Ai Product
PrunaAI Launches P-Video-Avatar with Multilingual TTS and Cinematic 1080p Output
PrunaAI's P-video-avatar model generates cinematic-quality video avatars up to 3 minutes long at 1080p resolution for $0.025 per second. It features built-in TTS supporting over 20 languages with voice selection, dynamic backgrounds, and body/camera controls. The model is now live on Replicate and f…
PrunaAI Launches P-Video-Avatar: Cinematic 1080p Avatars with TTS and 3-Minute Generations at $0.025/sec
PrunaAI's P-video-avatar model enables cinematic-quality video avatars with built-in TTS supporting 20+ languages and voice selection, dynamic backgrounds, and body/camera control. It generates up to 3-minute videos in 1080p resolution at a cost of $0.025 per second. The model is now live on Replica…
Perplexity Pro Users Gain Immediate Access to Model 5.5 on Desktop
Perplexity Pro subscribers can utilize the 5.5 model on computer platforms right now. This availability is highlighted in an hourly poll from Aravind Srinivas's X feed. It indicates early rollout prioritization for paid users on desktop before broader access.
Perplexity Builds Functional Google Earth-Flight Simulator Hybrid with GTA Aesthetics Using Subagents
Perplexity Computer has developed a fully operational Google Earth and Flight Simulator emulation featuring GTA-style visuals, powered by Codex and CC as integrated subagents. This marks a complete implementation of immersive geospatial simulation within Perplexity's AI framework. The project demons…
ChatGPT Dictation Persists Through Incoming Calls After Fix
ChatGPT's voice dictation now maintains transcript continuity when an incoming call interrupts the session. A fix has been implemented and confirmed, preventing transcript loss during such events. Deployment is scheduled for an upcoming weekly release.
Grok Imagine Agent Mode Beta Unifies Creative Workflow in Single Infinite Canvas
Grok's Imagine Agent Mode beta integrates brainstorming, writing, image generation/editing, and video creation into one seamless infinite canvas interface. Users access all features without page navigation on desktop via grok.com/imagine. Elon Musk promotes it hourly in his X feed, signaling strong …
Embiricos Launches Community Poll on Codex AI's Communication Style Preferences
Alexander Embiricos is running an hourly poll via his X feed to gather user feedback on Codex AI's personality and communication style. The poll solicits preferences for conciseness vs. explainability and directness vs. collaboration, along with specific examples of effective or poor interactions. R…
NotebookLM Introduces Auto-Labeling for Sources with 5+ Items
NotebookLM now automatically labels and categorizes sources when a notebook has 5 or more, reducing manual organization efforts. Users can rename, reorganize, and personalize these labels, including adding emojis. This feature streamlines source management to prioritize analysis and insight generati…
AI Platform Adds Autocomplete for Fast/Thinking Model Switching
An engineer implemented and landed a PR for autocomplete functionality enabling seamless switching between fast, thinking, and pro AI models. The feature was promptly recognized by user Josh Woodward, who plans to integrate it over the weekend. Deployment is underway via an imminent update, highligh…
AI Platform Enables Rapid User-Driven Feature Deployment for Model Switching
Developers at an AI platform respond swiftly to user requests for autocomplete in switching between fast, thinking, and pro models. A PR for fast model switching was recently landed and is being merged for the next update. This demonstrates real-time feature iteration based on user feedback, with on…
GPT-5.5 Launches in Cursor, Tops CursorBench at 72.8% with 50% Discount via OpenAI Partnership
GPT-5.5 is now accessible in Cursor and leads CursorBench with a 72.8% score. Cursor has partnered with OpenAI to provide the model at 50% discount until May 2. CursorBench serves as the evaluation benchmark for model performance in Cursor.
OpenAI Clones Notion's Custom Agents Amid Notion's Iterative AI Editor Dominance
Notion pioneered custom agents integrated with its powerful editor, prompting OpenAI to launch similar "workspace agents" in ChatGPT for complex, team-based workflows. OpenAI lacks an equivalent editor, limiting full replication. Notion refined its AI through five rebuilds, as detailed in recent pod…
Cohere Fine-Tunes Command Model with RWS to Power Language Weaver AI Translation
Cohere partnered with RWS Group to fine-tune its Command model, enhancing language and cultural expertise for translation tasks. The optimized model now serves as the core engine for RWS's Language Weaver AI solution. This integration stems from a strategic collaboration highlighted in industry disc…
Codex Memories Enhanced with Recent Desktop Screen Context via Chronicle
OpenAI's Codex now incorporates memories augmented by recent desktop screen context through the experimental Chronicle feature. This allows Codex to assist with ongoing work without requiring users to restate context. Currently pro-only, it builds on last week's memories preview.
Codex Gains Native Support for Over 90 Plugins Enhancing Tool Integration
OpenAI has integrated support for more than 90 plugins into Codex, expanding its ability to access context and execute actions. This enables seamless interactions with existing tools for document management, project tracking, code review, creative tasks, and deployments. The update broadens Codex's …
Replicate Offers $5 Seedance 2.0 Credits to First 1000 Claimants via X Feed Promotion
Replicate is running a limited-time promotion through its X feed, providing $5 credit for Seedance 2.0 to the first 1000 users who claim it. The offer is accessible via a specific invite link and emphasizes urgency with "Claim it before it's gone!" This targets rapid uptake among followers monitorin…
Scoble Shifts to AI-Powered Twitter Analysis Tool
Robert Scoble has pivoted from curating an hourly poll of his X feed to developing an AI that reads Twitter. The tool is accessible at alignednews.com/ai. This represents a transition to automated content processing for social media monitoring.
Perplexity Launches Hourly Polls on X Feed with Live YouTube Integration
Perplexity AI has initiated hourly polls on its X (Twitter) feed to engage users regularly. These polls link to a live YouTube stream for real-time interaction. This strategy enhances audience participation through frequent, time-sensitive content tied to ongoing broadcasts.
Perplexity's Personal Computer Enables Secure Local File Access and Native Mac App Integration
Perplexity's Personal Computer feature securely connects to any local folder, allowing search, read, and write operations on files. It extends access across native Mac applications including iMessage, Apple Mail, and Calendar. This provides seamless AI interaction with personal data and apps on macO…
Perplexity's Personal Computer Enables Seamless Cross-Device AI Operation on Local Mac Files
Perplexity's Personal Computer runs continuously in the background on Mac mini, accessing all apps and files 24/7. Users can initiate tasks from iPhone, with the AI executing on desktop local files via 2FA authentication. Deployment requires the latest iOS App Store update.
Dual-Agent Architecture Overcomes Voice UI Latency-Reliability Tradeoff
Vocal Bridge employs a dual-agent system with a foreground agent enabling real-time, low-latency voice interactions and a background agent handling complex reasoning, guardrails, and tool calls. This resolves the core tradeoff where fast voice models sacrifice reliability and agentic pipelines intro…
Perplexity's Personal Computer Prioritizes AI Orchestration on Existing Hardware Over New Devices
Aravind Srinivas argues that true AI-native computing requires an intelligent orchestrator managing devices, apps, and models across existing interfaces like phones and watches, rather than novel hardware form factors. This "orchestra conductor" evolves through continual learning of user workflows a…
Codex Expands to Native Mac App Integration with Advanced Memory and Task Automation
Codex now integrates with Mac apps, connects to additional user tools, generates images, learns from prior actions, adapts to user work preferences, and handles ongoing repeatable tasks. This brings nearly all blog-described features to Mac, excluding Windows-specific computer use which is forthcomi…
Universal App Tagging Revolutionizes Desktop Computer Interaction
The concept of tagging (@-mentioning) any application on a desktop computer, highlighted in context of Codex's desktop use, elicits strong enthusiasm as a transformative idea. It enables seamless, natural interaction with software akin to social media mentions. This innovation is praised for its imm…
Perplexity AI Launches Personal CFO with Plaid Integration for Comprehensive Financial Tracking
Perplexity AI's Computer now integrates with Plaid to connect users' bank accounts, credit cards, and loans. This enables detailed spending tracking, custom budget tool creation, and net worth visualization tied to investment portfolios. The feature positions Perplexity as a unified AI-driven person…
Simon Willison Seeks Transparency and Enhancements in ChatGPT Voice Mode Capabilities
Simon Willison expresses a desire for OpenAI to disclose the specific model powering ChatGPT's voice mode. He proposes integrating this voice model with background agents leveraging GPT-5 for complex tasks, including verbal cues like "let me think a moment." Additionally, he advocates for an overall…
Wikipedia Reverts Unsupported China Availability Date Change for Apple Intelligence to March 2026
A Wikipedia editor attempted to alter the Apple Intelligence page to state unavailability in mainland China "as of April 2026," but this was promptly reverted by another editor citing lack of reliable sources (WP:RS, WP:CITE). The prior text specifies "As of March 2026," indicating ongoing unavailab…
Apple Intelligence: Hybrid On-Device/Server GenAI with Iterative Rollouts, Partnerships, and Privacy Focus
Apple Intelligence integrates generative AI into iOS 18+, iPadOS 18+, and macOS Sequoia via on-device foundation models outperforming small rivals and server models matching GPT-4, using Private Cloud Compute for privacy. Key features span writing tools, image generation (Image Playground, Genmoji),…
Gemini 3.1 Flash-Lite Demonstrates Real-time Website Generation
Gemini 3.1 Flash-Lite showcases rapid, on-demand website creation. This capability allows for dynamic page generation as users interact, search, and navigate. The system's efficiency is highlighted by its ability to build pages in real-time.