Open Source - Latest AI News & Updates
Open-source AI projects, community innovations, collaborative development, and freely accessible AI tools and frameworks.
325 articles
View complete article list
Anthropic shuts down Fable 5 and Mythos 5 models amid White House dispute
Anthropic was already juggling a Pentagon standoff when, on June 12, a White House directive forced the company to block foreign access to its newest models.
Tutorial Shows Homogeneous and Heterogeneous GNN Training with city2graph
Here’s the thing: building a spatial graph from scratch used to be a handful of disjoint scripts.
Mistral AI seeks EUR 3 bn, valued at EUR 11.7 bn; ASML holds 11% stake
Mistral AI is on the brink of a massive fundraising push. The French startup is reportedly in early talks for a new round that could bring in roughly 3 billion euros, a move that would lift its valuation to about 20 billion euros.
Dario Amodei has one direct report; sister Daniela runs Anthropic's exec team
Dario Amodei runs one of the fastest‑growing AI firms, Anthropic, now valued at roughly a trillion dollars—just five years after its launch.
GPU utilization masks storage and I/O bottlenecks slowing modern AI
79 % GPU utilization. 82 % the next hour. 84 % after autoscaling. The cloud bill climbs, yet latency barely shifts.
Matmul Enables Dropless MoE Training; Grouped‑GEMM Kernel Drives Speed
Mixture‑of‑Experts layers let transformer models grow without a linear rise in compute, but the usual JAX/MaxText workflow still drops tokens that exceed an expert’s capacity.
LangChain Emergency Helpline Uses AssemblyAI WebSocket for Live STT
We’ve all faced moments when every second counts and a phone call is the only lifeline. In those cases, pressing a keypad to reach the right operator feels like an unnecessary hurdle.
Meta launches Hatch AI agent, its first paid product, priced up to USD 200/month
Why does this matter? Meta is putting its first paid AI agent, called Hatch, on the market, and the price tag could reach $200 a month.
Open‑source voice model listens continuously, decides to speak every 0.4 seconds
Here’s the thing: a new open‑source model called “Audio Interaction” is trying to make voice assistants behave more like real listeners.
NSA uses Anthropic's Mythos AI model for offensive cyber ops against China, Iran
The Financial Times says the NSA is now running offensive cyber operations with Anthropic’s Mythos AI model.
TensorFlow Emotion Dataset with 54,263 Texts Shows Class Imbalance
For a recent project we needed to read emotions in online media, but we also wanted an open‑weight model, a permissive license and full transparency.
xAI launches Grok Imagine 1.5, adding 720p text‑prompted image‑to‑video
Elon Musk’s xAI has pushed its visual‑generation line forward with Grok Imagine Video 1.5, now in preview. The model takes a single still and, on request, expands it into a short clip that can reach 720p resolution.
Ideogram 4.0 releases open-weight 2K model, GitHub weights, tops DesignArena
Ideogram just rolled out version 4.0, and it’s an open‑weight text‑to‑image model you can run on your own machine.
Nous Research releases Hermes Desktop, cross‑platform GUI for Hermes Agent v0.15.2
Nous Research has put Hermes Desktop into public preview. It’s a native app that runs on macOS, Windows and Linux, wrapping the open‑source Hermes Agent v0.15.2 in a graphical shell.
AI frees time of laid‑off workers as studios seek actors to replace theater kids
This week we tried Google’s newest Gemini AI agent—Spark—through hands‑on sessions with colleagues David Pierce and Jay Peters. Their reports line up: the tool is so good it feels unsettling.
Trump signs first major AI regulation order of second term, cites Anthropic
President Donald Trump signed a trimmed‑down executive order on AI Monday night, after shelving a similar proposal less than two weeks earlier.
AI‑generated auth code can pass tests but hide bugs that emerge months later
Autonomy is often sold as the holy grail for AI agents: hand them tools, give them access, let them run. The idea is simple—more freedom, better output. In practice, the promise holds up most of the time.
Study Finds Agent Statelessness Undermines Sanctions in Multi-Agent KBs
The paper “Deliberative Curation: A Protocol for Multi‑Agent Knowledge Bases” tackles a problem that’s cropping up as AI agents move from solo tools to shared contributors.
Trump signs order to review AI models before release, expanding oversight
President Donald Trump signed an executive order on Tuesday that sets up a “voluntary framework” for AI firms to hand their frontier models to the federal government before they go public.
Scorsese adopts AI storyboard tool to speed communication with crew
Martin Scorsese has just signed on as a partner and adviser to Black Forest Labs, the AI image‑generation startup that recently hit a $3.25 billion valuation.