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
Box CEO Aaron Levie on how AI is changing the enterprise SaaS landscape
This story reveals how the integration of AI agents into enterprise software could transform business models and create new market opportunities for startups, highlighting significant shifts in the tech landscape.
- Box CEO Aaron Levie predicts a future where enterprise software combines SaaS with AI agents for enhanced decision-making and workflow automation.
- Levie emphasizes the need for a clear separation between deterministic and non-deterministic systems to prevent risks, such as data breaches.
- He argues that the traditional 'per-seat' business model will change as AI agents may outnumber human users significantly.
- Startups are positioned to innovate in this evolving landscape, as they can create agent-first processes without altering existing business workflows.
Box is a cloud content management platform that enables businesses to share and manage their files securely. As AI technologies advance, enterprise SaaS companies are exploring how to incorporate AI agents to streamline operations and improve efficiency. The shift towards more AI-focused solutions indicates a significant change in the software development and business sales approaches.
Read the original ↗
Grokipedia vs Wikipedia: Elon Musk’s AI Encyclopedia is Here!
Grokipedia, created by Elon Musk's xAI, presents a significant new alternative to Wikipedia, claiming to provide a bias-free encyclopedia powered by AI, which could reshape how information is accessed and perceived online.
- Grokipedia was launched by xAI on October 27, 2025, and aims to provide a bias-free alternative to existing encyclopedic content.
- The platform currently features around 800,000 articles, significantly fewer than Wikipedia's 65 million articles.
- Grokipedia utilizes the Grok language model to generate and fact-check articles in real-time, aiming for a neutral presentation of information.
- It offers structured, sourced responses on sensitive topics, although it has faced criticism for repetitive and occasionally plagiarized content.
- As of now, Grokipedia does not allow users to edit articles directly or access its API, limiting developer engagement.
Grokipedia is an AI-enhanced online encyclopedia developed by Elon Musk's company, xAI, designed to challenge perceived biases in Wikipedia content. This product builds on the existing landscape of informational resources, striving to present neutral viewpoints powered by AI technology. Future developments may address its current limitations in article coverage and user accessibility features.
Read the original ↗
How AI labs use Mercor to get the data companies won’t share
Mercor's approach to sourcing expert knowledge from former employees highlights significant shifts in data procurement practices for AI labs, potentially reshaping competitive dynamics across industries.
- Mercor connects AI labs with former employees of companies like investment banks to access industry knowledge without formal data contracts.
- AI labs, including OpenAI and Meta, are increasingly relying on Mercor to automate workflows in legacy industries.
- Mercor CEO Brendan Foody, 22, reports that the startup's annual recurring revenue has reached roughly $500 million and is valued at $10 billion.
Mercor is a startup that facilitates the hiring of former employees to provide valuable industry insights for AI training purposes. The company has rapidly grown by filling a gap in data sourcing, particularly for sectors resistant to sharing their proprietary information. Going forward, Mercor aims to expand its partnerships across more industries, leveraging their data for AI developments.
Read the original ↗
ElevenLabs CEO says AI audio models will be ‘commoditized’ over time
The potential commoditization of AI audio models could impact competition and innovation within the industry, shaping how companies develop and utilize these technologies.
- Mati Staniszewski, CEO of ElevenLabs, stated that AI audio models are expected to become commoditized over the next couple of years.
- Despite this foresight, Staniszewski emphasized the importance of building models now as they currently represent a significant advantage in the audio space.
- He noted that in the next year or two, an increasing number of audio models will utilize multi-modal approaches, combining audio with video or language models.
ElevenLabs specializes in developing AI audio technologies, which are currently at the forefront of the industry. As competition intensifies and models evolve, the focus on building proprietary models may shift to more collaborative and open-source methodologies. Industry watchers should consider how this trend may influence the market and the quality of AI-generated audio.
Read the original ↗
Artificial Intelligence (367)
This article highlights the top five open-source models for converting text to speech, a technology that is increasingly important in accessibility and communication tools.
- The article focuses on the top five text-to-speech open-source models.
- These models provide options for developers looking to implement text-to-speech functionality in their applications.
Text-to-speech technology converts written text into spoken voice output, which can enhance accessibility for visually impaired users and improve user engagement in digital content. Open-source models allow developers to modify and deploy these tools in various applications.
Based on the headline only — the full article could not be fetched.
Read the original ↗
Uniform Discrete Diffusion with Metric Path for Video Generation
URSA's innovative approach to discrete generative modeling in video generation could significantly enhance the efficiency and quality of video synthesis technologies.
- URSA is a discrete generative model that improves video generation by using iterative refinement and metric paths.
- It outperforms existing discrete methods and matches the performance of leading continuous diffusion techniques.
- The model supports multi-task video generation, including interpolation and image-to-video conversions, through a unified architecture.
URSA addresses challenges that discrete generative models face, such as error accumulation in video generation. Its advanced techniques enable generation of higher-resolution videos more efficiently. Continued advancements in this area may influence the future of video synthesis technologies.
Read the original ↗
Nvidia hits record $5 trillion mark as CEO dismisses AI bubble concerns
Nvidia's milestone highlights its dominance in the AI industry amid ongoing debates about market sustainability and investment risks.
- Nvidia became the first company to reach a $5 trillion market capitalization after announcing $500 billion in AI chip orders.
- The company plans to build seven supercomputers for the US government as part of its AI strategy.
- Nvidia forecasts shipping 20 million units of its latest chips by 2026, significantly up from 4 million units of the prior generation.
- CEO Jensen Huang dismissed concerns about an AI bubble, citing strong demand for AI services.
- Recent partnerships with companies like Uber and Palantir demonstrate Nvidia's expanding influence in the tech sector.
Nvidia is a leading tech company known for its graphics processing units (GPUs) and AI hardware. The recent valuation surge follows a booming interest in AI technologies post-ChatGPT. Observing market reactions to Nvidia's rapid growth will be essential for understanding industry health going forward.
Read the original ↗
“The Silent Strain We See Today Could Lead to Tomorrow’s Disasters”: India’s Sinking Cities Put Millions of Lives at Risk
India's major cities are experiencing significant ground sinking, threatening millions of lives and critical infrastructure as urban populations grow and natural resources dwindle.
- India's largest cities are sinking, impacting millions of lives and thousands of buildings, according to a study by Virginia Tech scientists.
- Over a third of India's population lives in cities, and groundwater overexploitation is the main cause of subsidence in urban infrastructure.
- The study analyzed satellite radar data from 2015 to 2023, finding 339 square miles of land sinking across five cities that host roughly 80 million people.
- Approximately 1.9 million people live in areas experiencing subsidence rates over 4 mm per year, and 2,406 buildings are at high risk in New Delhi, Mumbai, and Chennai.
- The number of at-risk buildings could exceed 23,000 in the next 50 years if current trends continue.
India is facing severe infrastructural issues due to urban subsidence, primarily caused by excessive groundwater withdrawal. This situation is compounded by increasing climate-related disasters, requiring urgent action in infrastructure management and groundwater policies to prevent impending catastrophes.
Read the original ↗
Train Your AI Agents Like a Pro with Microsoft Agent Lightning (Ful...
The release of Microsoft Agent Lightning could enhance the performance of existing AI agents, making them more adaptable and efficient in handling complex tasks through reinforcement learning.
- Microsoft introduced Agent Lightning, an open-source framework that improves AI agents using reinforcement learning.
- The framework allows developers to wrap existing AI agents with minimal changes to enhance their learning from real-world interactions.
- Agent Lightning's design separates execution from learning, resulting in more effective performance improvements over time.
Microsoft Agent Lightning is a new framework that utilizes reinforcement learning to train AI agents, addressing their limitations in complex task execution. It builds on existing AI frameworks such as LangChain and OpenAI Agents SDK, allowing for easy integration without major changes to existing code. Future developments should focus on how effectively it enhances AI performance in practical applications.
Read the original ↗
Phia’s founders on how AI is changing online shopping
Phia's development of an AI-driven shopping assistant could significantly enhance online shopping efficiency and sustainability, impacting consumer behavior and e-commerce dynamics.
- Phia's founders, Phoebe Gates and Sophia Kianni, aim to revolutionize online shopping by integrating AI technology into their startup.
- The startup's tool, which acts as both a browser extension and app, allows users to compare prices across more than 150 second-hand platforms.
- Phia's AI shopping advisor is in beta testing, helping users gauge the value of items and making informed purchasing decisions based on previous orders and returns.
Phia is an AI startup focused on enhancing the online shopping experience by helping users discover items and understand their value. The service began as a class project and has evolved with a focus on sustainability and efficient consumer decisions. Future developments to watch include the full launch of the AI shopping advisor currently in testing.
Read the original ↗
Character AI is ending its chatbot experience for kids
This shift may set a precedent in the AI industry regarding the safety and interaction of minors with chatbot technology, potentially influencing future regulatory measures.
- Character.AI is ending open-ended chatbot conversations for users under 18 due to safety concerns after two teenagers died by suicide.
- The changes include implementing an age verification system and shifting the platform focus from conversation to content creation.
- The adjustments will begin by November 25, starting with a two-hour daily limit that will eventually be reduced to zero for under-18 users.
Character.AI is an AI role-playing startup facing scrutiny over its chatbot interactions for minors. This decision follows lawsuits and public pressure regarding the mental health impacts of AI chatbots on teenagers. Future developments will likely revolve around compliance with emerging regulations regarding AI and minors.
Read the original ↗
Grammarly rebrands to ‘Superhuman,’ launches a new AI assistant
The rebranding of Grammarly to Superhuman signals a strategic shift towards integrating AI capabilities in a competitive productivity market.
- Grammarly is rebranding itself as Superhuman following its acquisition of the Superhuman email client in July.
- The company is launching an AI assistant called Superhuman Go, which is integrated into Grammarly's extension and offers writing suggestions and app connectivity.
- Superhuman Go allows users to connect with apps like Jira and Google Calendar to perform tasks related to emails and scheduling.
- Grammarly plans to explore rebranding its acquired products, such as Coda, to enhance its product suite's identity and functionality.
Grammarly, known for its writing assistance tools, has shifted its identity to Superhuman after acquiring the email client. This move indicates a commitment to expanding into a productivity suite competing with established platforms that have integrated AI features. Future updates may enhance the functionalities of Superhuman Go and other acquired products.
Read the original ↗
A Call to Action for AI Practitioners
The call for ethical AI practices highlights the role of scientists and engineers in ensuring technology benefits society, especially amid increasing concerns over AI's negative impacts.
- Researchers face challenges in maintaining optimism about artificial intelligence amidst rising authoritarianism and misinformation.
- Public investment in AI has been redirected at the expense of other scientific disciplines, leading to increasing control by Big Tech in the AI sector.
- Scientists can influence AI development positively by advocating for ethical norms and resisting harmful applications of the technology.
The discussion emphasizes the potential of AI to be beneficial while also acknowledging the skepticism within the scientific community about its impacts. Policymakers and researchers are urged to guide AI towards ethical practices to enhance public good.
Read the original ↗
RoboOmni: Proactive Robot Manipulation in Omni-modal Context
RoboOmni demonstrates a significant advancement in robotic manipulation by enabling robots to proactively understand and respond to human intentions without explicit commands, improving real-world interaction efficiency.
- RoboOmni uses a framework that combines auditory and visual signals to infer user intentions for robotic manipulation.
- The system builds upon a new dataset, OmniAction, which includes 140,000 episodes and diverse environmental sounds.
- Experiments show RoboOmni achieves higher intention recognition accuracy and faster inference speed compared to existing models.
RoboOmni is designed to enhance human-robot interaction using Multimodal Large Language Models (MLLMs), addressing limitations of current robotic models that depend solely on explicit instructions. It will be important to monitor further developments in omni-modal frameworks and their applications in various environments.
Read the original ↗
OpenAI Bets Big on AI Researcher by 2028 and $1.4 Trillion Infrastructure Push
OpenAI's plans for a fully autonomous AI researcher by 2028 highlight the potential impact on scientific research and innovation, signaling a significant shift in AI capabilities and infrastructure investment.
- OpenAI aims to develop an AI research assistant capable of performing at an intern level by 2026 and a fully autonomous AI researcher by 2028.
- The company has restructured into a public benefit corporation to increase its capital raising capabilities while maintaining a public-interest mission.
- OpenAI plans a $1.4 trillion infrastructure investment to build 30 gigawatts of compute capacity to support future AI development.
- Current AI systems can perform complex tasks for up to five hours and compete with top human performers in challenges, indicating rapid progress in AI capabilities.
OpenAI, known for its AI development, has shifted from a non-profit to a public benefit corporation to facilitate rapid growth and infrastructure investment. The company's focus on creating advanced AI researchers signals a transformative approach to scientific discovery and innovation in the coming years.
Read the original ↗
Liquid AI Releases LFM2-ColBERT-350M: A New Small Model that brings Late Interaction Retrieval to Multilingual and Cross-Lingual RAG
The release of LFM2-ColBERT-350M signifies a new advancement in retrieval techniques that enhance multilingual and cross-lingual capabilities for various applications.
- Liquid AI has introduced LFM2-ColBERT-350M, a new small model designed for late interaction retrieval.
- This model supports multilingual and cross-lingual retrieval-augmented generation (RAG) applications.
Liquid AI is focused on developing AI technologies that enhance data retrieval and processing efficiency. This release builds on existing models by improving multilingual capabilities, which are crucial in a globalized digital landscape.
Based on the headline only — the full article could not be fetched.
Read the original ↗
THE AI ISSUE
The increasing ubiquity of AI in daily life raises urgent questions about regulation and ethical implications as society navigates this significant technological transformation.
- Hundreds of millions of people are currently using AI technologies, particularly large language models.
- In 2025, the focus of discussions has shifted to the real-world applications and effects of AI rather than its potential.
- AI is now integrated into various settings, including education, therapy, and government, highlighting its pervasive influence.
AI encompasses technologies that can perform tasks typically requiring human intelligence, such as understanding language. Currently, there is minimal regulation guiding its usage. Ongoing discussions are essential as society adapts to the implications of this widespread technology.
Read the original ↗
Mappa’s AI voice analysis helps you find the best job candidates and will show off its tech at TechCrunch Disrupt 2025
Mappa's AI technology aims to enhance the hiring process, potentially reducing turnover rates and promoting diversity in recruitment, which could transform how companies approach candidate selection.
- Mappa is an AI-powered platform that analyzes voice patterns to assess job candidates' traits such as communication style and empathy.
- The startup was founded in 2023 and has raised $3.4 million, serving over 130 customers with an annualized recurring revenue exceeding $4 million.
- Mappa helps companies reduce turnover rates among employees hired through its platform to as low as 2%, compared to the typical 30%.
- More than 60% of hires facilitated by Mappa have been women, LGBTQ+, or immigrants, contributing to a more equitable hiring process.
Mappa aims to minimize biases in hiring by using AI to analyze voice patterns. Initially focused on video submissions, the platform has shifted to voice analysis as the most effective method for evaluating candidates. Future applications may include assessing candidates for loans based on their behavior analysis.
Read the original ↗
Super Teacher is building an AI tutor for elementary schools — catch it at Disrupt 2025
The development of AI-powered tutoring solutions like Super Teacher could significantly enhance educational accessibility for underprivileged students in the U.S., addressing a pressing inequality in private tutoring access.
- Super Teacher's AI tutoring app costs $15 a month, making private tutoring more affordable for families.
- Roughly 20,000 families have signed up for Super Teacher, which is now used in select public schools in New York, New Jersey, and Hawaii.
- The app features animated tutors with AI-generated voices that assist elementary school students with interactive lessons.
Super Teacher is a startup founded by former Google product manager Tim Novikoff, aimed at democratizing access to tutoring for elementary school students. Before launching Super Teacher, Novikoff also founded Fly Labs, which was acquired by Google in 2015. As the app gains traction, future expansions are planned to include older grades and more school partnerships.
Read the original ↗
Inside CampusAI’s mission to close the AI training gap for everyday workers — check it out at TechCrunch Disrupt 2025
Closing the AI training gap for everyday workers is crucial as employees increasingly need to leverage AI tools to improve productivity and stay competitive in the workforce.
- CampusAI offers an educational platform designed for non-technical people to learn how to integrate AI into their workflows.
- The startup provides courses and a metaverse-based virtual campus where users can develop skills and connect with others.
- CampusAI claims its courses improve employee efficiency by 40% and job satisfaction by 60%, catering to both individual and corporate users.
CampusAI is a Polish startup focused on AI education, providing learning paths tailored for various professions. It has gained rapid traction since its launch in 2023 and aims to expand into 40 markets by 2030.
Read the original ↗