Speaker 1: We've made significant con contributions to the scientific community like developing the transformer, which set the stage for much of the generative AI activity we see today, and we are continue committed to continuing to bring these technologies to the world in a responsible way that benefits everyone. This is the journey we've been on with large language models, which can make engaging with technology more natural and conversational. [00:00:30] Back at IO in 2021, we unveiled our Lambda AI model, a breakthrough in conversational technology. Next, we are bringing Lambda to an experimental conversational AI service, which we fondly call Bard. You'll be able to interact with Bard to explore complex topics, collaborate in real time, and get creative new ideas. For example, let's say you're in the market for a [00:01:00] new car, one that's a good fit for your family. Bard can help you think through different angles to consider from budget to safety and more and simplify and make sense of them. Bard's suggestion to consider fuel type might spark your curiosity, so you can ask it to explain the pros and cons of buying an electric car and get helpful insights. And we all know that once you buy a new car, you'll have to plan a road trip [00:01:30] Bar can help you plan your road trip so he can take your new car out for a spin.
Speaker 1: He might ask Bart to help you find scenic routes, interesting places to stop along the way, and fun things to do when you, you and your family get to your destination. Bard seeks to combine the breadth of the world's knowledge with the power, intelligence, and creativity of our large language models. It draws in information from the web to provide fresh, high quality responses. [00:02:00] We are releasing Bart initially with our lightweight model version of Lambda. This much smaller model needs significantly less computing power, which means we'll be able to scale it to more users and get more feedback. We just took our next big step by opening Bard up to trusted testers. This week we'll continue to use feedback from internal and external testing to make sure it meets the high bar, our high bar for quality, safety, and groundedness. [00:02:30] Before we launch it more broadly, human curiosity is endless, and for many years we've helped remove roadblocks to information so you can follow your curiosity wherever it takes you. From. Learning more about a topic to understanding a variety of viewpoints,
Speaker 1: People often turn to Google for quick factual answers like what is a constellation? Already today we give you fast answers for straightforward queries like, [00:03:00] but for many questions, there's no one right answer what we call neuro queries. Questions like, what are the best constellations to look for? When stargazing for questions like those, you probably want to explore a diverse range of opinions or perspectives and be connected to the expansive wisdom of the weapon. That's why we are bringing the magic of generative AI directly to your search results so soon. If you ask, [00:03:30] what are the best constellations to look for when stargazing new generative AI features will help us organize complex information and multiple viewpoints, write in search results. With this, you'll be able to quickly understand the big picture and then go on to explore different angles. So say this new information on constellations peaks your interest. You can dig deeper, for instance, to learn what time of year is best to see them and explore further [00:04:00] on the web. Open access to information is core to our mission, and we know people seek authentic voices and diverse perspectives. As we scale these new generative AI features like this in our search results, we continue to prioritize approaches that will allow us to send valuable traffic to a wide range of creators and support a healthy open web. In fact, we've sent more traffic to the web every year, each year [00:04:30] than the year prior.
Speaker 1: The potential for generative AI goes far beyond language and text. As we mentioned earlier, one of the most natural ways people engage with information is visually. With generative ai, we can already automate 360 degrees spins of sneakers from just a handful of still photos, something that would've previously required merchants to use hundreds of product photos and costly technology. As we look ahead, you could imagine [00:05:00] how generative AI might enable people to interact with visual information in entirely new ways. They might help a local baker collaborate on a cake design with a client or a toy maker, dream up a new creation. They might help someone envision what their kitchen looks like, but with green cabinets instead of wood, or describe and find the perfect complimentary pocket square to match a new blazer. In our quest to make search more natural and intuitive, [00:05:30] we've gone from enabling you to search with text, to voice, to images, to combination of modalities like you saw with multis search today that Liz talked about. As we continue to bring generative AI technologies into our products, the only limit to search will be your imagination. We are on our own products. It's important to make it easy, safe, and scalable for others to benefit from these advances. [00:06:00] Next month, we'll start onboarding developers, creators, and enterprises so they can try a generative language. API initially powered with Lambda, with a range of models to follow.
