Edition: August 11, 2026

Accel has closed a $550 million India fund, exceeding its target and marking a significant investment in the country's startup ecosystem, with a focus on AI, consumer internet, fintech, and advanced manufacturing. The fund was oversubscribed and closed within weeks, despite Accel still having ample capital remaining in its earlier vehicle. Accel plans to begin deploying capital from the new fund in 2027, betting on India's next startup wave driven by AI and other technologies.

ChatGPT and Google's Gemini have both surpassed 1 billion users, marking a significant milestone for AI technology. Gemini is Google's fastest-growing product ever, while ChatGPT achieved the milestone a few weeks prior. This achievement highlights the rapid growth and adoption of AI-powered chat services.

Google's Gemini app has reached 1 billion monthly active users, making it one of the company's fastest-growing products and keeping pace with OpenAI's ChatGPT. The app's growth is attributed to its integration across Google's products and the introduction of new features such as Gemini 3.5 Flash. Gemini's user base is diverse, with over 100 million active users on iOS and 63% of users utilizing the voice feature.

The concept of a 'censorship-industrial complex' has emerged, suggesting a network of government agencies, academics, and tech platforms are suppressing conservative speech online. This idea has gained traction and influenced US policy, particularly under the second Trump administration. The narrative has significant implications for the global online community.

River AI, a 2-month-old startup, has secured $1.1 billion in funding to develop personally trainable AI assistants, aiming to revolutionize the way AI models are trained and used. The company, founded by xAI co-founder Igor Babuschkin, plans to rebuild the AI stack from scratch to enable capable agents that can be trained by individuals. With this significant investment, River AI is poised to make a significant impact in the AI industry.

Google Research's medical AI system, AMIE, has demonstrated real-time clinical video consultation capabilities in a first-of-its-kind study, showing expert-level capabilities in interpreting visual and auditory cues and guiding virtual physical exams. The system was assessed favorably by clinical evaluators and preferred by patient actors over text chat. AMIE is still in the research phase but offers a glimpse into the future of health AI.

An unreleased Anthropic AI model has made significant progress on the Riemann hypothesis, a 150-year-old unsolved math problem, by testing 650 different ideas and coordinating across 60 subagents. The model's findings were confirmed by in-house mathematicians and formalized using the Lean proof assistant. This breakthrough is the latest in a string of mathematical advancements led by large language models.

Apple is developing a feature to verify the authenticity of iPhone photos, allowing users to prove their origin and distinguish them from deepfakes. The 'Apple Reference Image' system embeds metadata into photos at capture. The feature is currently in beta and will be off by default.

Researchers have uncovered a major security vulnerability in Zoom that could allow attackers to hijack devices during meetings, using fewer than 20 AI prompts to exploit the flaw in Zoom's annotation feature. The vulnerability has been patched by Zoom. The exploit could have allowed attackers to steal data, access cameras and microphones, or install malware.

Researchers from IBM introduce ALTK-Evolve, a system that enables an agent to learn from its own trajectories, similar to ACE, but with a more efficient approach to delivering lessons to the model, resulting in fewer tokens required. Both systems agree on the importance of not compressing an agent's lessons, but differ in how the memory is built and delivered. This difference in delivery leads to a significant reduction in the token bill.

Spotify will introduce 'AI Persona' labels to identify AI-generated artists and exclude their music from recommendations, aiming to promote authentic artists and improve user experience. The labels will appear on artist profiles, search results, and playlists, and users can choose to follow AI Personas to include their music in recommendations. This move is part of Spotify's efforts to balance AI-driven innovations with the need to maintain music quality.

Spotify will start labeling AI-generated artists with an 'AI Persona' badge and remove their music from user recommendations, aiming to distinguish between human and AI-created content. The change will roll out in mid-September, using a combination of human review and AI tools to identify AI personas. This move reflects the growing need to address AI-generated content in the music industry.

A security vulnerability in Zoom's screen-sharing feature allowed anyone on a call to take over other devices without interaction or indication, but Zoom has since issued fixes to address the flaw. The bug was discovered using AI models, highlighting the increasing accessibility of advanced hacking capabilities. The vulnerability affected devices running all operating systems supported by Zoom.

Anthropic's AI model Claude will begin applying invisible watermarks to generated text and images, making it easier to detect AI-created content. This move aims to comply with European rules for AI transparency. The changes will be implemented in the future.

Anthropic will watermark text generated by its AI models, including Claude, to comply with European regulations. The watermark will be applied at the model level and will travel with the text when copied and pasted elsewhere. This move is part of a larger trend of AI companies watermarking generated content to avoid regulatory scrutiny.

Startups are working on new ideas to replace transformers in large language models, which are becoming a bottleneck as LLMs get bigger and better. Meanwhile, AI professors are navigating the changing landscape of academic research in the field. These developments could lead to faster, more efficient, and potentially smarter LLMs.

OpenAI has made a significant breakthrough in mathematics by solving 10 long-standing problems using AI, leaving mathematicians like James Maynard to ponder the future of their field. This development marks a significant shift in the traditionally slow-moving discipline. The AI takeover of mathematics has begun, with potential implications for various fields.

Computer scientists have discovered a method to extract the hidden 'thinking' of advanced AI models, revealing potential vulnerabilities and similarities between US and Chinese models. The findings suggest that certain Chinese models may have been trained using 'distilled' reasoning information from US models. This vulnerability could lead to personal information leakage and large-scale reasoning distillation attacks.

Scientists are creating human brain organoids, lab-grown brain tissue with millions of neurons, to study diseases and develop new biocomputing systems. These organoids can produce brain waves, respond to electrical signals, and even play video games. The research aims to create a new form of intelligence built from living neurons, potentially revolutionizing the field of artificial intelligence.

Novo Nordisk and AWS are partnering to bring AI into drug discovery, aiming to shorten the path from identifying a drug target to a first human dose. The partnership combines Novo Nordisk's scientific expertise with AWS's cloud infrastructure, AI services, and life sciences technologies. This collaboration is expected to transform how medicines are discovered, developed, and delivered.

A new study using AI-based computational models has identified 766 genes associated with schizophrenia, providing a more detailed picture of the complex genetic architecture of the disease. The findings suggest that the genes involved in the disease function as an interconnected network, contributing to the risk of developing schizophrenia. This breakthrough could lead to more precise investigations into the behavior of the disease and its potential treatments.

Artificial intelligence is revolutionizing the vulnerability response timeline by helping security researchers identify flaws that conventional tools may overlook, with a recent case showing AI-assisted development of a zero-day exploit. The use of AI in vulnerability response is becoming increasingly significant, with Google's data suggesting a record number of zero-days exploited in the wild in 2025. Complex container environments can make it harder to trace exposure, emphasizing the need for clearer software records and reduced package dependencies.

OpenAI has completed a $7 billion employee tender offer, providing liquidity to its workforce and valuing the company at $852 billion. The move suggests that an initial public offering (IPO) may not be imminent, despite the company's recent filing with the Securities and Exchange Commission. OpenAI's decision may be influenced by its desire to showcase strong financial results before going public.

Brad Lightcap, OpenAI's longtime COO, is leaving the company to start a new venture, marking another high-profile departure as the AI lab prepares for a significant IPO. Lightcap had been with OpenAI since 2018, serving in various roles including CFO and COO. His departure follows other recent executive exits from the company.

OpenAI has launched a dedicated ChatGPT desktop app for Linux, expanding its reach to the open-source community. The app, currently in preview, supports various Linux distributions including Ubuntu, Debian, and Fedora. This move follows Anthropic's release of a similar app for Linux last month.

Brad Lightcap, OpenAI's special projects lead and former COO, has announced his departure from the company after an eight-year stint, citing a desire to work on new projects that align with the company's mission. Lightcap will be starting something new, although details are scarce. His departure marks another high-profile exit from the AI lab.

Google is set to unveil its new Pixel devices on August 12th, including the Pixel 11 lineup, a next-gen Pixel Fold, and a Pixel Watch 5. The event, hosted by comedian Trevor Noah, will feature notable guests and will be live blogged. Preorders for the new devices are expected to begin on August 12th at 10AM ET.

Amazon order confirmation emails have become less informative, listing only item categories instead of specific items, leaving customers to visit the website for details. The change has been noticed by shoppers who have shared examples of the redacted emails online. The reason behind the change is not explicitly stated.