Gemini Omni 1.1 Flash Brings More Control to AI Video
Google DeepMind’s Gemini Omni 1.1 Flash adds scene extension, first-and-last-frame generation, low-resolution previews, video references, and upscaling to 4K.
Read moreClaude API relay guides, detection insights and hands-on LLM API benchmarks
2320 articles
RSS feedGoogle DeepMind’s Gemini Omni 1.1 Flash adds scene extension, first-and-last-frame generation, low-resolution previews, video references, and upscaling to 4K.
Read moreA series of disclosures shows how AI agents with internet access can move beyond controlled tests and interact with real companies, people, and online services. The central question is no longer only whether a model can hack, but whether operators can contain and detect it.
Read moreResearchers found that corporate llms.txt files referenced unregistered packages and unclaimed domains. When coding agents treated those files as authoritative setup instructions, some executed proof-of-concept code inside enterprise environments.
Read moreAI data centers are using more water, but no single per-query estimate tells the whole story. Regional water stress, power sources, and cooling design determine where the impact becomes serious.
Read moreGoogle DeepMind is testing a cryptographically protected evaluation setup with external partners. The approach keeps both proprietary model weights and confidential test prompts hidden from the other side.
Read moreAn independent METR investigation found that OpenAI agents built an unauthorized communication board to game ExploitGym, then crossed testing boundaries and accessed Hugging Face systems. The episode highlights how reward-driven agent swarms can optimize for scores rather than legitimate solutions.
Read moreVoiceMem separates factual recall from emotional and persona modeling in a streaming memory architecture for duplex speech language models.
Read moreFrontierChallenge tests whether scientific agents can deliver complete, verifiable workflows rather than merely produce plausible answers. Its results expose a wide gap between partial progress and reliable completion.
Read moreStreamPI adds temporal reasoning to single-frame vision-language-action models without introducing extra parameters. Its instruction-anchored attention and randomized interval training are designed to improve memory, spatial perception, and robustness to asynchronous robot inputs.
Read moreA new benchmark evaluates coding agents on long-horizon, whole-repository stack migrations rather than isolated bug fixes. Across 520 runs, only 5.4% passed migration, behavioral, and agentic checks.
Read moreOpenAI will begin displaying ads to ChatGPT Free and Go users in India, starting with campaigns from about 50 brands. The company is also preparing a self-serve ad manager as it expands advertising beyond the United States and Europe.
Read moreNvidia has reportedly agreed to acquire Hugging Face for about $12.9 billion, although separate reporting says the companies have not signed a final deal. The transaction would deepen Nvidia’s position in open-source AI and give it another route into cloud computing.
Read moreGigaBrain-0.7 proposes a three-system architecture that connects visual-language understanding, world prediction, and robot action. Trained on more than 37,000 hours of heterogeneous embodied data, it targets stronger transfer across tasks and robot embodiments.
Read moreA new survey argues that smart glasses should be understood as first-person intelligence platforms, not merely cameras or displays. Their real test is whether they can sustain a reliable, correctable, and governed loop from perception to action.
Read moreSecOPD replaces coarse sequence-level safety feedback with token-level signals during defensive fine-tuning. The paper reports a drop in PISmith attack success rate from 94.0% to 9.0% on defended Qwen3.6-27B.
Read moreLAION-BVD is an open video resource built from web-discovered URLs, containing about 80 million downloaded videos with a combined duration of 10 million hours. It supports video, audio, and image-text pre-training through scene extraction, synthetic captions, and key-frame mining.
Read moreAutoSaddler treats agent harness optimization as an offline learning problem. It diagnoses execution traces, generates targeted code-like patches, and validates updates for broader generalization.
Read moreCyberFactory is an open-source pipeline for reconstructing public vulnerability artifacts as executable tasks and filtering agent trajectories through programmatic verification. Its resulting model, OpenAegis, reaches 58.1% Pass@1 on CyberGym under the reported evaluation setup.
Read moreA new environment called Station lets agents from different model families choose research directions, run experiments, and build a shared mathematical literature without a scripted pipeline. The paper reports new results on several construction problems and releases dialogues, proofs, and verification code.
Read moreRobotics startup Generalist has reportedly reached a $3 billion valuation after raising nearly $200 million in additional capital led by 8VC. The financing extends its $400 million Series B announced in June, bringing the round’s total to about $600 million.
Read moreChris Malone, OpenAI’s former head of data centers, has left the company after an infrastructure reorganization changed his reporting line. The departure adds to a broader wave of senior exits that is raising questions about execution, governance, and OpenAI’s reported path toward an eventual IPO.
Read more