thinking machines lab releases inkling open-weights model
mira murati's thinking machines lab released inkling, a 975b parameter mixture-of-experts multimodal model under apache 2.0 license, aimed at fine-tuning rather than frontier performance.
aiplain 200-word summaries of important ai and data science news, updated every few hours.
mira murati's thinking machines lab released inkling, a 975b parameter mixture-of-experts multimodal model under apache 2.0 license, aimed at fine-tuning rather than frontier performance.
ainvidia releases open embedding models that rank first on rteb and improve agentic retrieval efficiency.
aigoogle vids now lets you generate and edit videos with text prompts and create a digital avatar from a selfie and voice recording.
aigoogle's ai mode now connects with instacart, canva, and youtube to complete tasks like shopping and design directly in search.
aipi is a coding agent that strips away built-in features like sub-agents and permission popups, relying on a tiny system prompt and user-built extensions instead.
aialexandre lebrun of ami labs rejects agi and superintelligence terms as poorly defined, focusing instead on world models for physical ai.
aixai released the full grok build codebase under apache 2.0 following user reports that the cli tool uploaded entire directories to google cloud.
aia curated list of youtube channels covering ai research breakdowns, coding tutorials, and industry analysis for data professionals.
aidharmaocr outperforms mistral ocr4 and unlimited-ocr on brazilian portuguese by concentrating training on one language.
aihugging face disclosed an intrusion driven by an autonomous ai agent that exploited dataset processing to access internal systems, detected and analyzed using ai tools.
aigoogle deepmind and isomorphic labs outline a joint program using ai to prevent misuse, detect outbreaks, and speed up responses to biological threats.
ailondon startup applied computing raised $20m to expand its foundation model orbital, which combines time series, physics, and language models to predict facility states and speed up investigations.
ainew bias correction method improves convergence rate of non-expansive stochastic approximation from t^{-1/4} to t^{-1/3}.
aia black-box method tests whether llm reasoning steps truly depend on stated premises by swapping predicates and checking if conclusions change.
aia new framework measures the symmetry information discarded by machine learning models under lie group actions, with implications for privacy and fingerprinting.
aia step-by-step trace of pytorch's automatic differentiation for pinn training, showing how it handles two levels of differentiation with explicit numerical values.
aia systematic review of federated explainable ai, covering roles, architectures, evaluation, and open challenges.
aia new method combines tumor growth and dropout predictions using empirical bayes variational autoencoders with genetic covariates.
ai