The day's AI news, briefed
The Daily Prompt
Our AI editor reads each day's AI and robotics stories, pulls out why they matter, and links every original source — so you can keep up in minutes.
Join over 250,000 subscribers getting the latest AI news
Latest AI & Robotics News Today
DirecTV screensavers will show AI-generated ads with your face in 2026
The introduction of AI-generated ads featuring users' faces represents a significant shift in how streaming services utilize advertising, which may affect customer satisfaction and revenue streams.
- DirecTV will begin showing AI-generated screensaver ads on its Gemini streaming devices in early 2026.
- Users can create a personalized AI avatar, which will be used to tailor product recommendations through the screensavers.
- The ads will display after the TV has been idle for 10 minutes, replacing the current Google wallpapers on Gemini devices.
DirecTV, a satellite TV provider, is adapting its services to include an advertising model amidst declining subscribers. Glance, the partnering ads company, specializes in AI-driven content and previously provided similar technology for Samsung Galaxy phones. As streaming services increasingly pivot towards ad revenue, DirecTV's new strategy symbolizes broader industry trends.
Read the original ↗
Google will let Gemini schedule meetings for you in Gmail
The introduction of the Help Me Schedule feature in Gmail could significantly improve productivity for users by simplifying the meeting scheduling process through AI assistance.
- Google has unveiled a new Gmail feature called 'Help Me Schedule' which utilizes Gemini AI to suggest meeting times based on email context.
- The feature offers an inline meeting widget that allows recipients to choose a suitable time, streamlining the scheduling process.
- Initially, Help Me Schedule will not support group meetings, focusing only on two-party interactions for now.
Gemini is Google's generative AI technology that enhances various products, including Gmail. This new feature is part of a broader trend of integrating AI to improve productivity tools, with a gradual rollout expected to begin in late October or early November.
Read the original ↗
Mozilla’s Firefox adds Perplexity’s AI answer engine as a new search option
The integration of Perplexity's AI search engine into Firefox allows users to access conversational search results, potentially reshaping how individuals find information online.
- Mozilla has added the AI answer engine Perplexity as a search option in its Firefox browser, available globally for desktop users.
- Perplexity offers cited answers in a conversational format, distinguishing it from traditional search results provided by engines like Google.
- The integration follows positive user feedback from previous tests in select markets including the U.S., U.K., and Germany.
- Perplexity will be available for mobile users in the coming months, expanding its reach further.
- Mozilla also announced broader availability of browser profiles, allowing users to switch between different setups for various purposes.
Mozilla's Firefox is a widely used web browser known for its emphasis on user privacy. The addition of Perplexity comes amidst a growing trend of AI integration into web browsing. Future developments may see more AI tools integrated into Firefox based on user feedback.
Read the original ↗
Nvidia sells tiny new computer that puts big AI on your desktop
The DGX Spark offers a compact, high-performance AI workstation that could empower developers to run complex AI models independently, reducing reliance on cloud services.
- Nvidia has introduced the DGX Spark, a $4,000 desktop AI computer capable of one petaflop of performance and featuring 128GB of unified memory.
- The system can handle AI models with up to 200 billion parameters locally, which exceeds the capabilities of typical consumer PCs.
- Nvidia will start taking orders for the DGX Spark on October 15, both through its website and select retail outlets in the US.
- This product addresses the challenge of running larger AI models without resorting to cloud services, which can be costly and limited.
- The DGX Spark runs on Nvidia’s GB10 Grace Blackwell Superchip and includes an ARM-based Ubuntu Linux operating system optimized for GPU processing.
Nvidia, known for its graphics processing units (GPUs), has launched the DGX Spark as a desktop workstation designed to facilitate AI development. Previously, AI developers faced limitations with standard PCs, often needing to resort to cloud services. The success of the DGX Spark could determine the viability of desktop AI computing against cloud alternatives.
Read the original ↗
OpenAI unveils “wellness” council; suicide prevention expert not included
The formation of OpenAI's Expert Council on Wellness and AI highlights the critical need for improved safety measures in AI applications, especially regarding children's mental health.
- OpenAI established the Expert Council on Wellness and AI to enhance the safety of ChatGPT after a lawsuit linked the platform to a teen's suicide.
- The council consists of eight experts focused on how technology impacts emotional and mental health, including children's development.
- Key members include David Bickham and Mathilde Cerioli, who research the effects of technology on youth mental health but do not specialize in suicide prevention.
- OpenAI's previous consultations with experts aimed at implementing parental controls and ensuring safe interactions for young users.
- The council will explore both the risks and potential benefits of AI technologies for user well-being during recurring meetings.
OpenAI is a leading artificial intelligence research company known for developing ChatGPT, a chatbot that engages users in conversation. The formation of this Expert Council follows scrutiny regarding AI's impact on mental health, particularly among children. Moving forward, observers will look for how these efforts shape safeguards against risks and enhance positive interactions with the technology.
Read the original ↗
Andrej Karpathy Releases ‘nanochat’: A Minimal, End-to-End ChatGPT-Style Pipeline You Can Train in ~4 Hours for ~$100
This development provides an accessible tool for quickly training a conversational AI model, which could democratize AI development for users with limited resources.
- Andrej Karpathy has released 'nanochat', a minimal pipeline for training AI chat models.
- The pipeline is designed in the style of ChatGPT and takes approximately four hours to train.
Andrej Karpathy is known for his work in artificial intelligence and neural networks. 'nanochat' aims to simplify the process of developing chatbots by providing an efficient training pipeline. This release could impact the field by lowering entry barriers for AI developers.
Based on the headline only — the full article could not be fetched.
Read the original ↗I sent ChatGPT Agent out to shop for me
ChatGPT Agent's slow performance and glitches highlight challenges in advancing AI capabilities for automated tasks, impacting user reliance on AI tools for everyday functions.
- OpenAI introduced ChatGPT Agent as an AI tool designed to automate complex, multi-step tasks using its own virtual computer.
- The agent has a subscription cost of $200 per month and has faced delays in rollout due to high demand.
- During testing, the agent took 50 minutes to find options for a vintage lamp but failed to add items to the user's actual Etsy cart.
- ChatGPT Agent cannot assist with any sensitive financial transactions, focusing instead on consumer purchases like groceries and flowers.
- Users have reported that the agent's task history does not always save, causing confusion about its actions and outputs.
OpenAI's ChatGPT Agent combines features from previous tools and aims to handle various consumer tasks more efficiently. Initial responses indicate it struggles with reliability and pacing, suggesting users still need to remain engaged. Future updates are expected to address these issues and expand its functionalities.
Read the original ↗
Sheryl Sandberg-backed Flint wants to use AI to autonomously build and update websites
Flint aims to revolutionize website creation and updates using AI, potentially changing how marketers engage with customers and optimize online content.
- Flint, co-founded by Michelle Lim and Max Levenson, has raised $5 million in seed funding led by Accel and supported by Sheryl Sandberg's investment.
- The platform allows users to create self-updating websites that can generate designs and layouts automatically, though content writing functionality is still a year away.
- Current capabilities enable Flint to analyze existing websites and deploy fully coded pages consistent with their design, completing the process in about a day.
Flint is an AI platform focused on automating website creation and updates, targeting marketers in startups and large corporations. Previously, website design was a resource-intensive process, often taking months. Future iterations aim to include AI-generated content, further enhancing functionality.
Read the original ↗
Google updates Search and Discover with collapsible ads, AI features, and more
Enhanced navigation and ad control in Google Search and Discover could influence user engagement and advertising effectiveness, as well as shape the future of digital advertising strategies.
- Google is implementing changes to its Search and Discover pages, introducing AI-powered features and navigation improvements.
- Users can now collapse ads in Google Search by selecting a new button, but the 'Sponsored' label will remain visible throughout scrolling.
- New sports updates and AI-driven topic previews will launch first in the U.S., South Korea, and India.
Google Search and Discover are essential platforms for information retrieval and advertising. Recent updates aim to enhance user experience by incorporating AI while maintaining visibility for advertisers. Continued changes can be expected as Google adapts to user needs and evolving digital landscapes.
Read the original ↗
OpenAI and Broadcom partner on AI hardware
This partnership between OpenAI and Broadcom highlights a significant investment in AI infrastructure, potentially impacting the future development of advanced AI technologies.
- OpenAI will purchase 10 gigawatts of AI accelerator hardware from Broadcom to enhance its data centers.
- Deployment of this hardware is planned to occur from 2026 to 2029.
- The estimated cost of the deal could be between $350 billion and $500 billion, according to the Financial Times.
- This agreement follows OpenAI's recent infrastructure deals, including a $300 billion cloud deal with Oracle, not yet confirmed by the companies.
OpenAI is a leading AI research lab focusing on developing advanced machine learning models. The partnership with Broadcom is part of OpenAI's broader strategy to enhance its operational capabilities through significant hardware investments. Ongoing developments in AI infrastructure indicate increasing competition among tech giants.
Read the original ↗
You’ll soon be able to shop Walmart from ChatGPT
This integration will change how consumers engage with online shopping, making it more convenient and personalized through AI technology.
- Walmart is partnering with OpenAI to allow users to shop for products through the ChatGPT chatbot.
- Customers can link their Walmart accounts to ChatGPT to browse items and checkout instantly, with third-party products also supported.
- The new feature will enable Sam's Club members to plan meals, restock essentials, and discover new items via the AI.
Walmart seeks to improve online shopping experiences by leveraging AI capabilities through its partnership with OpenAI. This follows OpenAI's broader push into e-commerce and Walmart's recent launches of AI tools like the generative shopping assistant, Sparky.
Read the original ↗
Coco Robotics taps UCLA professor to lead new physical AI research lab
The establishment of a physical AI research lab at Coco Robotics could enhance the automation capabilities of delivery robots, impacting efficiency and cost in last-mile logistics.
- Coco Robotics has appointed UCLA professor Bolei Zhou as chief AI scientist to lead a new physical AI research lab.
- The lab will focus on utilizing five years’ worth of delivery data to improve autonomous navigation for the company's robots.
- Coco Robotics aims to enhance its delivery bot services by leveraging its data for better automation and efficiency.
- Previously, teleoperators assisted the bots, but the company now seeks full autonomy in its delivery operations.
- The startup has collaborated with OpenAI to access AI models while maintaining exclusive control of its robot-collected data.
Coco Robotics, founded in 2020, focuses on last-mile delivery with its autonomous robots. It has previously relied on human operators to navigate urban environments. The new lab's research is expected to further accelerate the company's efforts to automate deliveries.
Read the original ↗
Google Meet launches an AI-powered makeup feature
The introduction of AI-powered makeup features in Google Meet signifies a growing trend in virtual communication, enhancing user experience and increasing competition among video-conferencing platforms.
- Google Meet has launched a new AI-powered makeup filter designed to provide users with virtual makeup options during video calls.
- The feature offers 12 different makeup choices and can be found in the ‘Appearance’ section under ‘Portrait touch-up’ alongside existing options.
- The virtual makeup feature remains fixed regardless of user movement, ensuring a natural appearance throughout the video call.
- It is set to remember a user's makeup preferences for future meetings and is disabled by default, requiring user activation.
- The feature began rolling out to Google Meet users on October 8, 2025, for both mobile and web platforms.
Google Meet is a video conferencing tool developed by Google that allows users to connect remotely. This new AI-powered makeup feature is designed to compete with similar offerings from other platforms like Microsoft Teams and Zoom. Future developments may focus on enhancing user personalization and functionality in virtual environments.
Read the original ↗
Boeing Unveils New “CxR” Tiltrotor: A Drone Wingman Aiming to Redefine Army Aviation
Boeing's unveiling of the CxR tiltrotor could significantly impact future military aviation by introducing advanced unmanned systems that enhance operational efficiency and safety.
- Boeing introduced the Collaborative Transformational Rotorcraft (CxR) concept at the AUSA 2025 conference on October 13.
- The CxR is a modular unmanned tiltrotor designed to operate alongside crewed helicopters, including the AH-64 Apache.
- Projected to weigh between 5,000 and 7,000 pounds, the CxR will have a payload capacity of 1,000 to 2,000 pounds and speeds of 200 to 250 knots.
- Two variants, the Collaborative Combat Rotorcraft and Collaborative Logistics Rotorcraft, will support combat and transport roles respectively.
- Boeing aims to integrate the CxR with existing Army systems, extending its capabilities within the Army’s Future Vertical Lift roadmap.
The CxR is part of Boeing's new effort to modernize army aviation with unmanned systems. This follows the Army's increased focus on unmanned missions after shelving its Future Attack Reconnaissance Aircraft program. Further development and feedback from the Army are needed before moving to prototyping.
Read the original ↗
Aquawise will show off its AI-driven water quality tech at TechCrunch Disrupt 2025
Aquawise's AI technology could provide cost-effective solutions for monitoring water quality, significantly impacting aquaculture productivity in Southeast Asia, a region struggling with traditional methods.
- Aquawise uses satellite images and a physics-based AI model to monitor water quality at aquaculture farms, without requiring hardware purchases.
- The technology allows for continuous monitoring, unlike traditional methods that check water quality daily or weekly.
- Founded in 2024, Aquawise aims to address the estimated $30 billion annual losses due to water quality issues in aquaculture, especially in Southeast Asia.
Aquawise is a startup based in Bangkok focused on improving water quality monitoring for aquaculture farms using AI and satellite data. Traditional methods are often too expensive for farmers in developing regions. The company will showcase its technology at TechCrunch Disrupt 2025, highlighting its potential benefits and plans for growth.
Read the original ↗
Stable Video Infinity: Infinite-Length Video Generation with Error Recycling
The development of Stable Video Infinity (SVI) allows for the creation of endlessly long, coherent videos, which could transform content creation across various media formats.
- Stable Video Infinity generates infinite-length videos with high temporal consistency, plausible scene transitions, and controllable storylines.
- SVI employs Error-Recycling Fine-Tuning to improve video generation by correcting its own errors during the creation process.
- The technology can effectively scale videos from seconds to infinite durations without additional inference costs.
Stable Video Infinity (SVI) is a new video generation model that utilizes a Diffusion Transformer to create long videos with consistent narrative quality. It addresses limitations of previous methods by recycling errors to enhance output. Future applications may involve diverse content types, including animations and films.
Read the original ↗
InternSVG: Towards Unified SVG Tasks with Multimodal Large Language Models
The development of InternSVG could significantly enhance capabilities in handling SVG tasks, which are essential for graphic design, web development, and data visualization.
- InternSVG is a unified multimodal large language model (MLLM) designed for SVG understanding, editing, and generation.
- The model leverages the SAgoge dataset, the largest for SVG tasks, which includes static graphics and dynamic animations.
- SArena, a benchmark accompanying InternSVG, provides standardized task definitions for evaluating SVG tasks.
- InternSVG employs a training strategy that progresses from simple to complex SVG tasks, improving its overall performance.
SVG, or Scalable Vector Graphics, is a format for rendering two-dimensional images using XML. Current challenges involve fragmented datasets and difficulties in task transferability. The introduction of InternSVG represents a move toward more efficient and versatile SVG modeling.
Read the original ↗
OpenAI wants to stop ChatGPT from validating users’ political views
OpenAI's efforts to reduce political bias in ChatGPT could shape user perceptions of AI neutrality and influence how such technology is employed in politically sensitive contexts.
- OpenAI released a paper outlining methods to reduce political bias in ChatGPT, aiming for the model to avoid acting like it has personal political opinions.
- The company found that its latest GPT-5 model shows 30 percent less bias than previous versions, with reported political bias in less than 0.01 percent of responses.
- OpenAI's evaluation focuses on reducing behaviors such as amplifying users' emotional political language and providing one-sided coverage of contested topics.
- The adjustments are designed to prevent ChatGPT from validating user perspectives, particularly in emotionally charged political discussions.
- OpenAI associates the timing of its research with a recent executive order from the Trump administration demanding ideological neutrality in federal AI contracts.
OpenAI is an artificial intelligence research organization known for developing AI language models like ChatGPT. The company is currently facing scrutiny over the political biases in its systems, particularly in response to government pressures for ideological neutrality. Future developments should be monitored to see how OpenAI's approaches affect user interactions and the broader discourse on AI neutrality.
Read the original ↗
Why We Love to See the Mighty Fall: The Science Behind Humanity’s Obsession With Seeing Others Fail
Understanding schadenfreude can illuminate how societal perceptions of success and failure affect social relationships and emotional health.
- Schadenfreude refers to the pleasure derived from others' misfortunes, particularly among those who are wealthy or influential.
- Research indicates that envy, especially malicious envy, triggers schadenfreude, leading individuals to take joy in the failures of those they perceive as superior.
- Belichick's recent struggles are seen as poetic justice by some, reflecting how public sentiment shifts toward satisfaction in the downfall of admired figures.
Schadenfreude, a term from German meaning 'malicious joy', reflects complex human emotions surrounding envy and social status. Research has shown varied factors influence feelings of schadenfreude, revealing the social dynamics of success and failure. Understanding these emotions highlights their impact on personal relationships and mental health.
Read the original ↗
“Drilling to This Depth is Unprecedented on Mars”: How the ESA’s New Robotic Rover Will Take the Search for Alien Life to New Depths
The ExoMars Rosalind Franklin Rover's unprecedented drilling capability may significantly enhance the search for signs of past or present life on Mars.
- The ExoMars Rosalind Franklin Rover, launching in 2028, can drill two meters below the Martian surface to collect potential evidence of life.
- The rover, designed by the European Space Agency, targets subsurface material that may contain organic compounds, shielded from surface radiation.
- The Enfys infrared spectrometer, developed in the UK, will help select drilling sites and analyze samples in tandem with another device called PanCam.
The ExoMars mission represents a significant advancement in planetary exploration, focusing on areas likely to harbor evidence of life. The choice of drilling depth is a strategic balance of scientific goals and the rover's mechanical capabilities. Future developments include rigorous testing of components like the Enfys spectrometer in simulated Martian environments before launch.
Read the original ↗