With the new Bing, Microsoft seems to be running a marathon at sprint speed. It’s been a little over a month since the presentation of the new version of its search engine, chatbot included and as the main star, and since then we have not stopped receiving information on new features and improvements in it on a constant basis. If we continue at this rate, I don’t rule out that by the end of the year Microsoft’s artificial intelligence has reached self-awareness and, finally, someone has gotten down to work to solve the truly important problems of humanity and our planet.
Yes, I have exaggerated a bit, but what you cannot deny is that the Prometheus generative artificial intelligence model, which, as we told you a few days ago, is based on GPT-4, adds new functions every so often, which substantially expands what we can expect from him, and that the starting point was already quite high in this sense. And that, despite being based on the latest OpenAI model, at least for the moment it still does not support the combined input of text and images, something that GPT-4 does provide through its API.
The generative artificial intelligences that we have seen in recent times are capable of creating texts, images, music and more, so if we think about the possibilities of a model that combines these features, and therefore offer unified access to themthe result can be something truly advanced, since it should also allow relating the different types of output based on the same input data and the generated outputs.
We still need to see something like this, I imagine, but for now we already have a step forward in that regard. And it is that, as we can read in the official Microsoft blog, Bing chatbot will be able to create images. Yes, just like the image-generative AIs that we’ve been talking about so much in the second half of last month, we’ll just have to tell it we want it to create an image and give you directions for it.
To this end, as we can read in said publication, Bing uses an advanced version of DALL-E, the OpenAI image-generative AI that we already talked about here. Of course, this is not really completely new, since it is based on Bing Image Creator, a service available since October 2022 that, as its main limitation, since its launch and so far only allows prompts in English, despite the fact that DALL- E 2 is able to understand Spanish. The novelty, the truly interesting thing, is its integration into the chatbot.
Microsoft claims that the new feature It has already started rolling out already, although it is not yet available to all users who have access to the preview of the new Bing. Of course, if you are going to check if you already have access to it, make sure to configure the chatbot in Creative mode, since at the moment it is the only one in which it will be available. However, Microsoft has confirmed that it intends to also bring it to Balanced and Precise mode in the future, although without setting a date.