If there’s one area of technology that’s moving at breakneck speed, that’s it. generative artificial intelligence. Just in the last week, Microsoft has announced the integration of the GPT-4 language model in Office, Google will do the same with its language models in Workspace and Midjourney has released a new version that brings the capabilities of this text-to-image model closer to photorealism.
The tech industry has never been shy when it comes to talking about revolutions, a term that is often used more to promote than to describe, but this does not seem to be the case. The novelties that are reaching users imply profound changes in the way of working, creating and communicatingare constant and between the proliferation of apps It’s easy to get lost in consumer products that make use of different language models for all sorts of tasks. We list ten generative AI applications that you can already use for a wide variety of purposes.
One of the main people responsible for so much talk about generative artificial intelligence. This language model create text responses on virtually any topic. Thus, it serves both to search for information (prior to 2022, since the data set with which it has been trained covers up to 2021) and to create all kinds of written content, from a joke to programming code. It’s been openly available since last November, but if you want to use the most advanced version with the GPT-4 language model, you’ll need to subscribe to ChatGPT Plus, which costs $20 per month, or stick with GPT 3.5.
Although it is not yet available in the open, it is possible to sign up for the beta that Wonder Dynamics is carrying out. Wonder Studio is presented as a visual effects studio accessible through the browser and that allows insert CGI, Computer Generated Images, characters and animations into any video. The AI is in charge of integrating them by automatically adjusting the lighting, composition and animations with results, which is why Wonder Dynamics has been taught, as remarkable as it is easy to obtain.
Wonder Studio from Wonder Dynamics on Vimeo.
It’s about a virtual photo studio in which there are no models or cameras or studio because all content is created by artificial intelligence. Thus, the user only has to select a model among those available in the gallery and provide a description of the image that he wants so that Deep Agency creates it in high resolution with photorealistic models generated by AI. It is also possible to create a virtual model of yourself, although this option is only available in the paid subscription for $29 per month. Deep Agency is currently in open beta phase.
Both of these models do not exist.
You can hire them on https://t.co/6JENJJS0KJ once it’s live. pic.twitter.com/L7BeXK4nuz
— Danny Postma (@dannypostmaa) March 5, 2023
This tool allows create realistic AI voices from text for use in podcasts and videos. The user only has to enter the text or the URL where it is and Fliki will summarize the content, find the right images or video from those available in its library and create a video with a human voice-over (there are more than 900, in 75 languages) and personalized subtitles. It can be used with a free, limited account, or through a paid subscription of which there are three modalities. The superior one, for 66 euros per month if the payment is annual, also allows the cloning of voices.
The most popular among Models of text to image just upgraded to su fifth version with surprising results again. Closer than ever to photorealismThough still with that heavily rendered image look that betrays it, Midjourney’s AI is available for free, with a 25 image limit, on the company’s Discord server. There are three payment methods: Basic for $8 per month, Standard for $24 and Pro for $48, all with the option to expand by buying GPU time at $4 per hour.
midjourney tip: v5 is impressive at doing split images with different angles of a person!
—Julie W Design (@juliewdesign_) March 16, 2023
Runway is a complete video editor whose AI capabilities allow you to generate videos from text instructions, images or a combination of both inputs. You can also stylize images and videos based on another image that is entered, rework images and expand them, among other AI-based editing tools. It has a limited free plan and two paid plans: Standard for $12 a month and Pro for $28 that give access to 1080p quality and higher, among other benefits.
Going viral is what those responsible for this text-generating AI that you can leave in control of your Twitter account propose. With Postwise the user only has to enter the topic you want to talk about so that the AI generates a series of tweets with different approaches and schedule your post. It is a paid tool, $29 a month for the Basic plan and $49 for the Boss, but it offers a 7-day free trial.
We return to Open AI, the company that has developed ChatGPT and DALL-E, to talk about Whisper. It is an automatic speech recognition system that employs language models to transcribe what the user says, translate it into English, or process files to transcribe the audio contained within them. Released last May, the tool is slightly complicated to access through Google Colab, but it’s easier to use a apps like Buzz who uses Whisper and which you can find on Github.
Bing and Edge
After 13 years of indifference on the part of the majority of users, Microsoft seems to have hit the key for its search engine to grow compared to Google. Integration of the OpenAI language model, GPT-4, with the Edge browser and browser has moved out of the beta channel to the stable version in which it is already possible both to chat with the AI visiting Bing and to use it as a writing assistant in Edge. The next disruptive novelty that has already reached the beta channel is the integration with a second open AI model, DALL-Efor creating images from text that has already made it to the Edge beta channel.