The Archive

Meta releases Muse Wonder, positioned as first Fable-class model, scoring GPT-6.1 equivalent on Artificial Analysis; llm-meta-ai plugin updated.

"No more hand-written code" or AI-free art for delayed Marvel/Star Wars shooter.

“I may have information regarding this case," Claude AI told Philly police.

Anthropic presents dynamic workflows for coordinating Claude agents, applying Unix/Linux design principles to multi-agent orchestration.

Python 3.15 released with lazy imports (PEP 810) and performance improvements; end-of-life for 3.10.

Meta's Superintelligence Lab released a pair of new state-of-the-art models on Monday, Muse Wonder and Muse Wonder Full, the company's first attempt at a frontier model.

Russian bloggers warn Russia’s hyper-centralized digital economy is vulnerable.

A Silicon Valley startup that uses weather balloons and AI to predict the weather won a US military contract worth almost $50 million.

Meta is making up for lost time in the AI race. After launching Muse Spark, the first model from its revamped Superintelligence Labs division, and buzzy AI agent Muse, it's now launching two new models it claims can go toe-to-toe with the industry's best. The new models, Muse Wonder and Muse Wonder Full, are available for Muse Code subscribers starting today. Wonder can also be accessed through the Meta Model API, Meta said. Meta says it eventually plans to power its Muse AI agent with Wonder, starting with subscribers, and will develop specialized models it "can bring to everyone using Muse....

Simon Willison demonstrates Python 3.15 profiling tools (flame graphs, heatmaps) on a web-edit app using GPT-6 Astra for code generation.

Hybrid models interleave state-space model (SSM) layers with a few attention layers. NVIDIA Nemotron 3, Qwen3.5 through Qwen3.8, Kimi-K3, and GLM-5.3-Flash are... Hybrid models interleave state-space model (SSM) layers with a few attention layers. NVIDIA Nemotron 3, Qwen3.5 through Qwen3.8, Kimi-K3, and GLM-5.3-Flash are all built this way, with Mamba-2, Gated DeltaNet (GDN), or Kimi Delta Attention (KDA) as the SSM layer. SSM layers are designed to make decoding fast: they use a fixed-size state in place of a key-value (KV) cache that grows with… Source

Simon Willison releases web-edit 0.1a0, a Python web UI for remote file editing over Tailscale, built with GPT assistance.

Journalists, entrepreneurs, and even some public relations professionals will tell you that the PR process is broken. But could AI actually make things better?

Microsoft President Brad Smith calls for urgent AI regulation, saying milk and washing machines have more oversight than AI companies.

Blueprints AI, a bootstrapped startup based in Santa Monica, developed software that produces code-compliant construction blueprints from a photo, PDF, or sketch.

Microsoft’s Maia 300 chip. | Image: Microsoft Microsoft is announcing its third-generation Maia in-house AI chip today, designed to be the "primary AI inference system" for Microsoft's products and services. Maia 300 not only improves the cost and efficiency of running AI workloads in the cloud, it's also more capable of handling the latest reasoning models and agentic tasks. Maia 300 will start being rolled out to production in 2027, and it has been designed with efficiency in mind. "We expect Maia 300 to deliver 40 percent more tokens per megawatt compared to leading GPU-based systems," say...

Ambi is releasing apps for iPhone, Mac, and Apple Watch that enables ambient capture

The second-largest corporate employer has bet big on robot workers to scale.

Import AI newsletter roundup: robotics for lunar missions, open-weight model competitiveness gaps vs. China, and AI creativity applications; RSI concerns noted.

Stratechery announces no update published; administrative notice.

AutoNovo uses agentic program search to discover nucleic acid design algorithms that transfer across independent predictive models.

GUIDE framework enables LLM-based agents to facilitate multi-party communication tasks with dynamic adaptation across diverse group contexts.

NetRL applies LLM-based reinforcement learning to network decision tasks like bitrate streaming and congestion control with improved generalization.

GenRev introduces agentic image generation that selectively revisits prior intermediate states during iterative refinement to improve compositional outputs.

Analysis shows participant-disjoint cross-validation can yield optimistic estimates in Parkinson's gait classification when acquisition site confounds diagnosis.

OneShot weights time series anomaly detectors from a single labeled confirmation, adapting detector selection in deployed monitoring systems.

SafeDSE design space exploration framework optimizes safety architecture patterns and scheduling for mixed-criticality streaming dataflow applications.

MP-MNCA uses morphology-prior cross-attention and self-supervised learning for efficient spatial domain identification in transcriptomics.

Lüshi-1840 is a 983M-parameter classical Chinese language model trained on pre-1840 texts, addressing historical language and knowledge cutoff analysis.

MSC framework combines NMPC planning and RL execution to handle nonlinear caster dynamics and state uncertainty in autonomous mobile robot control.