Pupsic

ChatGPT continues to hold a strong presence as the most widely used AI assistant with over 1.1 billion monthly users. However, its market share has fallen below 50% for the first time, as competitors Gemini and Claude gain traction with 662 million and 245 million users respectively. This shift indicates growing diversity in AI assistant preferences worldwide.

Respond.io, a notable Malaysian startup, has raised $62.5 million to expand its AI-powered messaging platform. The company uses AI agents to handle numerous customer interactions, charging clients per conversation instead of per user seat, with plans to grow further through acquisitions.

Italy's competition authority, AGCM, previously found Apple abused its market dominance regarding third-party developers. Following this, the AGCM has now initiated an investigation under the Digital Markets Act (DMA) to examine Apple's competitive behavior.

This article explores the concept of Probabilistic Design, highlighting how embracing uncertainty and interpreting AI predictions with nuance can empower UX and product teams to make smarter, more adaptive design decisions.

WPP and Madison & Wall project significant growth in AI-driven search ads, expecting it to be the fastest expanding advertising channel. The advertising market is anticipated to grow steadily despite economic uncertainties, highlighting resilience in ad spending.

Z.ai's GLM-5.2 is a 753-billion parameter open-source language model that outperforms GPT-5.5 on long-horizon coding tasks while costing only one-sixth as much. Licensed under MIT, it offers enterprises unrestricted use and local deployment. Innovations like IndexShare reduce compute requirements substantially. Available on Hugging Face and APIs with flexible plans, GLM-5.2 delivers top benchmark scores, cost efficiency, and developer-friendly features, marking a significant advancement for affordable, high-performance AI coding technology.

Databricks announced Lakehouse//RT and LTAP, two groundbreaking technologies that unify operational and analytical data storage to remove latency and pipeline complexity, enabling AI agents to access live data in real time. This approach eliminates the traditional data duplication and delays caused by separate systems, supporting faster and more efficient AI applications. The innovations mark a shift from fragmented data architectures to streamlined, unified infrastructures that meet the demands of modern AI workloads.

Stanford's DeLM framework reimagines multi-agent AI coordination by removing the need for a central controller. Through a shared knowledge base and decentralized task management, agents collaborate more efficiently, cutting task costs by 50% and boosting accuracy by 10.5%. This innovation improves performance in complex reasoning and software engineering tasks, proving faster, cheaper, and more reliable than traditional centralized systems.

Bernard Hampton of Bank of America reveals how the institution upskills 200,000+ employees for an AI future, balancing technical prowess with critical human skills like empathy and judgment. Through a three-level AI adoption strategy and AI-driven simulations, the Academy enhances workforce agility and client service. The approach emphasizes internal talent growth, measured AI use, and ongoing learning to adapt in a fast-changing environment.

Coach, with input from Gen Z, has unveiled a new brand platform that shifts the focus from fixed luxury narratives to collaborative storytelling. This innovative strategy leverages celebrity involvement and Gen Z perspectives to craft a more authentic and engaging brand experience.