Short

Midjourney has released an update to version 7 of its image generation model. According to the company, the new version delivers improved image quality, more accurate rendering of hands and bodies, and better alignment between text prompts and output. The update also includes a redesigned user interface for the image editor. Users now have direct access to functions like "Vary" and "Upscale," along with a right-hand image preview and refined tools such as intelligent segmentation. A new parameter, --exp, has been added to control image aesthetics. Higher values can produce more detailed and dynamic visuals, but may reduce the accuracy of the prompt interpretation. Recommended values range from 5 to 50.

Ad
Short

Jack Krawczyk, who led product development for Google’s Bard and Gemini AI projects, is taking on a new role at Meta. The move comes as Meta introduces its standalone Meta AI app for the first time at its developer conference. On LinkedIn, Krawczyk outlined his perspective on building AI assistants.

Personality is product. People want to talk to someone that resembles a friend or a coach; they don’t want a sanctimonious assistant. Insightful, knowledgeable, non-judgmental, humble, and a little witty are table stakes for excellent interactions. Trust hinges on guiding people to their own informed conclusions—not telling them what or how to think.

Ad
Short

NotebookLM’s "Audio Overviews" feature is now available in approximately 75 languages, including less commonly spoken ones such as Icelandic, Basque, and Latin. The audio for each language is generated by AI agents using "metaprompting," with the Gemini 2.5 Pro language model as the underlying system. At the same time, Google is moving to an audio production technology based entirely on Gemini’s multimodality, a development that does not bode well for providers focused exclusively on audio models.

As with AI-generated text, audio created by language models can also contain inaccuracies. This issue is especially pronounced in AI-generated podcasts, where large amounts of audio may be produced from minimal source text, and the conversion from text to dialogue constitutes a significant alteration of the original material.

Ad
Google News