Search the full wire by company, model, lab, or keyword. Every story we have ever aggregated.
Democratic ICAI improves preference-based alignment by eliciting multiple competing rationales through structured debate to extract nuanced steering principles from human preferences.
Nuclear physics paper using interpretable neural networks to model binding energies via SU(3)/SU(4) symmetries; domain-specific, not AI-frontier work.
Lighthearted suggestion that Anthropic should use friendlier model naming conventions.
PAC-Bayesian bounds for quadratic control with finite-sample guarantees; theoretical controls work, limited immediate AI systems relevance.
HORIZON: agent framework for hardware design via repository-level code evolution, evaluated on ChipBench and RTLLM.
Google proposes taxonomy for AI-assisted scientific peer review at four collaboration levels to address peer-review bottleneck.
PEHT transformer for cellular network traffic prediction via urban congestion features; domain-specific infrastructure application.
Mechanistic study of vision-language model conflict resolution: visual grounding dominates by default; knowledge priors require specific attention heads.
OpenAI launches GPT-5.6 series (Sol, Terra, Luna) with tiered performance/cost; limited preview underway, general availability in weeks.
Agent-Native Immune System (ANIS): runtime defense framework for autonomous agents against memory poisoning, tool-chain manipulation, multi-agent attacks.
TopoTTA: test-time adaptation for anomaly segmentation using topological structure; domain-specific computer vision application.
Stratechery weekly digest covering Apple regulatory challenges in Europe and miscellaneous tech commentary.
Less than 24 hours after news broke that OpenAI would stagger its next model release at the request of the Trump administration, that model, GPT-5.6, is here. On Friday, the company unveiled the limited preview of its new GPT 5.6 model suite: Sol, the flagship; Terra, a…
Hybrid classical-quantum neural networks for oral cancer detection on edge devices; quantum-computing application, niche deployment focus.
HPRO: preference-based reward optimization for emotional TTS via hierarchical separation of content and emotion features.
Theoretical analysis of generalization scaling laws in quadratic neural networks as functions of model width and sample count.
AI models have progressed to the point where their capabilities have real political consequences. Dealing with those consequences will require collective action.
Study shows autonomous AI coding agents pass individual benchmarks but accumulate integration friction in shared codebases, revealing ecosystem-level risks.
Identifiability theory for continuous-time latent SDEs using diffusion covariance shifts in causal representation learning.
DEFAR framework reduces exposure bias in Flow Matching generative models by exploiting dynamic signals during training.
Analysis of estimation-prediction tradeoff in probabilistic temporal graph models for link prediction.
Physics-informed neural networks with transfer learning applied to lithium-ion battery state estimation.
Framework for value-constrained credit assignment and reward allocation in delegated AI cooperatives using traversal learning.
HAT-4D agentic framework reconstructs 4D multi-object interactions from monocular video using VLMs for embodied AI data collection.
As context windows grow longer, moving large model weights efficiently becomes critical to performance. A common way to address this is quantization, an... As context windows grow longer, moving large model weights efficiently becomes critical to performance. A common way to…
Lagrangian duality-based method approximates best-response in non-linear strategic classification settings.
User critique of Anthropic's customer support responsiveness following Fable suspension due to US government directive.
COCOLogic-V2 benchmark dataset for visual logical reasoning on real-world images with fine-grained hard negatives for concept bottleneck model evaluation.
Time-series and ML models predict Sri Lankan remittances using exchange rates and oil prices; domain-specific economics work, not AI frontier.