- Projects of AI
- Posts
- šUse Glue to Keep Cheese on Pizza
šUse Glue to Keep Cheese on Pizza
WhatsApp Tests AI-Generated Profile Picture Feature
Amandeep Singh
May 24, 2024
Get the most important AI, tech, and science news in a free daily email.
We are testing a new format with Headlines linking to Source and a comprehensive summary of the News.
Your Feedback is Important on it. Kindly Poll
Do You Like the New Format? |
š»ļø
Headlines
TL;DR: Speaking at VivaTech 2024, Elon Musk predicted AI will make most jobs obsolete, making work optional with "universal high income" replacing traditional employment. While AI's rapid advancements are still being adapted to, sectors requiring emotional intelligence are slow to automate. Musk emphasized responsible AI usage and urged parents to monitor children's social media due to manipulative AI algorithms.
TL;DR: Apple is rumored to release foldable MacBooks in 2026 with 20-inch or 18.8-inch screens. Analyst Ming-Chi Kuo predicts over one million units shipped. Collaborating with LG Display for a crease-free screen, Apple aims to use the M5 processor, with production possibly starting by the end of 2024. Previous reports suggest a dual-screen device without a physical keyboard, and Apple may also be developing foldable iPhones and iPads.
TL;DR: NASA and IBM Research co-developed the Prithvi-weather-climate AI model to enhance weather and climate forecasting. Utilizing NASA's MERRA-2 data, it predicts severe weather events, refines local forecasts, and improves climate simulations, functioning well with incomplete data. Available on Hugging Face, it emphasizes open data. Collaboration included NASAās IMPACT, Oak Ridge National Laboratory, and various universities to advance global climate research.
TL;DR: Arc Search, a new app from The Browser Company, introduced Call Arc, an AI-powered voice feature for quick, on-the-go answers via phone calls. Users simply ask questions and receive instant voice responses. Launched in January, Arc Search creates web pages with information from various sources. The Browser Company recently raised $50 million to advance its products.
TL;DR: Alphabet, Meta Platforms, and OpenAI are negotiating with Hollywood studios to license content for AI video generation, offering financial incentives. Studios see cost-saving potential but worry about IP control. Notable deals include News Corp. licensing content to OpenAI for $250 million over five years. Actors and creators seek regulation to prevent unauthorized use of their likenesses, while the music industry is firmly against AI misuse.
TL;DR: Spotify is testing an AI DJ feature that speaks Spanish, expanding from its existing English AI DJ, DJ X. Tech veteran Chris Messina discovered references to "DJ Livi" and language-switching functionality in the code. While no definite release plan is confirmed, Spotify acknowledges routine testing to enhance user experience. With the global popularity of Spanish, particularly in the U.S., this expansion makes sense. Spotify is also exploring other AI applications such as host-read ads and personalized AI playlists.
ā”ļø
Quick Hits
Steve Kramerāthe political consultant who sent AI-generated robocalls to New Hampshire voters, mimicking President Bidenās voice telling them ānot to voteā, ahead of the presidential electionsāfaces a huge fine and criminal charges.
WhatsApp's latest Android beta introduces a feature enabling users to create custom profile pictures using AI from text descriptions, potentially replacing the traditional photo selection process. Discovered in beta version 2.24.11.17, it functions like existing AI image generators, producing varied images from text inputs.
Cohere For AI has launched Aya, featuring models with 23, 8, and 35 billion parameters, available as open weights. Additionally, the latest and most powerful model in the Command R family, named Command R+, has been introduced. These developments were announced by the Cohere For AI team on May 23, 2024.
Google's AI Overviews, a feature generating search summaries, suggested using glue to keep cheese on a pizza, reflecting issues with AI-generated information. Launched in 2024 in the US and UK, it aims to simplify searches but has yielded inaccuracies like confusing the Tower of London with Big Ben. Despite errors, Google asserts most summaries are accurate after extensive testing. A legal suit by Axel Springer and other media groups adds scrutiny to the tech giant.
š
Miscellaneous
OpenAI paused using the AI voice "Sky" for ChatGPT due to criticism that it sounded like Scarlett Johansson's voice, despite being recorded by another actress. Johansson initially declined to license her voice to OpenAI. The decision to pause Sky was made out of respect for Johansson's privacy concerns, despite legal compliance. This action coincides with growing awareness of privacy issues surrounding generative AI and deepfakes.
Anthropic released a research paper on interpreting large language models, focusing on the inner workings their AI model, Claude 3 Sonnet. They found that specific features activate within Claude when encountering relevant text or images, like the Golden Gate Bridge.
TikTok introduced new AI tools in its creative suite, TikTok Symphony, for advertising. These include an AI video generator and assistant to simplify ad creation for brands. Paired with the marketing hub TikTok One, they aim to boost productivity and business outcomes like sales and app installs.
Reply