1. ChatGPT Now Lets You Interrupt and Refine Queries
ChatGPT users can now interrupt ongoing queries, making it possible to refine their research or adjust requirements without needing to restart the entire conversation. This new feature enhances efficiency, allowing for more dynamic and flexible interaction with the AI model. Users can update their prompts mid-stream, saving time and improving the overall user experience by enabling real-time adjustments.
Read full story
2. Tinder To Scan Camera Rolls With AI for Better Matches
Tinder is integrating a new “Chemistry AI” feature that will access users’ camera rolls to analyse their photos. This technology aims to understand users’ interests and preferences, thereby improving matchmaking algorithms and suggesting more compatible profiles. While this initiative is designed to enhance user experience and increase engagement, it raises significant privacy concerns among its user base. Additionally, the company is battling potential revenue loss as users spend less time searching for matches manually.
Read full story
3. Google Maps Now Hands-Free with Gemini AI
Google Maps has integrated its advanced AI model, Gemini, to enable a completely hands-free navigation experience. Users can now verbally ask Gemini questions about their route, real-time traffic conditions, and nearby points of interest without needing to touch their device. This update also introduces landmark-based directions, making it easier for users to follow instructions by referencing prominent structures along their journey. This new feature aims to enhance safety and convenience for drivers and pedestrians navigating with Google Maps.
Read full story
4. ElevenLabs CEO Reveals Secrets to Rapid Voice AI Development
The CEO of ElevenLabs shares insights into their revolutionary approach to building advanced Voice AI. The company leverages autonomous teams to rapidly innovate and develop cutting-edge technologies like text-to-speech, AI-generated music, and real-time voice agents. This strategy allows them to efficiently bring new features and products to the market, showcasing a model for accelerated development in the fast-evolving field of artificial intelligence.
Read full story
5. AI Agents Struggle in Secret Microsoft Test
A recent secret test conducted by Microsoft revealed that AI agents performed poorly when tasked with real-world scenarios within a synthetic shopping environment. This “Magnetic Marketplace” simulation exposed significant limitations in current AI agent capabilities, indicating they struggle with the complexities and nuances of actual consumer tasks. The findings highlight the need for further development to enhance AI’s ability to handle practical, everyday demands effectively.
Read full story
6. PewDiePie Builds Private AI with Custom Hardware
Popular YouTuber PewDiePie has reportedly developed a self-hosted artificial intelligence system. He achieved this feat by utilising custom configurations of Graphics Processing Units (GPUs) to run AI models directly on his local setup. This innovative project, dubbed “ChatOS,” allows him to operate advanced AI functionalities privately, demonstrating a unique approach to personal AI infrastructure.
Read full story
7. ChatGPT Suffers Multiple Security Flaws
A recent report has highlighted seven critical security vulnerabilities within the popular AI chatbot, ChatGPT. These flaws could potentially allow malicious actors to manipulate the chatbot’s responses, leading to misinformation, and gain unauthorised access to sensitive user data. Users are advised to remain vigilant regarding the information obtained from the AI model, and developers are working to patch these issues.
Read full story
8. Brain-Inspired AI Cuts Energy Use Drastically
Researchers have successfully developed an innovative Artificial Intelligence (AI) model that significantly reduces energy consumption by mimicking the human brain’s neural networks. This new AI design prioritises essential connections, much like how the brain processes information efficiently, thereby leading to substantial energy savings compared to traditional AI systems. This advancement could revolutionise data centres and AI applications, making them more sustainable and eco-friendly.
Read full story
9. OpenAI Unleashes GPT-2 1.5B Model to Public
Exactly five years ago, on this day in 2019, artificial intelligence research company OpenAI took a significant step by releasing its much-anticipated GPT-2 1.5B model to the public. This full release included the complete code and weights for the language model. The move generated considerable excitement and discussion within the AI community, particularly highlighting the potential of transformer-based language models (LMs). This release was pivotal in advancing research and development in natural language processing.
Read full story
10. Caira: World’s First AI Voice-Controlled Camera
Caira, a revolutionary new camera, integrates a built-in Large Language Model (LLM), making it the world’s inaugural camera of its kind. This groundbreaking feature allows users to control the device and its functionalities through simple voice commands. From capturing photos and videos to performing complex editing tasks, Caira responds intuitively to verbal instructions, offering an unprecedented hands-free photography and videography experience. This innovation promises to redefine content creation, making it more accessible and efficient for users across all skill levels.
Read full story
11. Kimi CLI: New AI Tool for Software Development
Kimi CLI is an innovative new agent specifically engineered to streamline and enhance software development tasks and various terminal operations. This advanced tool aims to provide developers with intelligent assistance, making coding, debugging, and system interactions more efficient and productive. It promises to integrate seamlessly into existing workflows, offering smart solutions for common development challenges.
Read full story
12. LMCache: Boost Your LLM Server Performance
LMCache is a cutting-edge KV cache layer designed specifically for Large Language Model (LLM) serving engines. It significantly enhances performance by reducing the Time-To-First-Token (TTFT) and increasing overall throughput. This innovative solution helps LLM developers and users experience faster response times and more efficient operation from their models.
Read full story
13. Google’s SpeciesNet: AI to Identify Over 2000 Wild Animals
Google has unwrapped SpeciesNet, an advanced artificial intelligence system designed to accurately identify a wide array of wildlife. This innovative AI leverages a powerful combination of MegaDetector, a tool known for detecting animals in camera trap images, and EfficientNet V2, a state-of-the-art image classification model. Together, these technologies enable SpeciesNet to classify over 2,000 different species of wild animals, offering a significant leap forward in biodiversity monitoring and conservation efforts. From spotting elusive big cats to identifying common birds, SpeciesNet promises to streamline wildlife research and tracking.
Read full story
