AI News

OpenAI is introducing a memory feature in ChatGPT that allows the bot to remember personal details from past interactions, enhancing user experience. Users can manage what the chatbot retains and can delete information at any time. Currently in beta, the feature aims for broader release.

Slack has introduced a generative AI toolset to boost productivity, featuring AI-generated channel recaps, quick catch-up for threads, and conversational search capabilities. Future updates will enhance personalized summaries and integrate third-party apps.

NVIDIA has launched a demo version of "Chat with RTX," a locally-hosted chatbot that accesses PC files for personalized responses. It supports various file formats, integrates YouTube content, and operates securely without cloud reliance, though it's limited to Windows PCs with specific GPU requirements. Feedback will improve its functionality.

Methane significantly contributes to global warming, responsible for 30% of temperature increases since the Industrial Revolution. Google and the EDF are addressing this with MethaneSAT, a satellite set to launch, which will use AI to map emissions and provide actionable data to energy firms globally.

Andrej Karpathy, former head of AI at Tesla, announced his departure from OpenAI on X, stating his plan to focus on personal projects. This is his second exit from OpenAI, having previously worked there from 2016 to 2017 and rejoining in February 2023.

Apple's "Keyframer" is a pioneering AI tool that animates static images using natural language prompts, powered by GPT-4. It enhances creative workflows, allowing users to generate CSS animations easily while maintaining control through iterative design. This innovation may transform animation accessibility and redefine creative processes with AI as a collaborative partner.

AI-driven behavioral analysis is transforming Extended Detection and Response (XDR) by enhancing telemetry signal interpretation and threat detection. As CISOs shift from Endpoint Detection and Response (EDR) to XDR, AI integration improves efficiency, accuracy, and cost-effectiveness in cybersecurity strategies while addressing critical security gaps.

Stability AI has unveiled Stable Cascade, a new image generation model that enhances efficiency and flexibility over its predecessor, SDXL. Utilizing a three-stage modular architecture, it excels in image quality, text alignment, and typography while significantly reducing computational costs. Currently in research preview, it's available for non-commercial use.

Bret Taylor has launched Sierra, a startup focused on conversational AI for businesses, partnering with Clay Bavor. Sierra secured $110M in funding and is already serving clients like Weight Watchers and SiriusXM. The platform provides empathetic, always-available AI agents that enhance customer service by integrating with enterprise systems.

OpenAI has introduced a persistent memory feature for ChatGPT, enabling the chatbot to remember user-provided information across sessions. Initially available for select free and Plus users, this aims to enhance interactions by reducing repetition. Users can manage what memory is retained or deleted, with future updates expected for Teams and Enterprise users.

Nvidia is launching Chat with RTX, a feature that enables personalized local AI chatbots on Windows PCs, enhancing user privacy and minimizing data center reliance. Utilizing GeForce RTX 30 Series GPUs, it delivers quick, contextual responses and supports multimedia integration. Users can download the tech demo for free.

Otter.ai is enhancing its platform with "Meeting GenAI" features, improving meeting intelligence by allowing users to query all meeting transcriptions for summaries and insights. The updates emphasize collaboration, enabling multiple users to interact with AI in group chats. Otter aims for proactive AI participation in meetings, providing real-time insights.

OpenAI's GPT-4 model has demonstrated remarkable capabilities in diagnosing eye problems, outperforming junior doctors and closely matching the diagnostic abilities of seasoned eye specialists in a recent study conducted by researchers at the University of Cambridge. The evaluation involved diagnosing 87 clinical scenarios related to various eye conditions, allowing the model to engage in a head-to-head comparison with doctors across different experience levels.

NVIDIA Research's Eureka, an AI agent using GPT-4, outperforms human trainers by teaching robots tasks through effective trial-and-error reward algorithms. It successfully trains various robots and has its algorithms available for public use, highlighting a shift towards AI-driven robotics education.

A 4-year-old boy named Alex, after years of undiagnosed illness despite many specialist visits, received a potential diagnosis of tethered cord syndrome (TCS) from ChatGPT. Following this guidance, a neurosurgeon confirmed the diagnosis, leading to successful surgery. This highlights AI's potential in healthcare but stresses the importance of professional evaluation alongside its use.