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
TikTok parent company ByteDance releases new open source Seed-OSS-36B model with 512K token context
The release of Seed-OSS-36B adds significant capabilities for developers and researchers, particularly with its long-context processing and open-source accessibility, impacting the competitive landscape of AI models.
- ByteDance's Seed Team has launched Seed-OSS-36B, an open-source large language model with a maximum input length of 512,000 tokens.
- The model includes three variants: a synthetic data-enhanced option for higher performance, a non-synthetic version for unbiased research, and an instruction-focused variant.
- Seed-OSS-36B posts strong results on benchmarks for math, coding, and long-context tasks, with competitive performances against other leading models.
- The model's open-source Apache-2.0 licensing allows free, unrestricted use, modification, and redistribution, making it accessible for commercial applications.
- The architecture is designed to handle extensive reasoning tasks and is equipped with a 'thinking budget' feature for tuning performance.
Seed-OSS-36B is part of a new trend where Chinese companies are releasing powerful open-source AI models, with ByteDance's aim to provide robust tools for researchers and enterprises. This follows the recent release of OpenAI's own open-source models, intensifying competition in the AI space as developers seek high-performance alternatives.
Read the original ↗
Don't sleep on Cohere: Command A Reasoning, its first reasoning model, is built for enterprise customer service and more
This development highlights the growing trend of AI models becoming more specialized for enterprise applications, which could significantly enhance efficiency and control for businesses.
- Cohere has launched its first reasoning large language model, Command A Reasoning, designed for enterprise customer service and automation tasks.
- The model supports up to 256,000 tokens and includes 111 billion parameters, enabling advanced processing of multilingual queries across 23 languages.
- Cohere's platform allows enterprises to deploy customized AI agents securely within their infrastructure, ensuring data control while leveraging advanced reasoning capabilities.
- The model's unique token budget feature enables users to optimize performance for various tasks by adjusting the reasoning depth based on needs.
- SAP SE is among the first partners to integrate Command A Reasoning into its SAP Business Technology Platform, enhancing its generative AI functionality.
Cohere, a Canadian startup co-founded by Aidan Gomez, specializes in creating generative AI models for enterprise use. The release of Command A Reasoning reflects the increasing focus on models capable of handling complex business tasks efficiently and accurately. Future developments may include further enhancements to their platform and additional partnerships in various industries.
Read the original ↗
Meta is partnering with Midjourney and will license its technology for 'future models and products'
The partnership between Meta and Midjourney highlights a strategic move to enhance AI capabilities within Meta's ecosystem, potentially influencing the future of AI-generated content.
- Meta has announced a partnership with Midjourney to license its aesthetic technology for future models and products.
- Alexandr Wang, Meta’s Chief AI Officer, emphasized the partnership as part of Meta’s broader AI strategy to integrate advanced capabilities.
- Midjourney remains committed to maintaining its independence despite teaming up with Meta, which could impact its plans for an enterprise API.
Meta is a leading social media company known for platforms like Facebook and Instagram. Midjourney is a San Francisco-based startup recognized for its AI image generation technology, which has 20 million users. The collaboration raises questions about the integration of Midjourney's technology and its implications for both companies' strategies.
Read the original ↗
Developers lose focus 1,200 times a day — how MCP could change that
Reducing context switching for software developers through the Model Context Protocol (MCP) can significantly enhance productivity by streamlining workflows and minimizing distractions in coding environments.
- Developers spend as little as 16% of their working hours on actual coding, with the rest consumed by operational tasks.
- Context switching occurs nearly 1,200 times a day for the average digital worker, disrupting focus and productivity.
- The Model Context Protocol (MCP) connects coding assistants to essential developer tools, which helps reduce context switching.
- MCP is rapidly gaining adoption, with a 500% increase in new servers and an estimated 7 million downloads in June 2024.
The Model Context Protocol (MCP) aims to improve developer productivity by minimizing interruptions caused by context switching between multiple applications. This protocol facilitates integration between AI coding assistants and external tools, allowing developers to streamline their workflows. Future developments may include improved security features and better management of multiple MCP tools.
Read the original ↗
Tencent Hunyuan Video-Foley brings lifelike audio to AI video
The introduction of lifelike audio technology to AI video could significantly enhance the realism and viewer engagement in multimedia content.
- Tencent has launched Hunyuan Video-Foley, a technology designed to create lifelike audio for AI-generated videos.
- The technology aims to improve the auditory experience in AI video production.
Tencent is a leading Chinese technology company primarily known for its social media and gaming services. The Hunyuan Video-Foley launch adds to the company's expanding capabilities in artificial intelligence and multimedia technology.
Based on the headline only — the full article could not be fetched.
Read the original ↗
Midoo AI Review: Meet the First AI Language Learning Agent
Midoo AI introduces a new model for language learning that focuses on continuous engagement and personalized support, potentially transforming how learners interact with language education technology.
- Midoo AI acts as an intelligent language learning agent, adapting in real time to learner needs instead of delivering set lessons.
- The system analyzes a learner's goals and context to create a personalized learning pathway, improving engagement and retention.
- Midoo AI continues to support learners even after sessions end, using data to prepare customized reinforcement tasks for future encounters.
Midoo AI represents a shift from traditional language learning apps, which often rely on static quizzes, to an adaptive, agent-based model. By focusing on real-life communication scenarios and continuous engagement, it aims to enhance the language learning experience. Future developments may include additional features to further personalize the learning process.
Read the original ↗
Beyond Code: What Happens When You Hire AI Engineers for Your Team
Hiring AI engineers can significantly enhance a company's innovation and efficiency, positioning it to leverage AI technology effectively in a competitive landscape.
- As of 2024, 78% of organizations reported using AI in at least one business function, up from 55% the previous year.
- AI engineers design, train, and deploy machine-learning models, automating tasks and deriving insights from data to improve business processes.
- AI engineering roles currently constitute about 22% of U.S. software development job postings, indicating high competition for talent.
- The median total compensation for AI engineers at Meta was approximately $400,000 per year, reflecting the demand for specialized skills.
- A McKinsey report states that 92% of businesses plan to increase AI investment in the next few years, underscoring the growing need for AI expertise.
AI engineers combine software engineering and data science skills to create intelligent systems. Their role is increasingly essential as organizations adopt AI technologies to enhance productivity and innovation. Companies should track labor markets and training needs in AI to remain competitive.
Read the original ↗
What Rollup News says about battling disinformation
The conversation around disinformation is crucial for understanding how technology can help mitigate misinformation in today's digital landscape.
- Rollup News discusses strategies for combating disinformation.
- The piece highlights the role of artificial intelligence in addressing misinformation.
Disinformation refers to false information shared with the intent to deceive. The rise of digital platforms has made addressing disinformation increasingly important, particularly through the lens of artificial intelligence solutions.
Based on the headline only — the full article could not be fetched.
Read the original ↗
CODA: Coordinating the Cerebrum and Cerebellum for a Dual-Brain Computer Use Agent with Decoupled Reinforcement Learning
CODA's development could enhance the effectiveness of autonomous agents in scientific computing, leading to improved performance in complex applications that require both planning and execution abilities.
- CODA is a trainable compositional framework that combines a generalist planner and a specialist executor for improved performance.
- The framework addresses the challenges faced by autonomous agents in scientific computing, where both long-horizon planning and precise execution are essential.
- CODA outperforms existing models by utilizing a two-stage training pipeline that includes specialization and generalization phases.
- The system was evaluated against four applications from the ScienceBoard benchmark, establishing a new state-of-the-art among open-source models.
CODA integrates a generalist planner (Cerebrum) and a specialist executor (Cerebellum) to enhance execution and adaptability. Previous models struggled due to static approaches and limited data availability. Future developments may focus on refining these frameworks for broader applications in various domains.
Read the original ↗
Australia’s Large Language Model Landscape: Technical Assessment
The emergence of localized large language models, like Kangaroo LLM, is essential for addressing specific Australian linguistic and cultural needs, but significant challenges remain.
- Australia lacks a competitive, locally developed large language model (LLM) akin to global models like GPT-4 and Claude 3.5.
- Kangaroo LLM, an open-source project aimed at creating an Australian-specific model, is still in early data collection phases as of August 2025.
- International models dominate Australian research and industry despite limitations in understanding local English and cultural contexts.
Australia's AI landscape is characterized by reliance on international LLMs, with projects like Kangaroo LLM seeking to establish local alternatives. Current efforts focus on application and evaluation rather than foundational model development. The future of Australian AI may hinge on sustained investment and advancements in local model training and deployment.
Read the original ↗
Nous Research Team Releases Hermes 4: A Family of Open-Weight AI...
The release of Hermes 4 signifies a breakthrough in open-source AI models that achieve high reasoning performance without proprietary data, potentially transforming the future of AI development and accessibility.
- Nous Research launched Hermes 4, a family of open-weight AI models with 14B, 70B, and 405B parameters, achieving top-tier performance through innovative post-training techniques.
- The Hermes 4 models can switch between standard responses and deep reasoning using specific tags, enhancing their capability to tackle complex problems.
- Key components include DataForge for synthetic data generation, rejection sampling for quality verification, and a method to control the length of reasoning outputs effectively.
- The 405B model scored 96.3% on MATH-500 and outperformed competitors on various benchmarks, reflecting its remarkable reasoning abilities.
- Training utilized 192 NVIDIA B200 GPUs, showcasing the efficiency of their methods in achieving high performance.
Hermes 4 represents an advancement in the development of open-weight AI models, which are designed to be transparent and reproducible. By using open-source methodologies, Nous Research aims to challenge the dominance of proprietary AI systems, while ongoing enhancements in data generation and model training techniques will be key to its future progress.
Read the original ↗
Maisa AI gets $25M to fix enterprise AI’s 95% failure rate
With 95% of generative AI pilots failing, Maisa AI's funding and new platform aim to provide more reliable AI solutions for enterprises, potentially reshaping industry practices.
- Maisa AI secured $25 million in a seed funding round led by Creandum to enhance enterprise AI solutions.
- The startup's platform, Maisa Studio, focuses on creating accountable AI agents to address AI pilot failures.
- Maisa employs a human-augmented process called HALP to enhance reliability while using AI for automation tasks.
Maisa AI is a startup focused on improving enterprise automation by developing accountable AI agents. The company targets the high failure rates of generative AI in businesses by offering an innovative platform for deploying AI-driven digital workers. Upcoming growth will focus on expanding its customer base and staffing to meet demand.
Read the original ↗
Remaker AI Pricing Review 2025
Understanding Remaker AI's credit-based pricing model is essential for users seeking cost-effective AI image and video editing solutions, enabling them to maximize value without overspending.
- Remaker AI utilizes a flexible credit-based pricing model where users purchase credits for image and video editing tasks rather than committing to a monthly subscription.
- A free plan offers 30 credits for new users to test the platform, although access is limited with features like watermarked outputs and lower resolution.
- Paid credits allow for high-quality outputs, no watermarks, and flexibility in editing without daily limits on the number of projects completed.
Remaker AI is gaining popularity as a platform for AI-driven image and video editing. This credit-based pricing model contrasts with traditional subscription services, offering more flexibility for users. Upcoming pricing changes or promotional deals are anticipated as the market evolves.
Read the original ↗
Nvidia reports record sales as the AI boom continues
Nvidia's substantial revenue growth highlights the increasing demand for AI technologies, emphasizing its central role in the evolving AI market.
- Nvidia reported $46.7 billion in revenue for the second quarter, a 56% increase from the same period last year.
- The company's data center sales generated $41.1 billion, fueled by AI demand for GPUs.
- Nvidia's net income rose to $26.4 billion for the quarter, marking a 59% increase year-over-year.
- The Blackwell chip generation accounted for $27 billion of the data center sales, indicating strong AI infrastructure demand.
- Nvidia has struggled to sell its H20 chips in China, reporting no sales during the past quarter.
Nvidia is a leading technology company known for its graphics processing units (GPUs), which are crucial for AI applications. The firm has seen increased demand due to the AI boom. Future developments to watch include Nvidia's financial performance and regulatory challenges related to sales in China.
Read the original ↗
911 centers are so understaffed, they’re turning to AI to answer calls
The use of AI in 911 centers may alleviate staffing shortages and improve response times for emergency dispatchers, ensuring more urgent calls receive immediate attention.
- Aurelian has developed an AI voice assistant designed to handle non-emergency calls for 911 centers, which have been overwhelmed by staffing issues.
- The company raised $14 million in Series A funding to further develop its AI technology, which triages non-urgent issues and transfers true emergencies to human dispatchers.
- Since launching in May 2024, Aurelian's AI assistant has been deployed in over a dozen 911 dispatch centers across the U.S.
Aurelian aims to reduce the burden on 911 dispatchers by automating non-emergency call handling, a sector that has historically faced high turnover rates. The technology responds to issues that do not require immediate policing, allowing human operators to focus on urgent emergencies. Similar companies are emerging, but Aurelian claims to be the only one currently handling live calls.
Read the original ↗
AI-Powered Phones Get Proactive, Robot Antelope Joins Herd, LLM Environmental Impacts Get Measured
Measuring the environmental impact of AI models offers critical insights for sustainable technology development, guiding both industry practices and policy decisions in AI deployment.
- Google has launched an AI assistant, Magic Cue, for its Pixel 10 smartphone that proactively anticipates user needs.
- Chinese researchers have created a quadruped robot disguised as a Tibetan antelope to study herd behavior.
- The French company Mistral has disclosed the energy, water, and material consumption of its large language model, Mistral Large 2.
AI is increasingly integrated into various technologies, such as smartphones and robotics, raising concerns about sustainability. Companies like Google and Mistral are pushing the boundaries of AI capability while also addressing environmental impacts. The emergence of parallel agents in AI represents a significant advancement in the field.
Read the original ↗
OpenAI co-founder calls for AI labs to safety-test rival models
The collaboration aims to enhance AI safety standards amid increasing concerns over the implications of AI technologies on users' mental health and safety.
- OpenAI and Anthropic opened their AI models for safety testing to identify blind spots in their evaluations.
- OpenAI co-founder Wojciech Zaremba emphasized the growing need for safety standards as AI use increases among millions.
- The testing revealed differences in how the companies' models handle uncertain answers and address sycophancy in responses.
OpenAI and Anthropic are prominent AI labs conducting joint safety testing to collaborate on improving AI safety standards. This move comes at a time when fierce competition and substantial investments in AI technologies raise concerns about potential safety compromises. Future collaborations among leading AI labs are anticipated as they work to tackle emerging safety issues.
Read the original ↗
Google and Grok are catching up to ChatGPT, says a16z’s latest AI report
The latest report reveals shifts in the competitive landscape of AI products, with Google and Grok gaining ground on ChatGPT, highlighting changes in consumer preferences and market dynamics.
- A report by Andreessen Horowitz shows that Google's Gemini and xAI's Grok are closing the gap on OpenAI's ChatGPT in the AI chatbot market.
- Google's products, including Gemini and AI Studio, rose in rankings among top generative AI web products, reflecting their growing user engagement.
- Grok, which launched its stand-alone app in 2024, has rapidly attracted over 20 million monthly active users as of July 2025.
Andreessen Horowitz publishes a regular report assessing the consumer AI landscape, compiling data from third-party sources. The fifth edition of this report highlights how various companies, including Google and xAI, are expanding their AI capabilities to compete with ChatGPT.
Read the original ↗
Here are the 33 US AI startups that have raised $100M or more in 2025
This significant funding trend highlights the ongoing investment and interest in AI technology, which may drive innovation and advancement in various sectors.
- A total of 33 U.S.-based AI startups raised $100 million or more in funding during 2025.
- Key examples include Unconventional AI, which raised a $475 million seed round, and Luma AI, which raised $900 million in Series C funding.
- Anthropic secured a massive $13 billion Series F round, enhancing its valuation to $183 billion.
- Eight startups raised multiple rounds this year, indicating sustained investor confidence in the AI sector.
The AI startup ecosystem has seen robust venture capital engagement, particularly during 2024 with 49 startups securing similar funds. The trend indicates a progressive growth trajectory, with implications for technology development in upcoming years.
Read the original ↗
DrugReasoner: Interpretable Drug Approval Prediction with a Reasoning-augmented Language Model
The development of DrugReasoner increases the accuracy and interpretability of drug approval predictions, which is critical for advancing drug discovery processes and optimizing research investments.
- DrugReasoner is a reasoning-based large language model that predicts small-molecule approval outcomes with high accuracy and interpretability.
- The model outperformed conventional methods, achieving an AUC of 0.732 and an F1 score of 0.729 on the validation set.
- It integrates molecular descriptors with comparative reasoning, providing step-by-step rationales and confidence scores alongside its predictions.
- In testing, DrugReasoner achieved competitive performance against the ChemAP model, with an AUC of 0.728 and an F1 score of 0.774 on an independent dataset.
- This model addresses limitations in traditional machine learning methods by enhancing transparency in AI-assisted drug discovery.
DrugReasoner is a large language model built on the LLaMA architecture and fine-tuned with group relative policy optimization. Traditional machine learning methods have shown limitations in interpretability for drug approval predictions. The success of this model could influence future AI applications in pharmaceutical research.
Read the original ↗