time finally It is getting closer to what Google originally planned and, therefore, it is now when the search engine company really begins to gain muscle in a field in which, as I have already commented on previous occasions, it already had the necessary technological substrate for years to stand out from its rivals. However, since the arrival of ChatGPT, and especially the new Bing, it has given the impression that Google was going to slipstream. Today, however, this Google I/O 2023 may mark the breaking point, especially depending on the time it takes for all the news presented by the company to arrive during the opening of the developer event.
At the end of last year, Google already set I/O 2023 as the stage to expand its proposals based on artificial intelligence, so that the advances in recent months have only been a timid response to the extensive advances of its competitors. Either due to the need for more time for its development, to Defining an ethical foundation on which to build your products, for a more complete definition of them, etc., their time had not yet come… until now. And it is that of the little more than two hours that the inaugural event of the meeting for company programmers lasted, more than an hour and a half was dedicated to artificial intelligence.
The announcements, which we will review below, and in which the degree of integration of the AI is somewhat diverse, they are an update, a blow on the table, a “enough is enough” Faced with a situation in which there were even those who saw the Internet giant doomed to irrelevance in the near future. It’s true that its rivals have the benefit of getting there first, but Google’s plans are so ambitious and, on paper at least, seem so well defined, that it looks like we’re approaching a very interesting resurgence of what we can already call “The AI War”.
We’ll still have to wait, though. to check the speed at which the deployment occurs everything seen in the presentation. And it is that an important point is that Google has revealed its plans in relation to artificial intelligence for the coming months (or at least a large part of them), which means that its rivals are already alerted about what they will have to compete with. and, consequently, they may already be at work preparing their response. Thus, those of the search engine must now act quickly, otherwise they risk that, when everything they have presented reaches the users, there is already an offer of similar services offered by other technology companies.
In summary, and just before aborting everything presented, I think it is interesting to make a final reflection on the current situation of Google with regard to artificial intelligence. The battery of ads brings Google back to the fore, but in the absence of these becoming a reality for all its potential users, its situation remains fragile., so that the coming weeks and months will be decisive for us to know if they have been able to make up for lost time and once again be at the forefront. Personally, I think they are more than capable of it, but it depends on so many factors that, in truth, everything is still a big unknown.
Integration of generative artificial intelligence in Google Search
Although the number of services offered by Google today is enormous, it is undoubtedly information searches on the Internet are one of the pillars on which its entire structure is sustained, in addition to being the very origin of the company. Thus, it was to be expected that there would be an announcement that related AI and the search engine, although most of the forecasts in this regard pointed to the integration of Bard in the search engine. The plans, however, are much more interesting.
Under the name of Search Generative Experience, Google will open access to some of the users who sign up for the Google Labs beta to test it (for now, again, only in the United States) to its integration project of a generative model that will be able to provide much more answers. rich in searches To this end, in addition to responding to our initial query, it will try to anticipate our interests by proposing subsequent queries, more or less in the same way that chatbots are capable of holding a conversation, but exclusively related to the initial query.
The answers will be complemented, of course, with a selection of related linksamong which we can find the sources that the AI has used to generate the response, so that we will not have to worry about whether the information it is providing us is real or, on the contrary, is the result of a hallucination.
Bard grows, in every way
Given that Bard has been Google’s biggest announcement so far in the field of AI, we were expecting some big announcement in this regard, and luckily we weren’t disappointed, although there are some aspects that we already knew about, and also some which we already expected. Since Google I/O 2023 is a developer event, one of the most important points of the presentation was related to Bard’s functions in the field of programming. Support for more than 20 programming languagesquick actions related to the generated code, fonts used… the list is the most extensive.
Now, the most important announcement of all related to Bard is, without a doubt, its integrations with other services, something that we have already seen before in ChatGPT and to a lesser extent in Bing, but that Google seems to want to take much further. Obviously, the chatbot will offer integration with the company’s ecosystem of services, something that could already be expected, but it turns out that during these months they have been working with other companies in order to also integrate third-party services. Thus, although some of them will still take a while to arrive, the list is very interesting and varied, with services ranging from from Adobe Firefly to Spotify, via Wolfram Alpha, Trip Advisor and Uber Eatsamong other.
Another novelty announced for Bard involves a direct attack on the waterline of GPT-4 and, of course, all services based on this model. Because? Well, because it will incorporate one of the star functions of the most recent version of the OpenAI model: lthe ability to understand imagesfor which he will use Google Lens technology Thus, in the presentation, we could see how Bard received a text and image prompt (a photo of a couple of dogs with the request that he write something witty about them), something that made me It’s impossible not to connect with him. paper from GPT-4, where one of the use cases asked you to explain why an image was humorous.
This is not the only novelty with regard to images, since Bard’s interface will become more visual, offering responses in which the text will give up part of its space to the images, as long as they are related to the queries made. by the users.
And who will be able to enjoy all these new features, along with the features already existing previously in Bard? Well, many more people than could do it until now, since Google announced that it is eliminating the waiting list and that it is expanding access to the chatbot to 180 countries (although it has not specified which ones). At the moment, yes, the service will only be provided in English, Japanese and Korean, but the “close” plans (although without a date, at least public) are to expand it to 40 languages, including Spanish.
Artificial intelligence comes to Google Maps
The part of the presentation dedicated to Google Maps was undoubtedly one of the most striking of the entire event, not so much for what was announced, which is interesting but was already partially anticipated a few months ago, as for the visual impact of seeing it on functioning. As we already told you then, when making a route we will be able to obtain a preview that is extremely rich in detail, ranging from the simulation of the volume of traffic based on statistical information, to a representation of the weather forecast for the moment in which we are going to make that move.
The deployment of Immersive View, yes, will be much slower in terms of its map coverage, something understandable, of course, since the service will have to generate three-dimensional representations of the cities, which will later be used for the immersive visualization of the cities. routes. Thus, for now and by 2023, Google has announced that this function will reach Amsterdam, Berlin, Dublin, Florence, Las Vegas, London, Los Angeles, New York, Miami, Paris, Seattle, San Francisco, San José, Tokyo and Venice. . We cross our fingers that we can enjoy this function in a Spanish city before the end of 2023 or, more likely, in 2024.
Magic Editor, more “magic” editing for Google Photos
Magic Eraser is one of the most interesting functions of Google Photos and, as we told you back in February, it ceased to be the exclusive property of Google Pixel smartphones, as it was included among the benefits of having a Google One plan, either from Android or from iOS. The company is well aware of the success of this AI-based tool, and as a result, they have decided to broaden its reach substantially, introducing Magic Editor, a set of features that allow you to perform advanced editing creative processes without any complications. To exemplify its operation, I think the most explanatory thing is that you see the animation below:
As you can see, Magic Editor is capable of making an intelligent selection, segregating it from the rest of the image, replacing it within it, and also generating the part of the image that was previously occupied by what we have displaced. The result is simply spectacular, since it solves in a matter of seconds a type of retouching that, as a general rule, takes much longer, in addition to requiring a high level of knowledge.
MusicLM with waiting list
Announced in January, MusicML is a model of music-generative artificial intelligence that, as shown by Google, aims to be well above other services dedicated to this purpose. However, as we told you at the time, the company had decided not to launch it as a service, perhaps out of concern that the model would reproduce fragments of the music used for learning in one of its compositions. It seems, however, that the company already has more confidence in it, so they have finally begun to allow access to it, although yes, in a controlled manner, with a waiting list.
More information
