The Archive

HyCNNs combine Maxout and input convex neural networks for learning convex functions with exponential parameter efficiency gains.

Select to Think identifies local sufficiency in SLMs to reduce LLM calls during reasoning divergence, lowering latency and cost vs. standard distillation.

Proposes learning online relaxation parameters for ADMM optimization in Model Predictive Control to improve convergence on problem classes.

Technical note on removing ln ln T term from Squint algorithm bound via shifted KT potentials and Krichevsky-Trofimov algorithm modifications.

ClassEval-Pro benchmark: 300 class-level code generation tasks across 11 domains, auto-constructed to address compositional coding between function and repo levels.

Theoretical analysis of learning curves for revenue-maximizing algorithms using distribution-free PAC framework, extending Cole-Roughgarden results.

DIRECT mechanism demonstrates neural assemblies can learn causal directionality through projection, plasticity control, and sparse winner selection.

Canonical's plan to add AI features to Ubuntu has some users asking for "a version of Ubuntu that does not include these features," while others say they'll stick with older versions of the Linux distro or even switch to a different one. After Canonical's announcement earlier this week that it's bringing AI features to Ubuntu, replies included requests for an AI "kill switch" or a way to disable the upcoming features, and comparisons to Microsoft's addition of AI features into Windows 11. Canonical's VP of engineering, Jon Seager, responded on Tuesday, stating that Canonical isn't planning to...

Reddit anecdote about Claude's response to a hypothetical moral choice scenario; no technical or policy substance.

Uninsurable war damage is forcing tech companies to rethink Middle East plans.

59 open-source Claude Skills library covering website lifecycle: brand, design, content, SEO, dev, ops, and growth.

inclusionAI releases Ling-2.6-1T, open-weight 1T-parameter model with MLA/Linear Attention hybrid for inference efficiency and coding tasks.

ClawGym framework for systematic development of file/tool-based agents, includes 13.5K synthesized tasks for agent training and evaluation.

Review of mm-wave and sub-THz oscillator designs for 5G/6G using CMOS, SiGe, III-V semiconductors; not AI-relevant.

Theoretical analysis proving pathwise convergence of finite-depth transformers to stochastic particle systems with quantitative bounds and noise-driven synchronization.

Extension of gradient boosting replacing decision trees with shallow neural networks (CNNs, Capsule Networks) for structured and unstructured data.

Unconfirmed report: DeepSeek and Kimi may merge as China consolidates AI labs following historical precedent in strategic industries.

Reddit post celebrates agentic workflows achieving reproducible outputs across consecutive runs; anecdotal observation without technical detail.

FaaSMoE: serverless multi-tenant MoE serving architecture decoupling control and execution planes to reduce resource underutilization in expert model deployment.

HealthNLP_Retrievers system using Gemini 2.5 Pro in cascaded pipeline for grounded clinical QA over electronic health records.

Claude Opus 4.7 user reports significant token efficiency improvement vs. 4.6, enabling heavier agent workloads.

Google recreates Cher's closet from "Clueless" with AI.

The next wave of enterprise productivity is being built on AI factories. As organizations deploy agentic AI systems capable of reasoning, automation, and... The next wave of enterprise productivity is being built on AI factories. As organizations deploy agentic AI systems capable of reasoning, automation, and real-time decision-making at scale, competitive advantage increasingly depends on the infrastructure that supports them. Success requires more than raw compute. It demands a scalable, predictable foundation that can orchestrate intelligent… Source

Google TV just got more Gemini features, including the ability to transform photos and videos with tools Nano Banana and Veo.

KAYRA is a containerized microservice pipeline for clinical karyotyping using EfficientNet, U-Net, and Mask R-CNN with dual cloud/on-premise deployment.

MoRFI identifies latent directions in sparse autoencoders that causally drive hallucinations during LLM fine-tuning on factual QA tasks.

Figure AI achieves 24x production scale, manufacturing 1 robot per hour, advancing toward commercial robotics deployment.

Knowledge distillation framework compresses YOLOv8-L to YOLOv8-S for edge automotive VRU detection with INT8 quantization on BDD100K.

Interview study of 22 recruiters reveals GenAI use in hiring workflows creates misperceptions of human agency despite recommendations for human-in-loop design.