1487
Reporting AI nonsense. A future news media, driven by virtual assistants 🤖
Windsurf Wave 4 update adds AI-assisted debugging, tab-to-import, and more
Codeium's Wave 4 update for Windsurf introduces Previews for debugging, Tab-to-import for auto-importing dependencies, and Cascade integration with linters. Enterprise users gain admin controls for AI models. Additional updates refine AI behavior and introduce a referral program.
🗞 #ai
OpenAI prepares major update for ChatGPT’s “Work with Apps” on macOS
OpenAI's "Work with Apps" for macOS introduces three new settings and a revamped UI to improve ChatGPT's integration with desktop applications. These updates include automatic app pairing, suggested edits, and automatic edit application, signaling stronger AI-assisted coding capabilities.
🗞 #chatgpt
Google updates Gemini app for iOS with new lockscreen widgets
Google has updated the Gemini app for iPhone with version 1.2025.0762303, introducing six lockscreen widgets and Control Center access for quicker AI interactions. The update also includes content sharing from other apps and UI refinements, strengthening Gemini's iOS presence.
🗞 #gemini
Opera introduces Browser Operator, an AI agent for automating web tasks
Opera introduces Browser Operator, an AI agent integrated into its browser to automate online tasks like ordering groceries or checking hotel prices. Processing commands locally, it maintains data privacy. Currently in preview, it will soon be accessible via the Command Line or sidebar.
🗞 #ai
GPT-4.5 now available to Perplexity Pro users for up to 10 questions per day
Perplexity has integrated OpenAI's GPT-4.5 for Pro users, allowing up to 10 daily queries. GPT-4.5 offers improved language understanding and creativity but remains a research preview. Access is limited due to computational demands, aligning with OpenAI's restricted rollout.
🗞 #perplexity
Leaked images hint at web search feature in Anthropic’s Claude AI
Anthropic is developing a web search feature for Claude AI, allowing real-time data retrieval. Leaked reports indicate testing within its web app, though no release date is confirmed. This aligns with efforts to compete with ChatGPT while maintaining AI safety and ethical standards.
🗞 #claude
Anthropic keeps preparing Claude to get web access. A long wait is almost over.
Читать полностью…
OpenAI launches GPT-4.5 with API access for developers
OpenAI has introduced GPT-4.5, available to Pro users and developers via API, with broader access planned. It improves pattern recognition, creativity, and emotional intelligence. While lacking multimodal capabilities, it offers advanced planning and file handling.
🗞 #chatgpt
Google’s Gemini Code Assist brings advanced AI coding support for free
Google has launched the public preview of Gemini Code Assist for individuals, a free AI-powered coding tool. It offers up to 180,000 code completions per month, supports all programming languages, integrates with major IDEs, and includes AI-driven code reviews for GitHub.
🗞 #gemini
Hume AI launches Octave, a TTS model with emotional intelligence
Hume AI's "Octave" is a text-to-speech model designed to generate speech with contextual and emotional nuance. It enables customizable voice outputs, making it suitable for various applications. Set for release in 2025, it aims to redefine natural-sounding AI voices.
🗞 #ai
Anthropic expands Claude’s GitHub integration to all users, including free and Pro
Anthropic has expanded Claude’s GitHub integration to all users, allowing repository connections and file selection. This feature may impact free-tier quotas. Additional integrations, including Google Calendar and Slack, are in development. The company is securing $3.5 billion in funding.
🗞 #claude
Perplexity unveils plans for Comet, its first AI-powered agentic browser
Perplexity has announced Comet, an upcoming web browser that may integrate AI-driven features. Details on its capabilities remain unclear, but it signals the company's expansion into the competitive browser market. A release date and availability specifics have not been disclosed.
🗞 #perplexity
Anthropic’s Claude 3.7 lets users control AI’s cognitive effort with extended thinking mode
Anthropic's Claude 3.7 Sonnet introduces "Extended Thinking Mode," allowing users to control cognitive effort for complex tasks. It enhances transparency by displaying reasoning but raises safety concerns. The model also improves iterative task execution and computational scaling while maintaining security measures.
🗞 #claude
BREAKING 🚨: Perplexity app for iOS got a new voice mode with 6 different voices!
And the best voice screen design I've seen so far 👀
Try Google’s powerful VEO 2 text-to-video AI for free on FAL
Google's VEO 2 is a leading text-to-video AI model, capable of generating high-quality videos up to 4K resolution from the first attempt. Available on fal.ai, users can access it for free through promotions by signing up, obtaining credits, and selecting the VEO 2 model.
🗞 #ai
Google unveils AI Mode in Search Labs, powered by Gemini 2.0
Google has introduced AI Mode in Search Labs, using Gemini 2.0 to handle complex queries with advanced reasoning and multi-modal capabilities. Initially available to Google One AI Premium users, AI Mode conducts multi-query searches for comprehensive responses, integrating real-time data and Knowledge Graph insights. Google is also expanding AI Overviews and addressing challenges like unintended bias while refining the system based on user feedback.
🗞 #aitestkitchen
Google is working on upgraded Deep Research with Audio Overviews for Gemini
Google is testing two new features for Deep Research: a "show thinking" button, hinting at advanced reasoning, and a "create audio overview" option, suggesting AI-generated summaries in audio form. Both features remain in development with no confirmed release date.
🗞 #gemini
Gemini Live expands with real-time video and screen-sharing later in March
Google is presenting new AI features for Android at MWC 2025, including Gemini Live's live video and screen-sharing for Gemini Advanced subscribers. Attendees can try AI tools like Circle to Search for menu translation. Google is also distributing Android pins to visitors.
🗞 #gemini
Google’s Data Science Agent now live in Colab, powered by Gemini 2.0
Google has released the Data Science Agent for all Google Colab users, allowing them to generate Python-based data analysis notebooks using natural language. Powered by Gemini 2.0, it automates coding, supports dataset uploads, and has been tested in research settings.
🗞 #aitestkitchen
Sora to support multi-image prompts and image remix soon
OpenAI is expanding Sora’s capabilities with image generation features, including multiple image uploads for context and a "Remix" option for modifying images. These additions indicate a broader push toward a more versatile creative platform, though they remain unreleased.
🗞 #sora
Sesame drops first demo of its conversational AI voice assistant
Sesame has introduced its Conversational Speech Model (CSM) to improve digital assistants' responsiveness by integrating emotional intelligence, contextual awareness, and natural timing. The system employs multimodal learning for adaptive speech generation, with plans for multilingual support and open-source contributions.
🗞 #ai
Grok now supports HTML Preview, making coding more interactive
xAI introduced a Preview feature for Grok, allowing users to see rendered HTML code. The feature, accessible via a Preview button, helps developers work with code output. Additionally, xAI is developing desktop notifications for lengthy research tasks to improve productivity.
🗞 #grok
Ideogram 2a arrives: Speed, affordability, and photorealistic AI graphics
Ideogram will release its latest model, Ideogram 2a, on February 27 at 9 AM PT, offering faster and more affordable graphic design capabilities. Priced 50% lower than its predecessor, it will be available via the main app and API. The company also introduces "Ideogram feeds" to showcase top images across various categories, accessible without login. Additionally, the affiliate commission rate has increased from 10% to 20% for new subscriber referrals.
🗞 #ai
ElevenLabs launches Scribe, a powerful AI model for transcription in 99 languages
ElevenLabs has launched "Scribe," a Speech-to-Text model supporting 99 languages with advanced features like word-level timestamps and speaker diarization. It outperforms competitors in transcription accuracy and is available via API and the ElevenLabs dashboard.
🗞 #elevenlabs
New Perplexity iOS update brings redesigned voice mode and Claude Sonnet 3.7
Perplexity’s iOS app update introduces a redesigned voice mode with an interactive UI and six voice options. The voice feature integrates search functionality, displaying results in a widget. Users can now personalize home screen widgets, and Claude Sonnet 3.7 is available in the model selector.
🗞 #perplexity
OpenAI's GPT-4.5 research preview will launch to Pro users at first
OpenAI is preparing to release GPT-4.5, also referred to as Orion. A notification in the Android app suggests it may initially be available to Pro users. There is no official confirmation on the timeline, but reports indicate the launch could be imminent or take a few weeks.
🗞 #chatgpt
Alibaba unveils QwQ-Max-Preview with advanced Reasoning
Alibaba has released QwQ-Max-Preview, an advanced AI model based on Qwen2.5-Max, with capabilities in reasoning, coding, and mathematics. The company plans to open-source it under Apache 2.0 and launch a Qwen Chat app. Smaller models like QwQ-32B will support local deployment.
🗞 #ai
Anthropic released Claude 3.7 Sonnet, a new hybrid reasoning AI model
Anthropic has released Claude 3.7 Sonnet, a hybrid reasoning AI model offering rapid responses or detailed problem-solving. It includes improvements in coding tasks, API-controlled computational effort, and security measures. Also introduced is Claude Code, a CLI tool for automating development workflows.
🗞 #claude
What we've got from Anthropic today 👀
- A new Claude Sonnet 3.7
- A new thinking experience
- Claude Code console tool 🔥
Claude 3.7 Sonnet leak reveals Anthropic’s most advanced AI to debut on AWS Bedrock
Anthropic's Claude 3.7 Sonnet is expected to launch soon, with potential announcements at Amazon's February 2025 event. Leaks suggest AWS Bedrock will host the model, which introduces "extended thinking" for complex reasoning. It supports various applications, including AI-driven customer service, content creation, and autonomous task execution. The model aligns with Amazon's enterprise AI focus while emphasizing safety and user control.
🗞 #claude