Competitors like ChatGPT and Claude have introduced advanced models and innovative applications enabling AI agents to handle coding, design tasks, and desktop operations. In contrast, Google's Gemini has remained relatively subdued, which could signal a strategic pause ahead of major developments.
Recent updates for Gemini include the capable Nano Banana 2 for image creation, Lyria 3 for generating music from user descriptions, and various enhancements within Google Workspace.
However, in areas such as on-device collaborative AI tools or independent agents performing user-directed actions, Gemini appears to trail behind ChatGPT and especially Claude.
Anticipation builds for shifts in this dynamic, likely unfolding later this month.
This edition of the AI newsletter introduces its official title: Prompt Mode.
Hosted by Ben Patterson, Prompt Mode delivers weekly insights into key AI developments relevant to general users. Expect useful advice on AI applications, direct testing of emerging technologies, and effective prompting strategies to optimize interactions with AI systems.
Appreciate your interest; subscribe via the provided link to continue receiving updates.
Preparations are underway for Google's yearly I/O event, scheduled to begin on May 19, with strong indications of substantial progress for Gemini. It seems Google has reserved key revelations for the main keynote, promising a series of impactful Gemini disclosures. Attendees should prepare for an engaging session.
Among potential highlights is Proactive Assistant, a capability that delivers timely, customized recommendations drawing from integrated Google platforms like Gmail, Calendar, and Drive, alongside SMS, contacts, and on-screen content.
A more ambitious prospect involves enhancements to the prior Gemini Agent. Reports from Business Insider suggest internal trials of a version dubbed 'Remy,' described as an around-the-clock aide for professional, educational, and routine needs. This upgrade aims to transform the Gemini application into a proactive personal helper capable of executing tasks independently and managing intricate responsibilities.
Additionally, expectations surround the debut or preview of the forthcoming primary Gemini iteration, possibly Gemini 4. This version is projected to emphasize agent-based functionalities, with speculation about built-in capabilities for producing images and videos. Currently, tools like Nano Banana 2 and Veo 3.1 operate distinctly from the core Gemini framework.
Overall, the I/O gathering may mark the peak of an ongoing shift observed this year: AI companions emerging from conversational confines to offer real-time, initiative-taking support in everyday scenarios.
Should Google launch such a comprehensive AI companion, it would likely accompany a premium subscription option.
Interactions with AI often lead to frustrations, such as when a tool like ChatGPT disregards nuanced directives—perhaps overhauling a cover letter despite requests solely for critique, ignoring pleas to avoid alterations.
The underlying issue stems from AI's tendency to prioritize the core query over subsequent qualifiers, leading to incomplete adherence to full instructions.
A practical solution exists in the form of the 'anti-goal' prompting technique. This method uses XML-like structures to clearly delineate the AI's role, objectives, and prohibitions, ensuring comprehensive processing of all elements in the request. Experiment with it for improved results.
Thank you for engaging with this installment of Prompt Mode. To catch the next edition, subscribe now for direct email delivery. Until then.
With over two decades in consumer tech journalism, Ben Patterson now centers on AI's influence on daily life. His work examines cutting-edge large language models and their practical integrations in professional and personal settings to equip readers for the evolving AI landscape. 'AI's transformative impact will arrive faster than anticipated,' he notes. 'Daily engagement is key to adaptation.' A contributor to PCWorld since 2014, Patterson has reported on diverse topics from portable computers to surveillance devices prior to spearheading the site's AI coverage. His pieces have featured in PC Magazine, TIME, Wired, CNET, Men's Fitness, Mobile Magazine, and others. He possesses a master's in English literature.