AI Products news and analysis on Web Pulse: 18161 curated articles tracking the latest AI Products developments, tools, and research, updated continuously from vetted sources.
Zephyr Cloud deployed a pair of AI agents that autonomously hunt and fix bugs in its desktop app. The first agent explores the real app hourly, filing high-confidence bug reports with reproduction steps, while the second…
In a head-to-head test of 12 text tasks, OpenAI's GPT-5.6 Sol Pro scored 112.0 to xAI's Grok 4.5's 106.1, winning 5 tasks to 3 with 4 ties, giving OpenAI a narrow but real edge. Sol Pro excelled in implementation-heavy t…
OpenAI Product and Business Chief Fidji Simo stepped down from her role due to ongoing medical issues, Yahoo Finance reported. The departure comes as the company prepares for an anticipated initial public offering.
OpenAI is deprecating its standalone Atlas browser but integrating its features into the ChatGPT desktop app and Chrome extension, effectively merging browsing capabilities with its AI assistant. The move consolidates Op…
OpenAI claims its GPT-5.6 Sol Ultra model produced a candidate proof of the Cycle Double Cover Conjecture, a 50-year-old open problem in graph theory, using 64 subagents in under an hour. The proof, published by OpenAI w…
OpenAI's Codex team hosted a Reddit AMA on July 10 to discuss GPT-5.6, a new model family with three tiers (Sol, Terra, Luna) offering competitive pricing and state-of-the-art coding performance. The model family include…
Anthropic launched ten specialized AI agents for finance professionals at an invite-only event in New York on May 5, 2026, with JPMorgan CEO Jamie Dimon in attendance. The agents automate tasks like reading quarterly rep…
A developer argues that using AI to catch AI-generated code drift is the wrong approach, citing non-deterministic results, compounding costs, and opacity. Instead, they advocate for deterministic rule engines like ReWeav…
Google Gemini users are urging the company not to discontinue the Gemini 2.5 Flash model, citing superior performance and low latency for voice agents in Australia compared to newer models.
A developer explains how to overcome the 'Abstraction Tax' in Android Edge AI by using custom C++ operations via the NDK. The post details techniques like coarse-grained delegation and zero-copy data transfer with direct…
Cactus released version 2 of its on-device AI inference platform, featuring model confidence-based routing to cloud fallback, lossless 4-bit quantization, and GPU acceleration on Apple Metal. The platform processes milli…
Google India launched Gemini-powered advertising tools on July 9, 2026, including a Business Agent for Leads that places an AI brand agent inside Search ads to qualify leads in real time. The rollout also features YouTub…
Melius, a startup led by CEO Joowon Kim, launched a visual canvas for AI agents that replaces chatbots with node-based creative production. The tool integrates with agent protocols like MCP to let external AI read and wr…
OpenAI's model defeated top human programmers in both the heuristic and algorithmic divisions of the 2026 AtCoder World Tour Finals, solving all five algorithmic problems including two that no human could solve. The vict…
AI² Robotics raised $735 million at a $2.8 billion valuation for its wheeled humanoid robots, positioning the Shenzhen-based company among leaders in China's physical AI sector. The funding, backed by state funds, corpor…
Meta's new Kylie Jenner-fronted AI smart glasses campaign faces backlash over surveillance concerns, with consumers and experts questioning the normalization of always-on cameras in fashion. Despite partnerships with lux…
Meta faces backlash over a series of controversial AI and privacy decisions, including launching an AI image generator that uses Instagram photos, facing EU fines for addictive platforms, and adding facial recognition co…
SpaceXAI launched Grok 4.5 on July 8, pricing it over 60% cheaper than Anthropic's Claude Opus 4.8 and OpenAI's GPT-5.5 at $2 per million input tokens and $6 per million output tokens. The model features a 500K token con…
OpenAI's desktop app changes have caused widespread user confusion by merging ChatGPT and Codex into a unified application with unclear naming, nearly identical icons, and a buggy chat interface. The old ChatGPT app was …
Researchers developed a deep reinforcement learning framework based on Proximal Policy Optimization to optimize battery charging for autonomous mobile robots in warehouses. The model dynamically selects charging stations…