Google AI Storm 🦾🌪️

Source: Ingame

Disclaimer:
I am highlighting some of the tools which have been unravelled by Google, 
and also I am discovering more on this. For this purpose I have kept this 
blog brief.

I will be highlighting some major updates from Google Research and what makes them remarkable and ahead of the competition.

Google is developing Project Magi, a new search engine that will eventually replace the one that currently powers more than 8.5 billion daily searches.

The recent success of ChatGPT has demonstrated to us the potential of AI-powered chatbots to offer an information access method that is more efficient than standard Google search. Magi serves as a signal to the public (and to investors) that Google is making an effort to stay up with its rivals and is capable of developing technology that is comparable to ChatGPT.

Google Magi might have a significant impact on the future of search and what it means for users, businesses, and marketers thanks to its more conversational interface, personalized results, and in-platform transactions. Here is everything we currently know about Google Magi.

• 160 engineers have been assigned to “Google Magi” by Google. They are developing a new search engine with personalized results and AI elements with the goal of giving people pertinent information that is tailored to their specific needs.
• Similar to ChatGPT, the new search engine will include a chat-like user interface that will enable users to ask follow-up questions and make searches more conversational and personalized to their interests.
• With hopes to reach 30 million people by the end of the year, Project Magi will debut in May with an initial rollout to 1 million users in the United States. In the autumn, more features are anticipated to be introduced.
• By integrating Google Pay, customers will be able to conduct transactions directly on Google and receive personalized product recommendations based on their search history and interests, such as making purchases or booking tickets.

Another project by Google is DREAMIX :

• Dreamix, a diffusion-based AI video editor from Google researchers, can create new videos from an initial image or edit old ones using text descriptions.
• Dreamix noises the source images before sending them to a video diffusion model for video editing. The video diffusion model then creates new images from the noisy source images based on text cues and assembles them into a video. As a result, the source photographs offer a form of sketch that, while allowing for alterations, represents, for instance, the outline of an animal or its movements.
• Dreamix has the ability to create new videos in addition to altering already-existing ones. Google displays these two programmes: By making small adjustments, like changing an object’s position, video synthesis from a single image first creates additional images before applying them to the video model. Additionally, Dreamix has the ability to produce subject-specific videos. For instance, a video showing a video game character lifting weights might be produced using multiple images of a toy character.
• You can find many more applications and it’s official research paper here: https://dreamix-video-editing.github.io/

source: subj_driv_lifting.mp4( video link [clickable] )

source: img2vid_turtle.mp4( video link [clickable] )

Also, there are many other AI softwares being created by Adobe, Microsoft, Nvidia, and many more. So we are moving into the age of “productive automation” which may replace ample amount of jobs, but also create new ones. I know I am being very candid here, but it’s the fact.

Hence, we need to be aware of such discoveries and try to implement it for a small use case and understand the technical aspects behind it, which will help us to learn and apply. I would be writing more on this once I get sufficient amount of time to research. For time being, I thought this would be a topic of interest for those who haven't came across these tools.

Happy exploring !!