As generative AI technology continues to advance, new tools are emerging that push the boundaries of what's possible with it. From Cheat Layer's custom-trained GPT-4 models that act as personal AI software engineers to Banter AI's ability to let you talk to celebrities and experts, we take a look at 5 new and trending generative AI tools this week. Cheat Layer “Cheat Layer is a machine learning tool that helps automate businesses by providing custom-trained GPT-4 models to act as personal AI software engineers,” reads the product’s website. In other words, Cheat Layer is an AI tool that lets you create automation by connecting various tools together. It utilises Project Atlas, which builds entire automations from end to end using prompts in simple language. The demo video on the website shows what happens when you ask Cheat Layer to “Post about AI automation to Twitter every day at 10 am.” The tool does just that flawlessly. WisdomAI WisdomAI is a conversational assistant powered by GPT-4. But unlike ChatGPT that’s been trained on data from around the web, WisdomAI lets you create chatbots trained on audio and video content selected by you. You can either upload your own content or link it to various media platforms like podcasts, YouTube, Facebook, Zoom and others. WisdomAI then analyses the content library from different sources and creates a conversational response with the key points. The company behind it, Searchie, says that the tool enables creators “to easily create a conversational assistant for their digital products, including courses, memberships, coaching programs, and podcasts.” Banter AI Banter AI is trained on the conversations of over 100 celebrities and experts, letting you have conversations with the likes of Elon Musk or Steve Jobs. The AI will exhibit some of their personality, while also retaining their voice and conversational style. The only catch is that there will be a delay of a few seconds before they respond as the AI processes your speech. Segment Anything (SAM) On April 6, Meta announced a new AI model called Segment Anything Model (SAM) that can “cut out” any object in any image with a single click. It differs from iOS’s photo cutout feature in that you ask SAM segment anything you want using different input prompts. SAM can take various input prompts, such as points, boxes, text, or even a user's gaze from an AR/VR headset to select an object, and generate output that can be used for other applications, such as video tracking, image editing, or 3D reconstruction. SAM is trained on a huge dataset of over 1 billion masks on 11 million images, and it can handle unfamiliar objects and images without any extra training. Meta says that SAM uses advanced computer vision technology that allows computers to analyse and comprehend images/videos just the way humans perceive them. Koe Recast Koe Recast is a simple but entertaining tool that transforms your voice into a different one. The example on the site shows a 15-second clip of Mark Zuckerberg on a podcast alongside altered clips where his voice is changed to a narrator, a female, and an anime character. The tool is available as an app you can download.
