All Stories — Page 40
Every story published on RuntimeWire, newest first.
- Black Forest Labs opens FLUX 3 Video API with native audio and draft rendering
The model makes 20-second HD clips from text, images or video, with Full HD upscaling, multilingual dialogue and synchronized sound.
- Pokee AI launches 28B agent model, claims 10 million tokens on one RTX 4090
Pokee has not released the model's weights, detailed test configurations or enough architecture information to independently assess its long-context and throughput claims.
- Lamb Labs unveils AI chip plan targeting 20,000 tokens a second
Lamb Labs says an 8B FPGA prototype runs below 10 watts, while its 63x efficiency claim lacks a reproducible benchmark.
- Omanta Health maps patient records and genomics to potential therapies
The YC S26 entrant says its system combines longitudinal records, personal genomics and scientific literature to map disease drivers and therapies.
- Warp's Agent CLI brings its coding agent to any terminal
Zach Lloyd is separating Warp's agent from Warp Terminal while keeping the session infrastructure that powers interactive and remote work.
- Microsoft spotlights Paige's year-old PRISM2 model trained on 2.3 million pathology slides
The model pairs digitized slides with clinical reports, but its benchmark results remain research evidence rather than clinical validation.
- vlt ships 1.0 with hosted registries that work across npm tooling
Darcy Clarke is turning a package manager built by npm veterans into a commercial registry business aimed at faster CI and safer installs.
- Anthropic appoints Tino Cuellar as chief global affairs officer
Anthropic says the former California Supreme Court justice will lead policy, strategic international engagement and government relationships worldwide.
- Cursor open-sources MoK after a claimed 41% training throughput gain
The Apache-licensed kernel fuses MoE communication and computation for training models on Nvidia GB300 NVL72 racks.
- Firecrawl ships open-source anydoc for faster AI document ingestion
The Rust engine handles PDFs, Word files, slide decks and 10 other formats, and now powers Firecrawl's existing /parse endpoint.
- Google Workspace guest accounts put outside collaborators under admin control
The feature, generally available since March 30th, gives clients and contractors limited access to Chat, Drive files and Meet calls.
- NVIDIA releases Alpamayo 2 Super weights for autonomous-driving developers
The 34-billion-parameter model pairs a 32B vision-language backbone with a diffusion action expert for robotaxis, trucks and delivery fleets.
- OpenAI hires reverse-engineering veteran Halvar Flake for cyber and efficiency work
The incoming OpenAI researcher previously founded security startup zynamics and profiling-software maker Optimyze, selling both companies.
- Blue Water Autonomy wins seat on Navy's $40 million ocean-mapping contract
Blue Water Autonomy secured one of six seats on a two-year contract with a shared $40 million ceiling; task orders will be competed among the awardees.
- Genome Computer offers private AI chat for exploring personal genome data
David Ball's $15-a-month service sends models only question-relevant variants while storing the full genome separately.
- Mistral ships Shieldstral, a 3B guard model with runtime policy controls
The Apache 2.0 release evaluates text and images against plain-language policies, giving developers a compact moderation layer they can operate themselves.
- Cloudflare opens Wallet handle reservations for AI agent payments
Users can reserve identities now, while wallet payments, Virtual Wallets and funding tools have no established launch date.
- Straiker launches an AI-agent kill switch after raising a $64M Series A
Ankur Shah is applying his cloud-security playbook to software that can write code, call tools and act across enterprise systems.
- Endeavor Optical Networks raises $10.75M to route data-center traffic through satellites
Endeavor Optical Networks plans a 20-satellite network for dedicated, multi-terabit links between data centers.
- Horizon3 raises $250 million to expand its autonomous cybersecurity platform
NightDragon and NEA co-led the Series E, which values the NodeZero maker at more than $2 billion.
- Celeris-1 tops AI speed ranking at 2,158 tokens per second
Tom Hamer and Jesse Clark's diffusion model pairs a top throughput score with a 12 on Artificial Analysis' Intelligence Index.
- ChatTJB advertises a human-powered LLM parody on an SF billboard
Poet and former Google marketer Tucker Bryant manually writes every reply and draws users' image requests by hand.
- OpenAI releases messages challenging Apple's hardware trade-secret claims
The records challenge parts of Apple's case while raising new questions about files moved during an engineer's departure.
- MVNT seeks $10 million to build AI tools around choreographers' motion data
Cofounders Joon Jung and Choi Youngjun are pitching motion data, creator attribution and tools for games, animation and casual users.