THE COLLECTION
EVERY SIGNAL, SORTED.
Search the whole shelf by topic, tool, term, or plain old curiosity.
China, OpenMOSS, MOSS-Live, Full-Duplex Voice AI, Realtime Assistants, Multimodal AI, Background Tools, Turn Taking, Task State, Human Control, Privacy and Evaluation
MOSS-Live kept talking while tools worked. Publish the two-timeline handoff
OpenMOSS split a live assistant into a foreground conversation model and a background task model that share memory, context and task state. That can make voice agents feel less like phone trees, but it also creates two timelines that can quietly disagree. The handoff between them needs a visible receipt.
United States, South Korea, Materials Science, Crystalline Materials, Multimodal AI, Synthetic Data, Scientific AI, Open Source, Laboratory Validation and Research Reproducibility
MIT taught a model to read crystal structures. The lab still has to make the material
CLaC connects natural-language requests to three-dimensional crystal graphs using property-grounded synthetic descriptions. That can make a huge materials database easier to search, but retrieval is still a shortlist, not a synthesized sample, a working battery or a safety certificate.
United States, Open Models, Multimodal AI, Long Context, Coding Agents, Memory, Model Governance and Developer Tools
Meta opened Muse Spark. The one-million-token model needs a memory map
Muse Spark Open can accept text, images, audio and video across a context window of up to one million tokens. That is enough room for an agent to carry a small world into one run. Developers still need a map showing what entered, what mattered, what disappeared and what the model merely inferred.
United States, Open Models, Vision Language Models, Acquisitions, Cloud Migration, Open Source Stewardship and Developer Continuity
Prime Intellect bought Moondream. Open models need a continuity receipt
Prime Intellect acquired the small-model specialist, hired its founders and started folding Photon into a larger training and inference stack. Moondream says its models will stay open. Its cloud is also shutting down in thirty days. That gap is where a continuity receipt belongs.
Decision Models, Multimodal AI, Audio, Video, Calibration, Edge Computing and Human Review
Cloudflare taught a decision model to hear and watch. Calibration still has to travel
Clef-omni turns text, images, audio and video into bounded probability scores without generating prose. That can simplify real-time routing. It also moves every microphone, camera, codec and operating environment into the reliability test.
Open Models, Vision and Creation
SenseNova published the recipe for one model that reads and draws
SenseNova-U1.5 uses one native visual system for understanding, generation, editing, and illustrated answers. The new report explains how the parts fit, and which parts are still missing.