In one of the examples shown yesterday
Posted: Mon Jan 06, 2025 5:45 am
The last Google developer conference (Google I/O) with thousands of attendees in the auditorium dates back to 2019. The pandemic forced a pause (of a forced nature) in what is probably the most important annual event organized by the Mountain View company. That is why the latest edition of Google I/O, which raised the curtain yesterday in California, was a kind of reboot for the internet giant.
During the conference, Google came up with a whole host of new things: a smartphone, a new version of Android and, above all, a whole bunch of "superpowers" born from the heat of artificial intelligence (AI).
The Google I/O conference , held at the Shoreline Amphitheatre, just a few metres from the company's headquarters in Mountain View, yesterday had a star player: AI , which has finally opened the floodgates of a furious war in the technology sector.
“As you may have heard, AI has been very busy over the past year ,” said Google CEO Sundar Pichai in his opening keynote at Google I/O.
If you cannot view the embedded video correctly, click here
The most important announcement directly related to AI that Google made yesterday was the presentation of the new AI language model PaLM 2 .
PaLM 2, Google's new AI language model
This new product will reportedly be integrated into "25 Google products and features." Compared to its predecessor, PaLM 3's logical and mathematical capabilities and programming skills have been substantially improved, Google executive Sissie Hsiao said. It is also worth mentioning that some variants of PaLM 2 will be able to work completely offline on smartphones, for example.
The new system that Google released yesterday will also be able to handle multilingual inputs with extraordinary accuracy. As an example of these abilities, Google showed how PaLM 2 was able to correctly translate the German phrase “Ich verstehe nur Bahnhof” (literally “I only understand station”) as “I don’t understand anything.” And when asked why this phrase might at first cause confusion in the listener, PaLM 2 correctly answered that the expression was actually a phrase with a different meaning from its 100% literal translation.
According to Hsiao, AI tools are particularly popular among programmers, so PaLM 2 has expanded some specific capabilities in this particular area . The system will, for example, be able to generate code with explanatory comments in different languages, detect errors and then correct them. And if this functionality really works as Google claims, it will effectively make programmers' jobs much easier.
Bard significantly expands his capabilities
As for Bard, Google's response to the famous ChatGPT, Google describes this tool as an experiment (probably to protect itself from any possible mistakes made by this AI). By calling it an experiment, those from Mountain View seem to want to exonerate Bard from possible mistakes ( for which it was already put in the spotlight a few months ago ).
Google wants to expand the usefulness of this system by allowing Bards to export the results of some queries to Gmail and Google Docs. This "superpower" will give users the opportunity to rely on Bard to write reports and emails and to use the chatbot directly in other apps.
by Bard, this tool is, for example, capable of generating the complete agenda of an event or creating a job description from scratch.
One of the most important innovations of Bard is that from now on the chatbot can be used in more than 180 countries and not only in the United States and the United Kingdom. And not only that. In addition to English, Bard's default language, the Google chatbot can also be used in Korean and Japanese. And 40 versions in other languages will soon be released.
In addition, the answers to searches on Bard will soon be enriched with videos and photos. When asked what places to visit on a trip to Berlin, Bard includes, for example, images of the Brandenburg Gate and Checkpoint Charlie. At the end of the results list, there is also the option to “Google it”.
Bard also has the ability to analyze photos. And by showing it a drawer decorated with hand-made elements, the Google chatbot makes suggestions for the user to create a similar piece of furniture.
Google's AI system is also equipped with an image generator , which creates photorealistic images from text commands.
Google is also working with Adobe to integrate Firefly , the software company’s image generator, into Bard “in the coming months.” In Bard, you will also have the option to upload a photo and ask the chatbot for a description of its contents.
During the Google I/O conference, Hsiao also announced that Bard would be integrated into apps such as Kayak and Open Table. The goal? To familiarize the largest number of users with the "wonders" of Bard.
AI comes to Google searches
Google is also calling the “conversational mode” that will soon be integrated into the search results pakistan number screening of its famous search engine an experiment. This “conversational mode”, dubbed “AI Snapshot” by Google, summarizes search results as a continuous text. And the text is crafted in such a way that it does not really seem machine-generated and is permeated with a certain humanity. Google also does not forget where it comes from and will display links next to the text of the “AI Snapshot” summary.
If you cannot view the embedded video correctly, click here
When asked to recommend Bluetooth speakers, for example, Google's new "conversational mode" provides tips on what features to look for when buying such devices, as well as a list of stores where they can be purchased. Unlike regular Google searches, in "conversational mode" searches can be conveniently refined and display, for example, models "under $100."
With this new feature, Google is clearly defending itself against the much-vaunted integration of ChatGPT into Bing, Microsoft's search engine. It is also worth noting that in this feature, the results are not simply displayed in black and white, but are accompanied by attractive coloured backgrounds and graphic elements.
Google's new "conversational mode" will be available for the time being only in a test environment called "Search Labs." The feature will also be accessible only to users in the
During the conference, Google came up with a whole host of new things: a smartphone, a new version of Android and, above all, a whole bunch of "superpowers" born from the heat of artificial intelligence (AI).
The Google I/O conference , held at the Shoreline Amphitheatre, just a few metres from the company's headquarters in Mountain View, yesterday had a star player: AI , which has finally opened the floodgates of a furious war in the technology sector.
“As you may have heard, AI has been very busy over the past year ,” said Google CEO Sundar Pichai in his opening keynote at Google I/O.
If you cannot view the embedded video correctly, click here
The most important announcement directly related to AI that Google made yesterday was the presentation of the new AI language model PaLM 2 .
PaLM 2, Google's new AI language model
This new product will reportedly be integrated into "25 Google products and features." Compared to its predecessor, PaLM 3's logical and mathematical capabilities and programming skills have been substantially improved, Google executive Sissie Hsiao said. It is also worth mentioning that some variants of PaLM 2 will be able to work completely offline on smartphones, for example.
The new system that Google released yesterday will also be able to handle multilingual inputs with extraordinary accuracy. As an example of these abilities, Google showed how PaLM 2 was able to correctly translate the German phrase “Ich verstehe nur Bahnhof” (literally “I only understand station”) as “I don’t understand anything.” And when asked why this phrase might at first cause confusion in the listener, PaLM 2 correctly answered that the expression was actually a phrase with a different meaning from its 100% literal translation.
According to Hsiao, AI tools are particularly popular among programmers, so PaLM 2 has expanded some specific capabilities in this particular area . The system will, for example, be able to generate code with explanatory comments in different languages, detect errors and then correct them. And if this functionality really works as Google claims, it will effectively make programmers' jobs much easier.
Bard significantly expands his capabilities
As for Bard, Google's response to the famous ChatGPT, Google describes this tool as an experiment (probably to protect itself from any possible mistakes made by this AI). By calling it an experiment, those from Mountain View seem to want to exonerate Bard from possible mistakes ( for which it was already put in the spotlight a few months ago ).
Google wants to expand the usefulness of this system by allowing Bards to export the results of some queries to Gmail and Google Docs. This "superpower" will give users the opportunity to rely on Bard to write reports and emails and to use the chatbot directly in other apps.
by Bard, this tool is, for example, capable of generating the complete agenda of an event or creating a job description from scratch.
One of the most important innovations of Bard is that from now on the chatbot can be used in more than 180 countries and not only in the United States and the United Kingdom. And not only that. In addition to English, Bard's default language, the Google chatbot can also be used in Korean and Japanese. And 40 versions in other languages will soon be released.
In addition, the answers to searches on Bard will soon be enriched with videos and photos. When asked what places to visit on a trip to Berlin, Bard includes, for example, images of the Brandenburg Gate and Checkpoint Charlie. At the end of the results list, there is also the option to “Google it”.
Bard also has the ability to analyze photos. And by showing it a drawer decorated with hand-made elements, the Google chatbot makes suggestions for the user to create a similar piece of furniture.
Google's AI system is also equipped with an image generator , which creates photorealistic images from text commands.
Google is also working with Adobe to integrate Firefly , the software company’s image generator, into Bard “in the coming months.” In Bard, you will also have the option to upload a photo and ask the chatbot for a description of its contents.
During the Google I/O conference, Hsiao also announced that Bard would be integrated into apps such as Kayak and Open Table. The goal? To familiarize the largest number of users with the "wonders" of Bard.
AI comes to Google searches
Google is also calling the “conversational mode” that will soon be integrated into the search results pakistan number screening of its famous search engine an experiment. This “conversational mode”, dubbed “AI Snapshot” by Google, summarizes search results as a continuous text. And the text is crafted in such a way that it does not really seem machine-generated and is permeated with a certain humanity. Google also does not forget where it comes from and will display links next to the text of the “AI Snapshot” summary.
If you cannot view the embedded video correctly, click here
When asked to recommend Bluetooth speakers, for example, Google's new "conversational mode" provides tips on what features to look for when buying such devices, as well as a list of stores where they can be purchased. Unlike regular Google searches, in "conversational mode" searches can be conveniently refined and display, for example, models "under $100."
With this new feature, Google is clearly defending itself against the much-vaunted integration of ChatGPT into Bing, Microsoft's search engine. It is also worth noting that in this feature, the results are not simply displayed in black and white, but are accompanied by attractive coloured backgrounds and graphic elements.
Google's new "conversational mode" will be available for the time being only in a test environment called "Search Labs." The feature will also be accessible only to users in the