In recent weeks, there have been numerous advancements in the field of artificial intelligence, including the introduction of new and updated AI systems. These developments have the potential to revolutionize the world of information technology.
Artificial intelligence (AI) systems are poised to transform the IT industry, with significant benefits expected in terms of profits, efficiency, product development and more.
With AI, companies can automate tasks that previously required human intervention, reducing operational costs and increasing profits. AI-powered chatbots and virtual assistants can improve customer service, resulting in higher satisfaction rates and customer loyalty.
Let’s examine some AI products that were launched in March 2023.
GPT – 4
On March 13, 2023, OpenAI officially announced the release of ChatGPT-4, the company’s latest and most advanced language model. It is currently available as part of the paid ChatGPT Plus subscription and as an API for developers.
With the ability to create a game in mere minutes, write complex code, and handle up to 25,000 words at a time, GPT-4 has surpassed previous limitations of artificial intelligence systems and accomplished tasks previously deemed impossible.
GPT-4 is more advanced in three key areas: creativity, visual input, and longer context. It is capable of generating more accurate, useful, and safer responses, with 40% more factual responses in OpenAI’s internal testing and 82% less likely to respond to requests for disallowed content.
Stanford Alpaca 7B
In March 2023, the Stanford Institute for Human-Centered Artificial Intelligence (HAI) released a new AI model called Stanford Alpaca 7B. This model is an instruction-following system based on the Meta AI LLaMA 7B.
To develop this AI model, researchers at the institute used the GPT 3.5 by OpenAI to create 52,000 demonstrations of instruction-following. These demonstrations were designed in the style of self-instruction, and they were used to train and instruct the Alpaca model.
It’s worth noting that the researchers have stated that Stanford Alpaca 7B is intended only for academic research purposes. Any commercial use of the model is prohibited.
Claude
Anthropic , an AI startup based in San Francisco has developed an advanced chatbot named Claude, capable of summarizing information, answering questions, assisting with writing, and even generating code. Interestingly, Claude shares many similarities with OpenAI’s ChatGPT, but with additional features such as the ability to customize its tone, personality, and behavior for a personalized experience.
Early feedback from clients suggests that Claude is less likely to produce harmful outputs and is more user-friendly than other chatbots currently available.
Microsoft 365 Copilot
Microsoft, one of the leading technology companies, has recently announced the integration of an AI-driven Copilot feature into its Microsoft 365 suite of applications. This innovative feature, which is powered by OpenAI’s advanced GPT-4 technology, has been designed to assist users in creating various types of documents, emails, presentations, and other tasks with ease.
The Copilot will be presented as a chatbot that appears in the sidebar of Office applications, enabling users to call upon it to generate text in documents, create PowerPoint presentations based on Word documents, and provide guidance on using Excel’s PivotTables feature, among others.
PaLM API
Google has announced the launch of PaLM (Pathways Language Model), a new AI system that aims to compete with OpenAI’s GPT in terms of its ability to generate and edit text, as well as function as a conversational bot.
PaLM has demonstrated impressive capabilities in a variety of challenging tasks, according to Google, and it can assist businesses in generating various forms of content, including text, photos, code, videos, and audio, using natural language prompts.
Google to add GenAI to Google Workspace
In a move to incorporate AI into their productivity services, Google has announced the integration of generative AI systems into Google workspaces. The updated version of workspace, equipped with AI models, is currently available to select testers.
The AI systems offers assistance in document writing and rewriting in Docs, formula generation in Sheets, note-taking in Meet, and content creation, including text, images, audio, and video, in Slides.
PyTorch 2.0
PyTorch Foundation has finally launched PyTorch 2.0 on March 15, 2023 after a long wait. PyTorch 2.0 is an open-source machine learning (ML) training project that is now available to everyone.
This version is a significant upgrade from PyTorch 1.0, which was launched in 2018, and it incorporates years of incremental improvements and contributions from 428 different contributors. The primary focus of PyTorch 2.0 is to improve performance, which is achieved through the introduction of a new feature called Accelerated Transformers.
Midjourney V5
Midjourney, an independent research lab based in the UK, has launched Midjourney V5 – an AI-powered art generator that creates various types of artwork and images based on the user’s input. Unlike traditional photorealistic image generators, this platform focuses on generating artsy images.
Midjourney V5 boasts several improvements, such as a wider range of styles, faster image production, better comprehension of user requests, higher resolution, improved dynamic range, and enhanced depiction of details like facial features and hands. Additionally, it offers an “enlarge” feature that enables users to load larger versions of the generated images.