Perplexity Labs Launches, Kling 2.1 Tries To Take On Veo 3, EVI 3 Gives You The Perfect AI Voice.

AI News - May 29

Perplexity Launches Perplexity Labs Kling 2.1 gets its full release Sketch out your ideal image with Runway and create your perfect AI voice with EVI 3 Here's today's AI news. Perplexity has launched Perplexity Labs, their Canvas style tool that allows users to create assets as part of the workflow.

[00:18.0]
This could include images, video, code, reports, dashboards and more. With all of Perplexity's competitors offering some kind of canvas feature, it felt like a big missing piece for the service. But now pro users will have access to this feature. Perplexity have given their Labs tool some nice little touches, including the ability to store all of the generated outputs into an asset space for easy reference.

[00:42.0]
With all the top tools offering such similar features, it's the little interface touches that can make a big difference. We've been seeing previews of the capabilities of Kling 2.1 for a little while now, and today the new Generative video model has dropped. The outputs that we've seen so far have been pretty stunning.

[00:59.4]
There are a few different models to choose from so that you can select the price to quality ratio that suits you best, and this goes all the way up to Kling 2.1 Master for the highest quality outputs. From what we've seen, Kling 2.1 seems to be producing visuals that are on par with Veo 3.

[01:15.6]
However, audio still has to be generated in a separate step, so Veo has an advantage in this aspect. Runway has introduced Layout Sketch for its Gen 4 model. You can sketch out any design on a blank canvas or sketch over an existing image to give the model its direction.

[01:33.1]
You can even combine the sketch with other reference images to get the exact composition you're looking for. Greater creative control is always a priority for Generative AI, so these kinds of tools are some of the most valuable to users. Layout Sketch is available now for all plans. And finally, Hume have unveiled EVI 3, their latest speech language model that can generate a custom AI voice tailored perfectly to your requirements.

[01:59.1]
The whole process is incredibly slick. You just talk to EVI 3 and describe what you want your voice to sound like, talking it through the type of characteristics that you expect the voice to have moments later. You have a unique voice character that you can have a little chat with. It's very impressive and you can try the demo now over on the Hume website.