May 20, 2025 News

Google Expands Project Mariner AI Agent to Handle Multiple Web-Browsing Tasks Simultaneously

Google is rolling out Project Mariner, an experimental AI agent that browses websites and completes tasks like purchasing tickets or groceries without users visiting sites directly. The updated version runs on cloud virtual machines and can handle up to 10 tasks simultaneously, addressing previous limitations that required users to remain idle while the agent worked.

Google Unveils Deep Think Reasoning Mode for Enhanced Gemini Model Performance

Google introduced Deep Think, an enhanced reasoning mode for Gemini 2.5 Pro that considers multiple answers before responding, similar to OpenAI's o1 models. The technology topped coding benchmarks and beat OpenAI's o3 on perception and reasoning tests, though it's currently limited to trusted testers pending safety evaluations.

Google Integrates Project Astra's Real-Time Multimodal AI Across Search and Developer APIs

Google announced Project Astra will power new real-time, multimodal AI experiences across Search, Gemini, and developer tools through its Live API. The technology enables low-latency voice and visual interactions, with plans for smart glasses partnerships with Samsung and Warby Parker, though no launch date is set.

Android Studio Introduces Autonomous AI Development Agents with Journeys and Agent Mode

Google is adding "agentic AI" capabilities to Android Studio, including Journeys for natural language app testing and Agent Mode for autonomous multi-stage development tasks. The AI can handle complex workflows like API integration, dependency management, and bug fixing without extensive manual coding.

Apple to Release AI Development Framework for Third-Party Developers at WWDC

According to Bloomberg, Apple plans to unveil a set of AI products and frameworks at its upcoming Worldwide Developers Conference (WWDC) in June. The new tools will allow third-party developers to build applications using Apple's AI models, initially focusing on smaller models, as part of the company's strategy to catch up with competitors in the AI space.

Amazon AGI SF Lab's Cognitive Scientist to Speak at TechCrunch Sessions: AI Conference

Danielle Perszyk, who leads human-computer interaction at Amazon's AGI SF Lab, will be speaking at TechCrunch Sessions: AI on June 5 at UC Berkeley. She will join representatives from Google DeepMind and Twelve Labs to discuss how startups can build upon and adapt to foundation models in the rapidly evolving AI landscape.

OpenAI Launches Codex as It Enters the Emerging Field of Autonomous Coding Agents

OpenAI introduced Codex, a new coding system designed to perform complex programming tasks from natural language commands, placing it among a new generation of agentic coding tools. Unlike traditional AI coding assistants that function as intelligent autocomplete, these agentic tools aim to operate autonomously without requiring users to interact directly with the code, though current systems still face significant challenges with reliability and hallucinations.