Google I/O 2024: The Future of AI Assistants Unveiled
Google I/O 2024: The Future of AI Assistants Unveiled
Google I/O 2024 showed that Google is doubling down on organic AI development both for Cloud and On-device AI processing. From advancements in AI and mobile technology to innovative new projects, Google demonstrated its commitment to shaping the future. This blog analyzes key highlights from the event.
The LLM Revolution Takes Center Stage
Google is pushing the envelope on advancements in large language models (LLMs) and small language models (SLM). Three models were introduced: Gemini Nano, Gemini Flash, and enhancements to Gemini 1.5 Pro. Nano, designed for on-device processing, promises to revolutionize mobile experiences with features like real-time translation and personalized text generation. Nano will run on Google Pixel devices. Gemini Flash, integrated with existing Google services, will empower users to interact with text, images, and documents more intelligently. Meanwhile, Gemini 1.5 Pro boasts advanced reasoning capabilities, making it ideal for tasks like code generation and scientific analysis (see this blog for more details).
Google Search Upgrades: Generative Experience and Search with Video
AI is coming to Google Search and with it comes AI overviews. This could be a game changer and it has been in Google Labs for a while. Search Generative Experience (SGE) is only launching in the US for now.
Along with SGE comes a new and improved ability to use a video to search. Users can upload a video and speak their questions such as “how do I fix this”? This will really revolutionize search – and learning – as it fundamentally represents a change to the way people and machines will search for answers.
Generate Rich Content with VideoFX, ImageFX and MusicFX
The Google DeepMind team is not standing still. Alongside the LLMs, Google unveiled Veo, a video generation LLM. Google Veo will power VideoFX its new video creation service Google also announced updates to both ImageFX and MusicFX – all now available in Google Labs.
ImageFX is now powered by Imagen 3, which is the most powerful google image generation LLM to date. Imagen 3 now will allow for more photo realistic images with fewer artifacts.
MusicFX now allows for multiple tracks of music to be mixed. It also offers DJ Mode which makes it easy to combine different tracks and musical instruments.
AI at the Edge with Pixel Phones and Nano
Google’s new Pixel phone stole the show with its integration of the Gemini Nano LLM. This on-device processing eliminates the reliance on cloud connectivity, enabling faster response times and enhanced privacy. With Nano, the Pixel phone transcends its traditional role, becoming a powerful AI-powered personal assistant.
Project Astra will power Google Live
Google I/O wasn’t just about refinements to existing technology, it also unveiled ambitious new projects. Project Astra, a groundbreaking initiative exploring the potential of ambient computing, envisions a future where technology seamlessly integrates with our surroundings. Astra will power Google Live (coming later this year) which will allow a user to point a smartphone camera at an object and have Google Live answer questions about it. In a way, Google Live is more of an assistant than the original Google Assistant that runs on iOS and Android. If anything, Project Astra and Live give hints of what the future of Assistants might look like.
Bottom Line
From expanding its LLM product line, to transforming search, and reimagining video generation, Google I/O 2024 painted a picture of a future that will feature a world in which AI Assistants help us more than ever, in a multimodal fashion. While the competition with Microsoft/OpenAI is fierce, Google’s speed of technological innovation, led by the DeepMind team, has definitely not gone unnoticed.
UPCOMING WEBINARS
The Race to Generative AI: Putting CoPilots to Work in Your Enterprise
Generative AI is here and with it the promise of increased productivity. But is that a promise or is it a reality?
In this webinar Aragon analyst, Adam Pease and CEO, Jim Lundy take on the challenge of putting Generative AI to work.
Have a Comment on this?