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
DeepSeek LLM: Scaling Open-Source Language Models with Longtermism
The development of DeepSeek LLM offers advancements in open-source language models, potentially reshaping performance standards in code, mathematics, and reasoning tasks.
- DeepSeek LLM develops a dataset containing 2 trillion tokens that is continuously updated.
- The model employs supervised fine-tuning (SFT) and Direct Preference Optimization (DPO) for enhanced performance.
- DeepSeek LLM 67B outperforms LLaMA-2 70B in various benchmarks, particularly in code and mathematics tasks.
- Open-ended evaluations indicate DeepSeek LLM 67B Chat performs better than GPT-3.5.
DeepSeek LLM is part of the growing trend of open-source large language models (LLMs). This trend has seen rapid advancements in LLM performance but has been complicated by varied scaling laws. Future developments in this area may focus on refining these models for improved applications across various domains.
Read the original ↗
Why It’s Time to Demystify AI for SMEs
Understanding how AI can empower small and medium enterprises (SMEs) is crucial for their competitiveness, enabling them to leverage advanced tools previously reserved for larger firms.
- AI now allows SMEs to access advanced capabilities, such as data analysis and predictive modeling, previously available only to larger organizations.
- The shift towards proactive growth enables SMEs to anticipate trends and make more informed decisions instead of being reactive.
- Implementing AI can help SMEs operate more efficiently, thereby allowing better utilization of time, budget, and resources.
Small and medium enterprises often face challenges when competing with larger organizations due to limited resources. As AI adoption increases, SMEs have the potential to enhance their operations and decision-making processes significantly. Future developments in AI's integration into day-to-day activities will further influence the operational landscape for these businesses.
Read the original ↗
AI Agents Are Racing Ahead of Financial Security. Is the Industry Ready?
The rise of AI agents in finance introduces vulnerabilities previously managed by humans, necessitating stringent safeguarding to protect client assets from potential exploitation by these technologies.
- The Model Context Protocol (MCP) enables AI agents to connect with various tools efficiently, enhancing financial transaction processes.
- Recent studies show AI agents are susceptible to prompt injection attacks, compromising their decision-making capabilities.
- Anthropic's Claude Mythos Preview has demonstrated superior offensive capabilities, raising concerns over security in financial settings.
AI agents are being integrated into financial systems to streamline trading and order processing. As technology evolves, the risks associated with connectivity and security practices need to be reassessed. Upcoming regulations, like the EU’s AI Act, will likely impact how these systems are governed in the finance sector.
Read the original ↗
Is Elise AI the Tool Healthcare Needed?
EliseAI aims to significantly improve healthcare's administrative efficiency by automating tasks, which could ease the burden on overworked staff and enhance patient experience.
- EliseAI automates administrative tasks in healthcare, particularly for outpatient specialties like dermatology and women's health.
- The platform claims to handle 95% of patient inquiries without human interaction, focusing on call handling and appointment scheduling.
- EliseAI has raised $250 million in funding and reported over $100 million in annual recurring revenue, indicating market confidence.
- The tool integrates with electronic health records to streamline data management and improve operational efficiency.
EliseAI, founded in 2017 and pivoted to healthcare in 2023, addresses communication issues in medical practices by automating scheduling and billing. While it shows promise for specific specialties, its capabilities in broader healthcare areas remain limited.
Read the original ↗
Robot Dogs, Teslas, and Rescue Helicopters: The UN AI Summit Was a Lot
The focus on AI governance at the UN AI for Good summit underscores the urgent need to address the ethical implications and access disparities inherent in AI technology deployment worldwide.
- The UN's AI for Good summit aims to explore how artificial intelligence can be used to benefit humanity, rather than harm it.
- Discussions at the summit highlighted concerns over the influence of big tech on humanitarian efforts and public policy.
- Activists protested against the use of technology in geopolitical conflicts, emphasizing a need for transparency and accountability in the tech sector.
The AI for Good summit, organized by the UN's International Telecommunication Union, has been held for ten years, focusing on the ethical use of AI technologies. Previous discussions have often centered on AI's potential risks and benefits, with ongoing debates about equitable access and governance in AI technology continuing to grow.
Read the original ↗
How to shrink the token budget without shrinking the team
This story highlights the need for companies to rethink their budget allocation for AI, emphasizing that optimizing operational costs can enhance workforce support and overall productivity.
- Nvidia plans to spend $2 billion yearly on tokens for its engineering team, reflecting a trend where companies prioritize token budgets over workforce stability.
- A survey shows that 80% of executives who cut jobs due to AI have not seen improved financial returns from those reductions.
- Companies are beginning to adopt strategies like prompt caching that can reduce token costs by up to 90%, enabling them to allocate more funds to hiring.
- Studies indicate a blended workforce model, combining AI with human roles, improves customer satisfaction compared to relying solely on AI.
- Hiring junior engineers is becoming critical as the tech industry faces a decline in entry-level opportunities, potentially jeopardizing future talent supply.
The article discusses the increasing trend of companies reallocating budgets from workforce salaries to AI token spending, amid rising capital expenditures in the tech industry. It examines how smart budgeting practices can lead to a more sustainable balance between technology and human resources. Following this trend, companies may increasingly realize the value of maintaining human roles alongside AI advancements.
Read the original ↗
Google Research Introduces SensorFM: A Wearable Health Foundation Model Pretrained on One Trillion Minutes of Sensor Data
The introduction of SensorFM could revolutionize wearable health technology by enabling multiple health predictions from extensive sensor data, potentially improving health outcomes across diverse populations.
- Google Research has developed SensorFM, a foundation model pre-trained on over 1 trillion minutes of data from 5 million consented participants.
- SensorFM is designed to predict 35 health outcomes using aggregated features from various sensors like PPG and accelerometers over a 24-hour period.
- The model significantly outperformed smaller variants in 33 out of 35 health tasks, demonstrating the importance of large-scale data in improving predictive accuracy.
SensorFM represents a significant advancement in wearable health technology, addressing limitations in traditional models that focus on single outcomes. This model uses a large dataset, providing a foundation for more comprehensive health monitoring and prediction in the future.
Read the original ↗
OpenAI says GPT 5.6 is the ‘preferred model’ for Microsoft Copilot 365 amid breakup chatter
The selection of OpenAI's GPT 5.6 as the preferred model for Microsoft Copilot 365 highlights the ongoing integration of AI technologies in workplace applications amidst uncertainty in their partnership.
- OpenAI has announced that its GPT 5.6 model will be the preferred model for Microsoft 365 Copilot.
- This model will power Microsoft’s productivity apps, including Word, Excel, PowerPoint, and Cowork.
- The announcement comes amid reports that Microsoft is exploring the use of its own in-house AI models, known as MAI, to reduce costs.
OpenAI is a leader in AI development, while Microsoft has integrated AI into its productivity software. Recent talks about their partnership becoming strained raise questions about future collaboration. Observers should watch for further developments regarding their business relationship and technology integration.
Read the original ↗
Maine’s Senate Race Implodes, Meta’s Threads Rivals Musk’s X, and the Trump Phone Arrives
The developments in Maine's Senate race and the shipping of the Trump T1 phone reflect significant shifts in political dynamics and public interest in political merchandise associated with Donald Trump.
- Democratic Senate candidate Graham Platner is stepping down amid allegations of sexual abuse and ongoing controversies.
- Meta's Threads reached 500 million monthly users, positioning it as a competitor to Elon Musk's X platform.
- The Trump-branded T1 phone has begun shipping after multiple delays, attracting attention for its features and branding.
Senate candidate Graham Platner's withdrawal follows a series of controversies, highlighting challenges within Democratic campaigns. Threads was created by Meta as a direct alternative to Twitter, aiming to capture users discontented with Elon Musk's leadership of X. The Trump phone, following prolonged anticipation, is seen as a product appealing to his political base and is marketed as an 'all-American' device.
Read the original ↗
OpenAI’s CEO of AGI Deployment Fidji Simo Is Stepping Down
Simo's transition to a part-time advisory role highlights ongoing health challenges while signaling shifts in OpenAI's leadership and product strategy as the company approaches a potential IPO.
- Fidji Simo is stepping down from her full-time role as OpenAI's chief executive of AGI deployment due to health issues stemming from a neuroimmune condition.
- Simo will remain with OpenAI as a part-time adviser following a monthslong medical leave related to her chronic illness diagnosis.
- Since Simo's departure, OpenAI has undergone an executive reshuffle and is reorganizing its product teams ahead of a planned IPO in 2027.
OpenAI, known for developing advanced AI technologies, appointed Simo in 2025 to lead product and business organizations. Recent company changes aim to streamline operations and improve focus on core products like ChatGPT, especially in light of challenges faced by Simo and other executives.
Read the original ↗
OpenAI is shutting down Atlas, but its AI browser ambitions are still growing
The discontinuation of Atlas marks a strategic shift for OpenAI as it consolidates its AI browsing capabilities into existing platforms, aiming to enhance user experience in a competitive landscape.
- OpenAI is shutting down Atlas, its AI-powered browser, launched in October 2025.
- The company will move agentic browsing features from Atlas to the ChatGPT desktop app and a Google Chrome extension.
- The updates allow users to interact with web pages and complete tasks without leaving ChatGPT.
OpenAI is a leading AI research organization known for its language models, including ChatGPT. Previously, the industry has seen an increase in AI-powered browsers as competition with traditional browsers intensifies. Going forward, users can expect more integrated browsing features in OpenAI's applications in response to competitive pressure.
Read the original ↗
An AI agent startup just let its agent run its $100M fundraise
This development highlights the growing capability of AI agents, showcasing their potential to streamline fundraising processes in the startup ecosystem, ultimately facilitating capital investments in technology.
- Lyzr, a startup based in Jersey City, raised $100 million using its AI agent, SivaClaw, to manage the fundraising process.
- SivaClaw interacted with over 130 investors, drafted investment memos, and tracked engagement with presentation materials.
- The company reportedly attracted $400 million in investor interest without founders needing to conduct traditional fundraising meetings.
Lyzr develops AI agents designed for enterprise use. It has gained attention for its innovative approach in leveraging its product for capital raising, reflecting the increased availability of venture capital for AI-focused startups. Moving forward, the efficacy of AI in such functions may influence industry trends in fundraising.
Read the original ↗
OpenAI launches its new family of models with GPT-5.6
The introduction of OpenAI's GPT-5.6 models has significant implications for various industries, particularly in enhancing efficiency and effectiveness in cybersecurity tasks.
- OpenAI has launched GPT-5.6, which includes three variants: Sol, Terra, and Luna, aimed at improving enterprise work, coding, and scientific research capabilities.
- CEO Sam Altman stated that Sol is 54% more token efficient for coding tasks than previous models and is identified as the strongest cybersecurity model yet.
- The models support defensive cybersecurity activities such as threat modeling and code review, responding to previous concerns about misuse.
- OpenAI claims Sol outperforms Anthropic's Fable model based on the Artificial Analysis Coding Agent Index, establishing new benchmarks for coding efficiency and cost.
- The GPT-5.6 models are now available across OpenAI's platforms, with pricing outlined for each variant based on input and output tokens.
OpenAI is a leading artificial intelligence company known for its advanced models. The launch of GPT-5.6 follows competitive advancements from companies like SpaceXAI and Meta. Future developments to watch include enhancements in model capabilities and ongoing competition with emerging AI firms.
Read the original ↗
Fidji Simo steps down from OpenAI’s no. 2 role
Simo's departure creates uncertainty in OpenAI's leadership at a critical time as the company prepares for a possible IPO while competing against rival Anthropic in the enterprise market.
- Fidji Simo is stepping down from her full-time role at OpenAI due to an extended medical leave and will transition to a part-time advisory position.
- Simo joined OpenAI in May 2025 as CEO of Applications, consolidating product and business operations under her leadership until now.
- OpenAI is reportedly seeking a successor for Simo at a pivotal moment as the company anticipates a potential IPO and needs stable leadership.
- Simo faced health issues leading to her medical leave, a condition she disclosed publicly in April 2026, when her leave began.
OpenAI is a prominent AI research organization known for developing advanced artificial intelligence systems, including ChatGPT. Simo's leadership followed a series of executive changes, with OpenAI actively gearing up for a potential initial public offering (IPO). The company is also responding to challenges in its consumer business while working to strengthen its enterprise offerings.
Read the original ↗
AI-generated ads on Google will hopefully get disclosures soon
Understanding the use of AI in advertisements is essential for consumers to make informed decisions about products and services, especially in an era of increasing media manipulation.
- Google is enhancing transparency around AI use in ads by adding a disclosure feature in My Ad Center.
- The 'How this ad was made' section will be available globally on Google's Search, YouTube, and Discover properties.
- If Google's AI tools are used, the disclosure will be added automatically; otherwise, brands can opt to disclose AI use if they choose.
Google's My Ad Center allows users to see how ads are tailored to them. Transparency tools have become a focal point for companies to align with evolving industry standards and improve consumer trust.
Read the original ↗
OpenAI releases a new ChatGPT tool for all things work related
The release of ChatGPT Work enhances productivity by integrating various OpenAI technologies into a unified application, allowing users to manage tasks and collaborate more efficiently.
- OpenAI has launched a new desktop application called ChatGPT Work, combining ChatGPT, Codex, and Atlas into a single interface.
- Powered by the new GPT-5.6 models, ChatGPT Work serves as a general-purpose productivity agent.
- The app supports task scheduling, enabling users to manage assignments remotely via their phones and track progress on a desktop or web browser.
- A unified plugin directory allows ChatGPT to connect to various apps and tools for context-driven task execution.
- OpenAI plans to discontinue Atlas, with the deprecation targeted for August 9.
OpenAI is a leading artificial intelligence research organization known for developing advanced AI models like ChatGPT and Codex. Earlier initiatives included mobile integrations for Codex, but ChatGPT Work marks a significant advancement by consolidating OpenAI's technologies. Users can look forward to further improvements and features as the app rolls out widely.
Read the original ↗
Can AI answer the $3 trillion question?
The projected need for the AI industry to generate $3 trillion highlights significant risks to investors and the broader economy, especially if major tech firms fail to achieve anticipated returns.
- David Cahn estimates that AI infrastructure spending will reach $1.5 trillion by 2026.
- To meet investments in AI infrastructure, the industry must generate $3 trillion in revenue, likely an underestimate due to rising costs.
- Hyperscalers like Google, Meta, Microsoft, and Amazon forecast substantial cash flow increases by 2028, contingent on AI market performance.
David Cahn is a partner at Sequoia who has been calculating the financial implications of AI infrastructure investments. Concerns about the sustainability of AI profitability may influence the strategies of tech giants and could impact the overall economy.
Read the original ↗
Elon Musk praises Mythos/Fable, promises not to ‘cut off’ Anthropic
Musk's endorsement of Anthropic signals a potentially stable partnership worth $40 billion, which could influence the competitive landscape in the AI industry.
- Elon Musk praised Anthropic and assured that he would not terminate their contract to host AI models in a way that would harm them.
- Anthropic signed a deal with Musk's xAI to purchase 300 megawatts of compute for $1.25 billion per month until May 2029, totaling about $40 billion.
- Musk stated that Anthropic is currently a leader in AI, and that its models, Mythos and Fable, have set high standards in the industry.
Anthropic is an AI lab that has gained significant market share and entered a high-value deal with Musk's xAI, which merges with SpaceX. Their partnership represents a strategic relationship within the competitive AI sector, particularly as both companies develop advanced AI systems. Moving forward, their collaboration may reveal more about their operational dynamics and competitive strategies.
Read the original ↗
Humanoid robots controlled by surgeons did world-first operation on live pigs
The successful use of humanoid robots in surgery could enhance access to surgical care in underserved areas, marking a significant shift in medical practices if proven effective for human patients.
- Human surgeons remotely controlled humanoid robots to perform surgeries on live pigs in a preclinical trial published in Nature.
- Two gallbladder removal surgeries were conducted, demonstrating the potential for humanoid robots to assist in surgical procedures.
- The humanoid robots used are significantly cheaper and more compact than existing specialized surgical robots, making them suitable for smaller clinics.
- Current limitations include the need for frequent recalibration and slower operation times compared to established surgical systems.
Humanoid robots are being tested for their feasibility in surgery, with teleoperation allowing skilled human surgeons to control them. While this experiment shows promise, the technology is still in the early stages and requires further development before clinical application.
Read the original ↗
OpenAI wants its new tool to do your work for you and with you
The introduction of ChatGPT Work could significantly change workplace automation, enhancing productivity while allowing for human oversight in complex tasks.
- OpenAI has released ChatGPT Work, a tool designed to automate workflows and stay engaged with projects for extended periods.
- The tool integrates with popular workplace platforms such as Slack, Microsoft Teams, Google Drive, and SharePoint.
- Scheduled Tasks in ChatGPT Work automatically manage repetitive tasks, allowing users to monitor them remotely.
- ChatGPT Work merges existing Codex functionality while introducing a new billing structure similar to Codex, impacting usage costs.
- The rollout accompanies the launch of the GPT-5.6 model, which promises better performance for complex tasks at a cost-effective rate.
OpenAI is a leading artificial intelligence research organization known for its development of language models like ChatGPT. The new ChatGPT Work expands capabilities for workplace automation, following earlier limitations in automated tasks. Future developments will likely focus on refining integration and performance based on user feedback.
Read the original ↗