Sam Altman-Backed Coco Robotics Raises $80 Million

2025-06-12

Coco Robotics, a Los Angeles-based startup developing emission-free delivery robots, raised $80 million in funding, bringing its total to over $120 million. Backers include Sam Altman, Pelion Venture Partners, and Offline Ventures. Since 2020, Coco's robots have completed over 500,000 deliveries for partners like Subway and Wingstop. The company also partnered with OpenAI to exchange technology and data. Founded in 2020 by Brad Squicciarini and Zach Rash.

Read more

Meta Releases J-VEPA 2 AI Model for Video Understanding

2025-06-12

Meta has unveiled J-VEPA 2, an AI model that enhances robots' and AI agents' understanding of the physical world by interpreting video data. Building on its predecessor, J-VEPA 2 enables AI to predict outcomes and plan actions, reducing real-world trials. This advancement aids in tasks like object manipulation, navigation, and accident prevention. Meta also introduced three new benchmarks for evaluating reasoning models.

Read more

Apple Enhances Image Playground via ChatGPT

2025-06-12

Apple's Image Playground faced criticism for limitations and low-quality results after launch. This year at WWDC, Apple announced enhancements by integrating ChatGPT, enabling advanced AI-generated imagery in new styles like oil painting, watercolor, and anime. Users can now choose an "any style" option for precise descriptions. Image Playground sends user inputs to ChatGPT with consent, expanding its capabilities. The updated app will launch with iOS 26 this fall, positioning it as a stronger c

Read more

Google Releases Android 16: New Features and What's Missing

2025-06-12

Google has launched Android 16, one of the earliest major OS updates in recent years, available first on Pixel devices. Key features include real-time notifications, enhanced security with Advanced Protection Mode, improved multitasking for tablets/foldables, better hearing aid support, and delayed Material 3 design updates until September. The update will expand to other devices later this year.

Read more

OpenAI Updates ChatGPT's Voice Mode for More Natural Conversations

2025-06-10

OpenAI has upgraded ChatGPT's voice conversation mode, making it more natural with improved tone modulation, pacing, and emotion expression. The feature supports easier language translation and is available to paying users across platforms. However, minor audio quality issues may still occur, and voice-related hallucinations remain unaddressed.

Read more

Run the Complete DeepSeek-R1-0528 Model Locally

2025-06-10

DeepSeek's latest R1 model (DeepSeek-R1-0528) requires 715GB disk space but can be reduced to 162GB using Unsloth's quantization. This tutorial covers setting up Ollama, downloading the quantized model, and running it with GPU/CPU configurations. Minimum requirements include 24GB GPU or 64GB RAM for CPU-only operation, enabling efficient local deployment of large language models.

Read more

Apple Unveils Elegant Glass-like User Design Experience and Powerful New OS Features

2025-06-10

Apple unveiled a major software redesign with Liquid Glass for a translucent UI, enhancing focus and interaction across devices. iOS 26 introduces dynamic lock screens, Live Translation, and Visual Intelligence for richer experiences. iPadOS 26 advances multitasking with a new windowing system and Exposé for task management. These updates unify and elevate user experiences across Apple's ecosystem.

Read more

Apple Launches New Smart Fitness Assistant for More Efficient Workouts

2025-06-10

Apple launched Fitness Companion, an AI-powered personal trainer, at WWDC 2025. It analyzes workout data to offer real-time guidance and integrates with Apple Intelligence for motivational coaching via synthetic voices modeled after Fitness+ trainers. Updates to the Fitness app include encouraging phrases, milestone tracking, and comprehensive summaries post-workout. The tool prioritizes user privacy and introduces a new Apple Watch layout with one-tap access to custom workouts and enhanced musi

Read more

From Spatial Widgets to Realistic Avatars: All visionOS Updates Announced by Apple at WWDC

2025-06-10

Apple released visionOS 26 for its mixed-reality headset, introducing customizable spatial widgets, enhanced avatars with volumetric rendering, and shared immersive experiences for consumers and enterprises. Updates include AI-driven 2D image depth, eyes-only mode for privacy, new languages, and improved accessibility features like eye navigation and call handover. Logitech's Muse accessory further boosts precision in 3D collaboration.

Read more

OpenAI Claims Annual Revenue Hits $10 Billion

2025-06-10

OpenAI reports $10 billion in annual recurring revenue, up from $5.5 billion last year, driven by consumer products, ChatGPT solutions, and its API platform. With 500 million weekly users and 3 million enterprise clients, it targets $125 billion in revenue by 2029. Despite significant spending on talent and infrastructure, operational costs and profitability details remain undisclosed.

Read more

Amperity Launches First AI Agent for Data Engineers

2025-06-10

Amperity launches Chuck Data, an AI agent trained on customer data from 400 enterprise brands. It automates tasks like identity resolution and PII tagging using natural language prompts, operates within Databricks via zero-copy architecture, and leverages Amperity's patented Stitch technology. This tool aims to streamline customer data engineering by eliminating manual coding and offering free trials for smaller datasets with paid plans for larger ones.

Read more

Apple Downplays AI Hype While Showcasing Next-Gen iPhone and Mac Features

2025-06-10

Apple's WWDC 2025 showcased a redesign of macOS and iOS with Liquid Glass, modest AI advancements via Foundation Model Framework, and updates across iPhone, iPad, Mac, Apple Watch, and Vision Pro. Notable features include Messages enhancements, Real-time translation, and professional-grade workflows in iPadOS. Despite progress, Apple Intelligence remains underwhelming compared to rivals like OpenAI and Google. Public beta arrives in July, with full release in fall.

Read more

Google Quietly Upgrades Gemini 2.5 Pro

2025-06-09

Google unveiled an upgraded Gemini 2.5 Pro model with improved coding, reasoning, and response quality. It excels in benchmarks like LMArena, WebDevArena, and Aider Polyglot, showcasing multilingual coding and scientific abilities. Expected for full release soon, the model addresses past feedback with more structured and creative outputs. Amid fierce AI competition, Google continues to position itself as a reliable enterprise-grade provider.

Read more

Cudis Bets on Wearables, AI, and Solana Tokens to Drive Longevity Movement

2025-06-08

Cudis, a wearable device developer, launched a Solana-based token to reward users for improving health habits via its AI-powered smart ring and app. The app offers personalized health insights and integrates with other wearables like Apple Health. Users earn CUDIS tokens by tracking metrics such as sleep and activity, with rewards based on consistent progress. The total token supply is 1 billion, allocated for community, investors, and ecosystem growth. Cudis also features an AI coach built on C

Read more

Orca Explores the Human-Guided Future of AI Agents

2025-06-09

Researchers at UC San Diego introduced Orca, an open-source system enabling LLMs to assist web users by guiding interactions. It accelerates tasks, improves accuracy, and promotes human-AI collaboration with features like summarization, data extraction, and spatial organization. Built as an Electron app with React, Orca uses Claude 3.7 for language tasks and emphasizes shared control to build trust. While promising, it has performance limits under heavy loads and is a research prototype. Other t

Read more

OpenAI to Retain Deleted ChatGPT Conversations per Court Order

2025-06-07

OpenAI will retain deleted ChatGPT data due to a court order related to a copyright lawsuit. The move aims to preserve potential evidence, storing the data securely and accessible only for legal obligations. This policy does not apply to zero data retention agreements or enterprise/educational versions. OpenAI plans to appeal the ruling.

Read more

Anthropic Open-Sources Tool to Track Large Language Models' "Thinking"

2025-06-09

Anthropic's research team has open-sourced a tool to track the internal operations of large language models. The circuit tracing Python library works with open-source models, enabling graphical exploration via Neuropedia. It identifies replacement circuits, generates attribution maps, and calculates the impact of transformer features on model outputs. This contributes to LLM mechanistic interpretability, offering insights into reasoning processes. The library can be run from tutorial notebooks o

Read more

Meta in Talks to Invest Billions in Scale AI, According to Reports

2025-06-09

Meta is reportedly negotiating a multi-billion-dollar investment in Scale AI, potentially exceeding $10 billion. This would be Meta's largest external AI investment and one of the biggest private company financings ever. Scale AI, led by Alexandr Wang, offers data labeling services to firms like Microsoft and OpenAI, generating $870 million in revenue last year and projected to reach $2 billion this year. Meta previously invested in Scale AI's Series F round, valuing it at $13.8 billion, and Sca

Read more

ChatGPT Memory Upgrade: Smarter Conversations and More Control for Free Users

2025-06-06

OpenAI's June 2025 update introduces conversation memory for free ChatGPT users, allowing the AI to reference recent chats for short-term continuity. Previously exclusive to paying subscribers, this feature enhances personalization while raising privacy concerns. Free users can manage memory settings and use Temporary Chat for private discussions. Long-term memory and deeper personalization remain available only to Plus and Pro subscribers.

Read more