stability ai releases audio model for six-minute songs
stability ai launches stability audio 3.0, a family of models that can generate professional music up to six minutes long, with open weights for smaller versions.
topic
ai
hellora adapts moe models by targeting hot experts
hellora fine-tunes mixture-of-experts models by adding low-rank adapters only to the most active experts, cutting parameters and compute while boosting performance.
nanoclaw creator rejects $20m buyout, raises $12m seed
nanoco, the company behind the secure openclaw alternative nanoclaw, raised $12 million in seed funding after turning down a $20 million acquisition offer.
dual-channel tensor neural networks with finite-sample guarantees
a new neural network processes tensor data by splitting it into low-rank and sparse parts, with proven error bounds and automatic structure selection.
ocean raises $28m to fight ai-powered phishing
ocean, an agentic email security startup, emerged from stealth with $28 million to detect ai-generated phishing attacks using a custom small language model.
conformal prediction via transported beta laws
a new framework uses beta distributions to measure calibration-conditional coverage in conformal prediction, providing bounds on coverage gaps and bad-calibration probabilities under non-i.i.d. settings.
ai speeds up hunt for anti-aging genes
google deepmind's co-scientist helps biologists find genetic factors that reverse cellular aging, cutting analysis time from months to days.
llms with personal health records improve patient answers
study tests gemini 3.0 flash on 2,257 patient queries using personal health records, finding better accuracy with full clinical notes.
google launches ai design app pics at io 2026
google unveils pics, an ai design app for workspace that generates and edits images from text prompts, competing with canva and claude design.
calibrated uncertainty cuts llm cascade costs by 31%
a calibration-first router uses isotonic regression to map token uncertainty to error probability, enabling cost-optimal escalation thresholds without per-workload tuning.
talk to your gmail inbox with gemini ai
google adds conversational ai to gmail, letting users ask questions aloud to find buried information like flight details or appointment times.
google genie adds street view grounding
google deepmind's project genie now uses street view imagery to create interactive virtual worlds based on real locations, available to google ai ultra subscribers.
musk trial reveals his own openai conflicts
court testimony showed elon musk used openai researchers for tesla without reimbursement, undercutting his lawsuit claims.
google search gets ai agents that work 24/7
google introduces customizable ai agents in search that continuously monitor topics and send push notifications, moving beyond one-time queries.
olmoearth v1.1 cuts compute costs up to 3x
a new family of remote sensing models reduces token sequence length to lower inference costs while matching previous performance.
google workspace adds voice ai, image tool, and agent
google announces conversational voice features in gmail, docs, and keep, a new image creation app called google pics, updates to ai inbox, and a personal ai agent named gemini spark.
gemini 3.5 flash costs more, ships everywhere
google released gemini 3.5 flash with a notable price increase and plans to use it across many consumer and developer products.
google launches gemini omni for video creation
gemini omni flash can generate and edit videos from text, images, audio, and video inputs, starting with the gemini app and youtube shorts.