Brief

The AI development landscape has shifted dramatically, with coding agents now capable of sustained, long-horizon tasks, a change noted by Andrej Karpathy since December 2025. This has led to new products like Perplexity Computer, an orchestration-first agent system, and advancements in tools like OpenAI's GPT-5.3-Codex and GitHub Copilot CLI. However, this rapid progress has also fueled a "productivity panic" and a form of "AI psychosis" among executives and VCs, who are investing heavily in agentic workflows and tools that may not yield measurable value. AI

Google DeepMind has launched two new autonomous research agents, Deep Research and Deep Research Max, powered by Gemini 3.1 Pro. These agents are designed to securely analyze user-provided or third-party data, with Deep Research optimized for speed in interactive applications and Deep Research Max for more exhaustive background tasks. Both agents can generate professional, cited reports and present-ready visuals, with Deep Research Max being the first to natively produce such graphics. AI

Together AI has enhanced its fine-tuning platform to support a wider array of large language models, including recent releases from DeepSeek, Qwen, and Meta, alongside OpenAI's gpt-oss. The platform now offers expanded context lengths, up to 131k tokens for some models, at no additional cost, facilitating tasks like long-document processing and complex code editing. Separately, Together AI researchers have explored LLM behavior using minimal, topic-neutral prompts to uncover inherent model preferences, finding that GPT-OSS favors programming and math, Llama leans literary, DeepSeek often produces religious content, and Qwen tends toward multiple-choice questions. AI

Several AI companies have achieved significant funding milestones, with ElevenLabs securing $500 million in Series D funding at an $11 billion valuation and Cerebras raising $1 billion in Series H at a $23 billion valuation. Google is integrating its Gemini 3 model across its products, including a new Chrome side panel, and reporting substantial adoption and cost reductions in serving the model. The coding assistant landscape is evolving with VS Code and GitHub Copilot introducing support for multiple agents, including Claude and OpenAI Codex, to enhance developer workflows. AI

Snowflake and OpenAI have announced a significant multi-year partnership, involving a $200 million investment, to integrate OpenAI's advanced AI models directly into Snowflake's data platform. This collaboration will enable enterprise customers to build AI agents and derive insights from their data using models like GPT-5.2 within Snowflake Cortex AI. The integration aims to simplify AI deployment for businesses, allowing them to leverage their existing data infrastructure securely and efficiently. AI

Venture capital firm Sequoia Capital has invested in Flapping Airplanes, a startup aiming to develop data-efficient AI models. The company, founded by brothers Ben and Asher Spector along with Aidan Smith, focuses on attracting raw talent and pursuing research-driven breakthroughs rather than solely scaling existing models. Flapping Airplanes believes that data efficiency, not just compute power, is the key bottleneck for achieving artificial general intelligence (AGI). AI

ServiceNow has entered a multi-year agreement to integrate OpenAI's advanced models, including GPT-5.2, into its enterprise workflow platform. This partnership aims to provide businesses with AI capabilities that can understand, decide, and act within their secure infrastructure, automating complex processes across various departments. The integration will enable features like natural language querying of enterprise data, AI-powered summarization, and intelligent search, facilitating end-to-end task execution with minimal human intervention. AI

Anthropic has launched "Cowork," a unified product suite that bundles prior work like Claude Code and Claude for Chrome into a single, cohesive agent experience. This initiative is part of a broader reorganization within the company, where Mike Krieger and Ben Mann will lead a new agent lab focused on productizing Claude, backed by over $1 billion in annual recurring revenue. The launch emphasizes agent UX converging on providing models with file system and shell access within a sandboxed environment, with a focus on iterative development and human review. AI

Nvidia has reportedly acquired AI chip startup Groq for approximately $20 billion, signaling a major investment in inference technology. New York has enacted the RAISE Act, a significant piece of legislation aimed at regulating AI safety. Zhipu AI has released GLM 4.7, an advanced open-source model specifically designed for coding tasks. AI

xAI has secured $20 billion in Series E funding, valuing the company at approximately $230 billion and exceeding its initial target. The investment round saw participation from major entities including Nvidia, Cisco Investments, and the Qatar Investment Authority. These funds are designated for enhancing AI infrastructure, such as the Colossus I and II supercomputers, training its Grok 5 model, and developing new products by utilizing real-time data from X's user base. AI

NVIDIA has introduced DGX Spark, a new cloud platform designed to accelerate AI development, particularly for agent-based systems. This platform integrates with NVIDIA's hardware and software ecosystem, aiming to simplify the creation and deployment of sophisticated AI agents. The announcement also highlighted the Reachy Mini, a compact robotic arm developed in collaboration with Pollen Robotics, which is intended to serve as a hardware platform for testing and demonstrating these AI agents in real-world applications. AI

PVH Corp. is integrating ChatGPT Enterprise across its global operations, including brands like Calvin Klein and TOMMY HILFIGER, to enhance fashion design, planning, and consumer engagement. The company aims to leverage OpenAI's frontier models to drive data-driven decision-making and innovation throughout its value chain. Separately, Deutsche Telekom is collaborating with OpenAI to bring privacy-first AI experiences to millions of European customers, starting in 2026, and will also deploy ChatGPT Enterprise internally for improved customer care and network operations. AI

General Intuition has secured $134 million in seed funding to develop world models, which are seen as the next frontier beyond large language models. The company plans to leverage a massive dataset of 3.8 billion gaming highlight clips from its sister company, Medal, to train these models. The goal is to create vision-based agents capable of spatial-temporal reasoning that can transfer from video games to real-world applications, including robotics. AI

OpenAI has announced its agreement to acquire Neptune, a company specializing in tools for tracking AI model training and experiments. This acquisition aims to enhance OpenAI's internal infrastructure for frontier research by integrating Neptune's capabilities into their training stack. The goal is to provide researchers with better visibility into model evolution, allowing for more efficient learning and decision-making during the development process. AI

Google's new Gemini 3 model demonstrates significant advancements in AI capabilities over the past three years, moving beyond simple text generation to complex tasks like interactive game creation and autonomous coding. This evolution, highlighted by Ethan Mollick, suggests AI is becoming a general-purpose tool capable of performing any computer-based task. While Gemini 3 shows promise, a comparison with ChatGPT's image generation, particularly in relation to a viral Ghibli trend, indicates potential weaknesses for Google's AI, despite ChatGPT's reported $70 million revenue from such trends. AI

OpenAI has finalized its restructuring into a capped-profit entity under a non-profit board, a move that also includes a timeline for AI researcher development extending to 2028. The company is also charting a course for its AI cloud product offerings and has outlined plans for acquiring the next trillion dollars worth of compute power. This strategic shift aims to align its research and development with long-term commercial and technological goals. AI

Google is integrating its Gemini models into a new personal health coach feature for Fitbit Premium users, offering personalized insights and adaptive guidance for sleep, fitness, and health habits. This initiative leverages Gemini's reasoning capabilities and a multi-agent framework to provide tailored advice grounded in scientific principles. Separately, a leaked Walmart 'Onn' smart speaker indicates Google is expanding its Gemini ecosystem to third-party devices, potentially reviving the market for external smart speakers. AI

Together AI has expanded its platform to include advanced multimedia generation capabilities, integrating over 40 new image and video models. This move aims to simplify development by offering a unified API for text, image, and video generation, eliminating the need for developers to manage multiple providers. The platform now hosts models like FLUX.2 for consistent character and product image generation, alongside video models from major players such as OpenAI, Google, and ByteDance. AI

OpenAI has launched ChatGPT Atlas, a new web browser designed to integrate ChatGPT directly into the browsing experience. The browser utilizes a new architecture called OWL (OpenAI's Web Layer) which separates the Chromium runtime from the main application, enabling faster startup and improved stability. Atlas aims to act as a co-pilot, assisting users with tasks, research, and analysis across the web, with features like 'agent mode' and optional 'browser memories' for context retention. The browser is initially available on macOS for various user tiers, with plans for Windows, iOS, and Android versions soon. AI

OpenAI is enhancing its safety protocols for advanced AI models by incorporating external testing and assessments. This involves collaborating with independent experts to evaluate capabilities, risks, and mitigation strategies, aiming to build trust and transparency. These third-party evaluations, including those conducted for GPT-4 and GPT-5, supplement internal testing and inform responsible deployment decisions. Separately, Wrtn, a Korean AI company, has successfully scaled its lifestyle AI applications to millions of users by leveraging OpenAI's models, including GPT-5. Wrtn's approach focuses on making AI accessible through persona-based prompts and localization, leading to significant user engagement and expansion into new markets. AI

OpenAI, in partnership with Oracle and SoftBank, is expanding its Stargate AI infrastructure platform with five new data center sites across the U.S. This expansion, which includes sites in Texas, New Mexico, and Ohio, brings the total planned capacity to nearly 7 gigawatts and over $400 billion in investment over the next three years. The company is on track to meet its $500 billion, 10-gigawatt commitment by the end of 2025, ahead of schedule. These new facilities are intended to accelerate AI research and make high-performance compute more accessible. AI

Alibaba Cloud announced a significant expansion of its AI capabilities, releasing seven new models over a four-day period. Among these were the Qwen3-Max, Qwen3-Omni, and Qwen3-VL models, indicating advancements in various AI domains. The company also unveiled a substantial $52 billion roadmap, signaling a long-term commitment to AI development and infrastructure. AI

Thinking Machines has appointed Soumith Chintala as its new Chief Technology Officer, succeeding Barret Zoph. Chintala, who has a decade of experience in AI research and development, is expected to lead the company's technological advancements. The company also announced its intention to build a vLLM team focused on advancing open-source vLLM and supporting frontier models. AI

Oracle's cloud infrastructure division announced a significant surge in revenue bookings, reaching $455 billion, largely due to a substantial contract with OpenAI. This deal positions Oracle as a key player in providing AI computing power, potentially challenging Microsoft's existing relationship with OpenAI. Meanwhile, Moonshot AI has released a new checkpoint engine for large-scale models, and research continues to advance reinforcement learning frameworks and efficient model training techniques. AI

Anthropic has secured $13 billion in Series F funding, valuing the company at $183 billion. This substantial investment marks a significant milestone for the AI safety and research company. The details of the investors and the specific allocation of the funds were not immediately disclosed. AI

OpenAI is enhancing ChatGPT's safety features to better handle users experiencing mental and emotional distress. The company is training its models to respond with empathy, offer support, and direct users to professional resources like crisis hotlines. For cases involving potential harm to others, conversations are escalated for human review and may be reported to law enforcement, while self-harm cases are handled with privacy in mind. These improvements are being developed with input from medical professionals and mental health experts. AI

Cohere has secured $500 million in funding at a $6.8 billion valuation, signaling significant investment in the open-source AI model space. Additionally, the Allen Institute for AI (AI2) received grants totaling $152 million from the National Science Foundation and NVIDIA. These financial injections highlight continued strong support for Western-based AI research and development initiatives. AI

OpenAI has released its most advanced open-weight reasoning models, aiming to democratize access to powerful AI technology globally. This initiative, part of the "OpenAI for Countries" program, seeks to foster AI development rooted in democratic values and support nations in building their own AI infrastructure. The release is intended to encourage innovation, support emerging markets, and strengthen US leadership in AI by promoting models developed within an ecosystem of transparency and democratic norms. AI

OpenAI has announced strategic partnerships with two major companies, Target and Intuit, to integrate AI-powered experiences into their platforms. Target will offer a new app within ChatGPT for personalized shopping recommendations and checkout, while also continuing to use OpenAI APIs for employee productivity. Intuit, a financial technology company, will deepen its use of OpenAI's frontier models under a new agreement exceeding $100 million, enabling AI agents for tasks like tax preparation and cash-flow forecasting within its apps and ChatGPT. AI

Genspark has launched Super Agent, a no-code AI assistant capable of automating real-world tasks such as making phone calls and generating presentations. The platform leverages OpenAI's GPT-4.1 and Realtime API, utilizing multimodal capabilities for text, image, and voice interactions. In a remarkable display of rapid adoption, Super Agent achieved $36 million in annual recurring revenue within just 45 days of its launch, driven by significant user demand for outcome-oriented automation. AI

Hugging Face has partnered with NVIDIA to launch a new Training Cluster as a Service. This offering aims to provide researchers and developers with access to powerful GPU infrastructure for training large AI models. The collaboration leverages NVIDIA's advanced hardware and Hugging Face's platform to streamline the model development process. AI

Anthropic has launched Claude Design, a new product that allows users to collaborate with Claude Opus 4.7 to create visual assets like designs, prototypes, and presentations. This tool leverages Anthropic's advanced vision model and offers features for refining designs through conversation, inline edits, and custom sliders, with the ability to integrate team design systems. Concurrently, Anthropic has made Claude Opus 4.7 generally available, highlighting its improved capabilities in software engineering and vision, while also implementing specific safeguards for cybersecurity-related tasks. AI

OpenAI has enhanced its Responses API by integrating a computer environment, enabling models to act as agents capable of executing complex workflows. This new capability allows models to interact with command-line tools, run various programming languages, and access restricted network resources within an isolated workspace. The update also introduces new built-in tools like image generation and improved file search, alongside features for background processing and encrypted data handling, aiming to boost reliability and developer efficiency. AI

Together AI has launched new GPU clusters featuring NVIDIA's Blackwell platform, offering significant speedups for AI training and inference. These clusters, powered by the Together Kernel Collection, achieve up to 90% faster training speeds compared to previous NVIDIA H100 hardware, processing over 15,000 tokens per second for large models. Early access customers like Salesforce and Zoom have reported substantial performance gains, with some experiencing double the training speed. Together AI's optimization efforts span custom kernels, inference engines, and speculative decoding, aiming to redefine efficiency in AI model development and deployment. AI

OpenAI has reportedly secured a $300 billion valuation, marking a significant milestone in the AI funding landscape. This valuation comes as part of a broader trend of substantial investment in AI companies, with Cursor also raising $9.5 billion and Etched securing $1.5 billion. The total funding across these entities today exceeds $41 billion, underscoring the intense investor interest in artificial intelligence. AI

OpenAI has announced leadership updates to support its growth and mission of developing AGI for humanity's benefit. Mark Chen is now Chief Research Officer, focusing on advancing AI capabilities and safety while integrating research with product development. Brad Lightcap, as COO, will expand his oversight to include business strategy, partnerships, and infrastructure. Julia Villagra has been appointed Chief People Officer to manage global scaling and maintain OpenAI as a top destination for AGI talent. AI

Anthropic has secured $61.5 billion in Series E funding, marking a significant financial milestone for the AI company. This substantial investment is expected to fuel further research and development in advanced AI systems. The funding round saw participation from major players, underscoring strong investor confidence in Anthropic's future growth and technological contributions. AI

Google has released "Deep Research," an AI product that functions as an agent, utilizing custom-tuned frontier models like o3 and Gemini 1.5 Flash. This tool is designed to perform complex research tasks rapidly, with users comparing its output quality to that of a PhD-level research assistant. Early adopters have praised its efficiency, noting that it can complete multi-day research projects in minutes, and have speculated about its potential to redefine internet navigation and knowledge work. AI

Chai Research, a startup founded by former hedge fund traders, has achieved over 1.4 million daily active users and $22 million in revenue with its consumer AI chat application. The company has developed a platform called Chaiverse that allows for rapid A/B testing and deployment of over 100 large language models per week. This system significantly shortens evaluation cycles from weeks to hours by leveraging user feedback from its large user base to rank and iterate on models. AI

Project Stargate, a proposed $500 billion datacenter, represents a significant investment in AI infrastructure, potentially consuming 1.7% of the US GDP. This initiative is linked to the development of advanced AI models, including Gemini 2 Flash Thinking 2. The project underscores the massive scale of resources being allocated to support the growing demands of artificial intelligence. AI

Anthropic has released an upgraded version of its Claude 3.5 Sonnet model, which reportedly matches the capabilities of its Opus 4.6 counterpart in some benchmarks and offers a 1 million token context window. Independent evaluations suggest the new Sonnet model performs comparably to human baseliners on certain tasks, though its token usage can be significantly higher than previous versions. Meanwhile, the AI coding assistant Cursor is reportedly valued at $28 billion, with OpenAI acquiring Windsurf for $3 billion, indicating significant investment and consolidation in the AI tooling space. AI

Hugging Face has partnered with AMD to introduce the 5th Gen EPYC™ CPU, codenamed Turin. This collaboration aims to enhance AI and machine learning workloads by leveraging AMD's latest processor technology. The new CPUs are expected to offer significant performance improvements for training and inference tasks, making them a compelling option for AI infrastructure. AI

OpenAI has launched GPT-5.4-Cyber, a specialized model for cybersecurity defense, alongside its "Trusted Access for Cyber" program. This initiative aims to provide verified defenders with advanced AI tools to accelerate vulnerability discovery and remediation, while implementing safeguards against misuse. The program includes $10 million in API credits for cybersecurity grants and partnerships with leading security firms and institutions like the UK AI Security Institute. AI

Researchers have developed a benchmark to test Large Language Models' ability to handle temporal changes in legal statutes, identifying issues like outdated information and recency bias. Meanwhile, the AI industry is seeing a significant shift as model labs increasingly focus on building agent-based products rather than just foundational models. This strategic pivot is exemplified by companies like AI21 and DeepSeek, and is further underscored by DeepSeek's aggressive pricing strategy for its V4-Pro model, making advanced AI more accessible. AI

Silurian, a startup founded by former Microsoft researchers, has launched Generative Forecasting Transformer (GFT), a 1.5 billion parameter model designed to simulate Earth's weather up to 14 days in advance. This deep learning model, which learns purely from data without explicit physics, has demonstrated strong performance in predicting hurricane tracks, outperforming traditional forecasting methods. The company aims to expand its simulations to model other weather-impacted infrastructure like energy grids and agriculture. AI

OpenAI has established an independent Safety and Security Committee, chaired by Zico Kolter, to oversee critical safety and security measures for its model development and deployment. This committee will review safety evaluations for major model releases, with the authority to delay launches if safety concerns arise. The committee's recommendations, focusing on governance, security enhancements, transparency, external collaboration, and unified safety frameworks, are being adopted by OpenAI. Additionally, OpenAI is exploring the creation of an Information Sharing and Analysis Center (ISAC) for the AI industry to improve collective cybersecurity resilience. AI

Apple has unveiled "Apple Intelligence," a new personal intelligence system integrating generative models into its devices like the iPhone, iPad, and Mac. This announcement was a key topic at WWDC 2024, highlighting Apple's push into AI capabilities. The discussion also touched upon "Advanced RAG," a technique for precise zero-shot dense retrieval, and explored hybrid search methods combining BM25 and semantic search for improved results, particularly with tools like Langchain. AI

Brightwave, an AI research assistant for investment professionals, has secured $6 million in seed funding led by Decibel. The company, which serves clients managing over $120 billion in assets, is focusing on practical LLM applications rather than solely on long context windows. Brightwave employs a strategy of breaking down complex tasks into subtasks and then synthesizing the results, a method they've found effective for generating detailed summaries and reports. AI

OpenAI has established a new Safety and Security Committee composed of board members and key technical experts. This committee will evaluate and recommend critical safety and security decisions for all OpenAI projects and operations, with initial recommendations due within 90 days. The formation of this committee coincides with OpenAI's ongoing training of its next frontier model, emphasizing the company's commitment to balancing advanced capabilities with robust safety measures. AI

Meta has released Llama 3, an open-source large language model, in an effort to democratize AI development. The models, available in 8B and 70B parameter sizes, are designed to be more capable and efficient than their predecessors. Meta aims to foster innovation by providing broad access to powerful AI tools, contrasting with the more closed approaches of some competitors. AI