Google I/O 2024 has been one of the most important events of the year so far. At this year’s conference, Google presented its vision for the future of artificial intelligence. The presentation took place shortly after OpenAI announced the revolutionary GPT-4o and featured updates for Android, WearOS, Android Automotive, and any device that powered by Gemini, Google’s family of AI models.
We've compiled everything that was announced at the Google I/O 2024 below, from improvements in its AI models to all the new features coming to Android phones once the stable version of Android 15 is released at the end of the year.
Gemini 1.5 Pro and Gemini Flash
On Tuesday, Google announced several improvements to the Gemini family for 2024. The lightest model, Gemini Flash, is set to be the fastest and offer the lowest latency within the Gemini family, running locally on devices. It’ll serve as an alternative to Gemini Nano and will be ideal for apps requiring a fast response rather than a denser model.
In addition to introducing the newest member of its AI family, the company also revealed enhancements coming to Gemini 1.5 Pro. These include improved computational speed. As such, 1.5 Pro will be able to analyze large documents. Specifically, it’ll have the ability to read 1,500-page files, summarize up to 100 emails simultaneously, and handle 1 hour of video or codebases with over 30,000 lines.
Gemini 1.5 Pro will also be smarter at understanding images and capable of real-time travel planning based on information from emails, flights, calendars, and more.
Project Astra
Understanding Google’s naming conventions can be a bit confusing, but here’s a breakdown. Google’s AI model family is referred to as Gemini, which includes the following models:
- Gemini Flash, a lightweight model
- Gemini Nano, a moderately lightweight model currently used in phones
- Gemini Pro, a more powerful model than Nano
- Gemini Ultra, the most powerful model. Google is dedicating the most effort to this model
- Gemini Live, not a model itself, but the name given to the conversational interface integrated into mobile apps
- Gemini Advanced, a paid Google subscription that includes access to the entire family
As if this wasn’t confusing enough, Google has announced Project Astra, which aims to “build a universal AI agent that can be truly helpful in everyday life.” The project is part of Google’s plan to create a perfect virtual assistant, powered by a low latency model, that can remember what it sees and hears even with different voice tones.
Essentially, it's Google’s own version of Her–a reference to Spike Jonze’s 2013 movie. Project Astra will eventually be integrated into Gemini, the voice assistant being tested to replace Google Assistant. One interesting integration we saw during I/O 2024 was Project Astra running on glasses, giving us a glimpse of what’s to come in the future.
Veo and Imagen 3
A photorealistic landscape. Image: Google
At I/O 2024, Google also announced Veo and Imagen 3, the latest versions of its text-to-image and text-to-video AI models. These new models are meant to compete with Midjourney, Sora, and similar technologies.
Imagen 3 is an improvement of Imagen 2, with enhanced natural language interpretation and the ability to incorporate very specific details from lengthy prompts.
For its part, Veo is Google’s largest generative video model to date. It can create Full HD (1080p) resolution clips over a minute long. Additionally, Veo is capable of producing timelapse videos and applying various cinematic techniques.
AI Improvements for Google Search
Google Search was another major player at I/O. During the event, Google showed off a feature called AI Overviews that showcases AI-generated search results. This feature will be rolled out in the U.S. first. Users interested in the feature will be able to test it out using Google Labs.
Android Updates
In addition, Google gave I/O attendees a sneak peek at the new features coming to its Android 15 operating system and apps. Here are the highlights:
- Android 15 Beta 2 has been released with several new features.
- Google Gemini is now directly integrated into Android 15, allowing the assistant’s interface to be integrated into the system’s own apps. This integration enables users to ask the assistant about the content it’s seeing.
- New anti-theft functions in Android, including Theft Detection Lock and Private Space.
- Gemini is integrated into apps such as Google Photos, enabling users to ask the AI assistant about their gallery.
- Google Cast has been introduced to Android Automotive.
- Improvements in the battery life of WearOS 5.
- Android TV has been updated to Android 14.
Image | Google
View 0 comments