Microsoft has formalized on social networks the integration of GPT-4 in Bing AI and in Edge.
Microsoft products have been at the forefront of innovation for several months in the integration of artificial intelligence. But with the appearance of ever more efficient and accessible AIs, the lead of the Redmond firm in this area is melting like sherbet in the sun. Indeed, Bard, Google’s AI, even if it arrived later on the market, is already multimodal. It was Mikhail Parakhin, the director of publicity and services of Microsoft himself, who announced this morning on Twitter that Microsoft’s AI takes it to the next level and also becomes multimodal.
To read – ChatGPT has found a serious competitor, what do we know about Claude 2?
Mr. Parakhin says on the social network: “OK, we finally deployed the multimodal/image understanding everywhere (well, almost – the deployment will be completed tomorrow, but 99% complete today)”. OpenAI had launched GPT-4, the “final step in its efforts to extend deep learning” in March 2023. Indeed, this new large language model includes not only texts, but also images. With this new capability, Bing AI will therefore be able to write a poem based on an image or even generate a melody based on a textual description, for example.
Bing AI acquires a multimodal capacity: it recognizes different types of content
Four months after the formalization of the new OpenAI LLM, Microsoft engineers are therefore in the process of implementing it in the company’s products. Search by image in Bing Chat and/or Edge is coming tomorrow… and it’s revolutionary. Despite the enthusiasm that such an announcement may arouse, there remains a few issues that suggest the launch won’t go without a few hiccups.
OK, we have finally rolled out the Multimodal/Image understanding everywhere (well, almost – the rollout will fully be done tomorrow, but 99% today). pic.twitter.com/CIuGIXQNSK
— Mikhail Parakhin (@MParakhin) July 18, 2023
A member of Twitter, who was able to test this new multimodal AI in preview, sent the following complaint to the company representative: “ the only big problem is that if I upload an image and then use the microphone to ask my question, Bing says there is no image. So the picture and the mic don’t work together, I guess”.