Google I/O is an annual developer conference by Google where the company showcases its latest technologies and products. This year’s event, held in May, was particularly exciting, with a strong focus on AI-driven innovations set to empower both developers and users to glimpse into the future of technology.
Some of the key announcements at the conference revolved around Google Bard, the implementation of PALM 2, and new features like fully realistic maps and the Magic Editor tool. Google stated at the conference that the company aims to fully pivot towards artificial intelligence, meaning that in the future, we can expect to see more products, services, and innovations based on AI.
In the following lines, we’ll present to you most of the news announced at the conference.
Google is currently testing a language model powered by artificial intelligence for its search engine.
“Users will see the next steps they can take when searching the internet. They will be presented with AI-generated summaries of information with links in case they want to delve deeper. When they click on the suggested next step, the search engine will take them to a new mode where they can ask Google more about the topic they are interested in. The context will change with the change of subsequent questions,” said Google.
The new search engine will allow you to explore various topics on the internet in a much more creative and dynamic way. Searching will be based on algorithms that generate similar topics, ideas, and images using the keywords you use.
Also announced at the conference was the new “Perspectives” filter, which will soon be available in the top corner above the search results. Through it, you’ll be able to see how other users search for information. For example, discussions on forums, Q&A sites, social media platforms, especially those containing videos.
Palm 2 is a large-scale language model designed for self-learning, without the need for additional data type classification, aimed at improving the accuracy of NLP tasks. In other words, the model can independently classify textual information and analyze the emotions and motives embedded by the user during interaction. The introduction of Palm 2 is highly technological and is attributed to the use of the most up-to-date databases. Additionally, the model can be trained on nearly all programming languages and domains.
Powering over 25 of the products announced at the I/O conference, Palm 2 is used daily by dozens of Google teams in their work. Moreover, Palm 2 also offers enhanced support for writing and debugging codes.
Proceed to the second part.
English 
Bulgarian