 #chatgpt
ChatGPT connects users through 1-800 number and WhatsApp
OpenAI now offers ChatGPT via phone calls in the U.S. and Canada or WhatsApp globally. Calls allow 15 free minutes monthly, while WhatsApp supports text-only chats. This initiative increases AI accessibility, especially for users with limited internet or devices.
🗞 #chatgpt
Claude working on artifact screenshot feature for web users
Claude has introduced updates, including direct screenshot capture of artifacts, math.js integration for advanced computations, Excel file analysis up to 30MB, and mobile data analysis. A project creation button and style selector are also being tested for wider release.
🗞 #claude
OpenAI developing React rendering and text Transform for ChatGPT Canvas
OpenAI is advancing ChatGPT Canvas with new features: React rendering for visualizing React components, aiding developers in prototyping, and a "Transform" option for reformatting text into lists or tables. Both are in testing but aim to enhance coding and editing efficiency.
🗞 #chatgpt
Gemini 2.0 Experimental Advanced launches for Gemini Advanced users
Google has launched Gemini-Exp-1206, an experimental AI model for Gemini Advanced users. This tool addresses complex coding, math, and reasoning tasks. Though experimental and with noted limitations, it underscores Google's push for advanced AI solutions in competitive spaces.
🗞 #gemini