Track model releases, AI agents, open-source projects, infrastructure, policy, and product updates as they publish.
The U.S. government has lifted the export ban on Anthropic's Claude Fable 5 and Mythos 5 models. These models, previously restricted from international sale, are now set to return to the global market. startup daily reports.
Google has opened up Gemini Omni Flash to developers via the Gemini API and Google AI Studio, combining multimodal reasoning with video generation and editing. It also released Nano Banana 2 Lite, an image model that generates 1K resolution images in about 4 seconds at half the cost of its predecessor.
Hugging Face and Cerebras announced a collaboration to power real-time voice AI using Google's Gemma 4 model on Cerebras hardware. The partnership combines Cerebras' wafer-scale chips with Hugging Face's open-source ecosystem for low-latency speech inference.
Chinese AI companies are turning World Cup match predictions into a new competitive arena, according to Bloomberg. Multiple firms are leveraging large language models and data analytics to forecast tournament outcomes, making sports forecasting the latest front in their technology showcase.
HackerNoon reports the release of Higgs-tts-2-3b-base, a text-to-speech foundation model. The model provides a pretrained TTS base for developers building speech synthesis applications.
The free, open-source agentic programming framework OpenClaw is now available on Android and iOS, bringing autonomous agent capabilities to mobile devices for the first time. Users can run and interact with AI agents directly from their smartphones without needing a desktop environment.
Anthropic announced Claude Science, a major new flagship product designed to support scientific research in the same way Claude Code supports software engineering. The agentic system can autonomously carry out meaningful work from concise, high-level instructions, targeting pharmaceutical executives and biotech researchers.
EquiLibre Technologies, a Prague-based AI lab founded by three former DeepMind researchers, is now valued at more than $500 million. The team, known for building a championship poker AI, has pivoted its reinforcement learning expertise to generate profits for quantitative hedge funds.
Anthropic has launched an AI drug discovery program, joining Google, Nvidia, and other tech giants investing heavily in healthcare.
OpenClaw has released its AI application, now available on both iOS and Android platforms, expanding its AI services to mobile users.
AWS announces a $1 billion forward deployed engineering organization to embed AI teams directly into enterprises, accelerating AI adoption.
Google updates its image generator to be faster and cheaper, providing a more useful tool for creators making AI content.
Anthropic releases Claude Sonnet 5, emphasizing cost performance and enhanced AI agent capabilities for enterprise automation.
IBM Research introduces ScarfBench, a benchmark designed to evaluate AI agents on enterprise Java framework migration tasks.
AI chip startup Etched reaches $5 billion valuation, with $1 billion under contract for its AI inference systems.
Alibaba's Qwen team released the Qwen 3.6-27B open-source model, offering developers a powerful new option for local AI deployment.
Startup Acti is betting the smartphone keyboard is the next home for AI assistants. Its new keyboard for iOS and Android works across apps and lets users create custom AI-powered shortcuts using natural language.
Cryptocurrency exchange OKX has launched a dedicated AI Marketplace for autonomous agents, marking a significant step in integrating AI with crypto ecosystems.
Huawei has officially announced the world's first commercial multimodal cultural tourism large model achieving scaled deployment, bringing AI to the cultural tourism industry.
Infrastructure company Firmus announces plans to build a 360MW Nvidia AI data center in Indonesia, further expanding AI computing infrastructure in Southeast Asia.
X launches a hosted MCP server, enabling easier integration of AI applications with its API.
AI company Anthropic is posting $85,000 AI positions with no college degree or work experience requirements, signaling a shift in AI talent acquisition.
NVIDIA releases an inference software stack designed to minimize cost per token for AI factories.
Riverside, the popular podcast recording platform, is launching AI-powered newsletter creation that lets users turn their recordings into email newsletters automatically.
Amazon has launched a new $1 billion Frontier Deployment Engineering (FDE) organization, embedding engineers inside enterprise customers to deploy custom AI agents.
Google has cut off Meta's access to Gemini AI services due to a compute capacity crunch, as reported by MoneyCheck.
According to reports, OpenAI has cut its AI model inference costs in half through deep system-level optimizations.
Proton's Lumo chatbot is getting a major upgrade to version 2.0 this week, bringing a broader range of capabilities to its privacy-first AI assistant.
IT services company Hexaware has launched Tensai for Reasoning Ops, a new product that brings Agentic AI capabilities to IT operations management.
OpenClaw has released AI Agent applications on both iOS and Android platforms, bringing AI agent capabilities to mobile users.
NVIDIA introduces three workflows via its Omniverse and Metropolis platforms, leveraging synthetic data generation and model fine-tuning to improve vision AI agent accuracy in factories and other physical-world settings.
Blockchain project ZetaChain is shutting down its cross-chain services to fully pursue AI memory infrastructure development.
A ZDNET survey reveals 61% of US adults now use AI for health information, skyrocketing from just 2% in 2024.
Cursor, the popular AI-powered coding assistant, has released a mobile application that lets users control AI coding agents directly from their smartphones.
LQR House has entered a two-year agreement with ByteDance to deploy AI computing power supporting quantitative research at Fusion Five Continents Securities.
Anthropic has publicly stated that its upcoming AI model is not too dangerous to release after rigorous safety evaluations. The company confirmed that the model meets internal safety standards across multiple benchmarks, paving the way for its planned launch.
MIT Technology Review reports that while AI offers transformative potential for agriculture—especially amid volatile fertilizer costs and unpredictable weather—the industry lacks the data infrastructure needed to deploy it effectively. Research shows AI-powered predictive models can improve crop yields, but scaling from pilot to production remains constrained by poor data quality and availability.
Ex-employees from Meta and OpenAI are raising $400 million to establish a new AI laboratory.
Blacklisted Chinese AI firm Z.ai (Zhipu AI) sees its open-weight GLM-5.2 model top multiple AI ranking charts, powered by Huawei chips.
Meta blocks engineers from using Claude Code and Codex due to concerns about model distillation.
Daily Archive