It seems that the biggest point of focus for Google’s application of Gemini’s machine learning capabilities is centered on quickly and easily answering users’ questions. You will be able to ask complex questions in the search bar and have Gemini screen articles to get answers and then use generative text to offer summaries, or else break tasks down into step-by-step instructions. “With our custom Gemini model’s multi-step reasoning capabilities, AI Overviews will help with increasingly complex questions,” Google claims. “Rather than breaking your question into multiple searches, you can ask your most complex questions, with all the nuances and caveats you have in mind, all in one go.”
Google also plans to integrate several other features, such as settings that allow you to adjust the AI Overview to customize the results to your liking, AI-organized page results, and video-based searches that use Gemini’s image recognition capabilities to answer questions. Most of the demonstrations showcased in the keynote were a little underwhelming, though.
Google VP of Product Rose Yao, for instance, came on stage and used the image-based search function to find the make and model of a record player, while also looking up what the tonearm is called and why it isn’t staying in place, combining three to four searches into one. This seems like a highly complex solution to something that wasn’t much of a problem. This also might have seemed more impressive if the make and model weren’t printed on the front of the record player.
