Artificial Intelligence
OpenAI’s Bing Chat Expands to Chrome and Safari Users

According to a report from The Verge, Microsoft’s Bing Chat is no longer confined to the Edge browser and is now being rolled out to users on Chrome and Safari as well. While initially being tested with select users, Microsoft plans to expand access to more users once the standard testing procedures are completed.
The rollout appears to be more than just a small-scale test, as a number of users, including some on Chrome, have already gained access. However, there are certain limitations to be aware of. Users need to be logged into their Microsoft account, and prompts are restricted to 2,000 characters compared to the 4,000 available on Edge. Additionally, conversations reset after five queries instead of 30, and users are frequently prompted to download Edge.
Bing Chat leverages OpenAI’s ChatGPT-4, providing users access to its powerful features without any charges. Despite the similarities, there are some differences between Bing Chat and ChatGPT-4. Bing Chat has access to Bing Search, granting it the ability to offer more up-to-date information and sources for responses. On the other hand, ChatGPT-4 generally delivers more detailed responses. Recently, a paper surfaced claiming that ChatGPT-4’s capabilities have declined, but experts have downplayed the conclusions drawn.
In addition to its expansion to other browsers, Microsoft has introduced a dark mode for Bing Chat. To enable it, users can click on the hamburger menu located at the top right and select “Appearance/Dark,” although this function might not be available to all users just yet.

Apps
Experience the ChatGPT ‘Make It More’ Trend: Generate Weird AI Images

AI art generators occupy a peculiar space, capable of rendering diverse scenarios, from a cosmic skateboarding dog to a coffee cup adrift in the ocean. Setting aside ethical considerations, some creations may not initially meet expectations, necessitating user prompts to refine the AI-generated output.
Yet, what if the aim isn’t to craft a polished piece of AI art? Enter the “make it more” trend, where ChatGPT users task DALL-E to generate an image and subsequently request the bot to amplify certain aspects. For instance, Justine Moore prompted DALL-E to create a bowl of ramen, then instructed it to make it spicier. The iterative process led to increasingly absurd results, culminating in a bowl of noodles shooting fire beams into outer space.
A Dad getting increasingly Dad-ier pic.twitter.com/1EUFECmnT3
— Justine Moore (@venturetwins) November 27, 2023
While the Make It More trend gains traction on ChatGPT and DALL-E, allowing users to experiment with various image transformations, it also raises concerns about resource utilization. The AI processing involved consumes power and energy, prompting considerations about the trend’s efficiency and environmental impact. Nevertheless, these prompts contribute to OpenAI’s ongoing efforts to enhance AI capabilities across platforms like ChatGPT and DALL-E.
Artificial Intelligence
Humane Launches AI Pin, a Wearable Powered by OpenAI

After months of anticipation, Humane has officially introduced its groundbreaking device, the AI Pin. Priced at $699, the wearable consists of a square device and a magnetic battery pack that attaches to clothing or surfaces.
Alongside the upfront cost, users will be required to subscribe to Humane’s service for $24 per month, providing a phone number and T-Mobile data coverage. Preorders kick off on November 16th, with shipments slated for early 2024.
Weighing in at 34 grams, the Pin, equipped with a Snapdragon processor, employs voice control, gestures, a camera, and a small built-in projector for operation. The 13-megapixel camera captures both photos and videos following a software update. Notably, the Pin, unlike the Rewind Pendant, isn’t designed for continuous recording, requiring manual activation through the touchpad.
Powered by AI Mic software, the Pin connects to AI models, incorporating technologies from Microsoft and OpenAI, particularly leveraging ChatGPT access. The operating system, Cosmos, streamlines user interaction by automatically directing queries to appropriate tools, eliminating the need for extensive settings or app management.
Humane’s vision for the Pin is to simplify the user experience, allowing seamless interaction through voice or touch without the clutter of traditional interfaces.
With features like voice-based messaging, email summarization, nutritional information retrieval through the camera, and real-time translation, the Pin aims to function as a wearable, AI-powered search engine. Future plans include adding navigation and shopping capabilities, along with providing developers the opportunity to create their own tools.
Humane sees the AI Pin as a foundational step in a broader project, anticipating advancements as underlying AI models improve. The company envisions the device evolving similarly to smartphones, where improved hardware enhances user experience, but true innovation arises from the expanding range of capabilities. While challenges lie ahead, Humane is poised to embark on this transformative journey.
Apps
ChatGPT Plus Subscribers Can Now Upload and Analyse Files

ChatGPT Plus members can now enjoy new beta features, including file upload and analysis, and multimodal support. Users no longer need to manually select modes like “Browse with Bing” from the GPT-4 dropdown; the chatbot will intuitively discern their needs based on context.
These features bring a touch of the office capabilities available in ChatGPT Enterprise to individual chatbot subscribers. While not all Plus subscribers have received the multimodal update, many have tried the Advanced Data Analysis feature. When a file is uploaded to ChatGPT, it takes a brief moment to process it before becoming operational. The chatbot can then perform tasks such as summarising data, answering questions, or generating data visualisations based on user prompts.
It’s worth noting that ChatGPT can handle more than just text files. In a Threads post, a user demonstrated uploading an image of a capybara and asked ChatGPT, via DALL-E 3, to create a Pixar-style image based on it. They further refined the concept by uploading another image, this time of a wiggly skateboard, and asked the chatbot to incorporate it into the image.