AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchHow Should Transformers Encode Numeric Values in Electronic Health Records?
arXiv:2607.01391v2 Announce Type: replace-cross Abstract: How do we encode numeric values in transformer-based sequence processing, particularly in electronic health record (EHR) data? We systematically compare discrete,…
- arXiv cs.AIresearchWAM-TTT: Steering World-Action Models by Watching Human Play at Test Time
arXiv:2607.06988v3 Announce Type: replace-cross Abstract: Steering robot foundation models (RFMs) toward new task variants or user-preferred behaviors remains challenging, often requiring additional robot demonstrations,…
- arXiv cs.AIresearchHow Do I Know What to Say Next? Barenholtz's Autogenerative Theory as an Enrichment of Harrisean Integrationism
arXiv:2607.07891v2 Announce Type: replace-cross Abstract: Roy Harris's Integrationist linguistics offers a compelling critique of the referentialist tradition embedded deep at the heart of computational approaches to lan…
- arXiv cs.AIresearchSemantic Anchoring for Robotic Action Representations
arXiv:2607.13597v3 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models inherit rich semantic representations from pretrained Vision-Language Models, yet fine-tuning on limited robot demonstrations…
- arXiv cs.AIresearchFinVerse: Toward More Realistic Evaluation for Financial Time-Series Forecasting
arXiv:2608.03259v3 Announce Type: replace-cross Abstract: Time-series foundation models (TSFMs) are increasingly evaluated on broad benchmarks that apply uniform conventional forecast evaluation metrics across heterogene…
- SiliconANGLE AIresearchApollo analysis shows no one besides the biggest tech giants is seeing benefits from AI
Despite all of the hype and the multiple millions of dollars invested in artificial intelligence, most of corporate America outside of its biggest technology firms is yet to see any financial payoff, according to an anal…
- Apple Machine Learning ResearchresearchGHARP: Real-Time Gaussian Head Animation from Large-Scale Reconstruction Prior
We present GHARP (Real-time Gaussian Head Animation from Large-scale Reconstruction Prior), a method that animates 3D human heads in real time from a few input images of a subject and a driving expression signal. We deco…
- OpenAI NewsresearchWorking with academic communities to build scientific understanding
OpenAI shares model results on the Kakeya conjecture, Exponential Time Hypothesis, and quantum chromodynamics—and how researchers can verify and build on them.
- LessWrongresearchDifferent eval awareness may track different graders
TL;DR: Qwen3.6 27B shows different kinds of eval awareness, most notably on safety and capabilities evals. The split seems to track how the eval is graded, by an LLM judge or by an automated checker. One explanation is t…
- LessWrongresearchTracing Subliminal Learning Through Pretraining
This post summarizes our paper , Tracing Subliminal Learning Through Pretraining, which was developed as part of EleutherAI’s SOAR program . TL;DR We study subliminal learning , the phenomenon in which a student model ac…
- Tom's Guide AIresearchWill your neighborhood flood during a hurricane? An expert explains how AI can predict vulnerable areas before extreme weather strikes
Consumer AI launches, tools, and practical coverage.
- LessWrongresearchIs GPT 2XL already clean?
In which I investigate if Opus is able to interpret neurons better than GPT 4. TLDR: Opus 5.5 is able to identify 3-19% of the neurons of GPT 2-XL based on activations. This is a big improvement over the 0.3% that GPT-4…
- The DecoderresearchAnthropic co-founder Jack Clark moves his prediction for self-improving AI forward to 2027
A new benchmark by P-Zero Research measures how well AI models design and evaluate experiments on their own. The results are improving so fast that Anthropic co-founder Jack Clark now thinks self-improving AI could arriv…
- LessWrongresearchAn Attempt at Top-down Interpretability
There are at least two distinct ways to conceptually organize a scientific endeavor: top-down or bottom-up . The bottom-up approach involves finding the smallest objects we can confidently understand, studying them, and…
- The DecoderresearchAI makes junior patent attorneys instantly better but may quietly stunt their professional growth
A three-month field experiment with 133 patent attorneys shows that AI boosts immediate performance, with junior lawyers gaining the most. But when the AI is taken away, only senior attorneys retain the benefits, suggest…
- AI Stack ExchangeresearchHow to deal with delay in reinforcement learning, an unclear case
According to the question in How to deal with the time delay in reinforcement learning? , we can tell the delay in the reinforcement learning can be observation delay, action delay and reward delay. I have a special case…
- The DecoderresearchFancy harnesses make coding agents more expensive but not better, study finds
A controlled study shows that switching between coding-agent harnesses like Claude Code, OpenCode, or mini-SWE-agent barely moves pass rates more than rerunning the same setup. Cost per task, however, can triple, driven…
- LessWrongresearchLM self-models are built from indirect sources
Imagine you ask a language model "What is your personality?" as the first message of a conversation, with no system prompt. The model has no data about its own behavior in its context, so its answer can only rely on cach…
- LessWrongresearchSelf-reward can elicit diverse model personas
Epistemic status: This is mostly a toy project rather than a thorough scientific investigation. Most design choices are not ablated. We list some caveats at the end of the post. We let Qwen3.5-2B choose its own reward du…
- LessWrongresearchLoReFT GRAFT Better at Preventing RL Grift
tl;dr. I show that the recently introduced GRAFT method can be improved upon by using a single layer rank 1 Low-rank Linear Subspace Representation Fine Tuning (LoReFT) adapter as the FORGET adapter, using only 0.029% of…
- AWS Machine Learning BlogresearchMulti-Turn Reinforcement Learning on Amazon SageMaker AI: A hands-on guide
Multi-Turn Reinforcement Learning (MTRL) on Amazon SageMaker AI trains models on the outcome of a whole multi-step task rather than labeled steps. This hands-on guide walks the full pipeline using Wordle: deploying the a…
- OpenAI NewsresearchRapidly scaling online storage to serve over 1 billion ChatGPT users (Part II)
Learn how OpenAI scaled Habitat with reliable multi-tenancy, global caching, and Azure Cosmos DB to deliver 12× faster reads for ChatGPT.
- The Register AI + MLresearchAI models think humans are a bunch of amoral sheep
Ask LLMs for their most controversial take on humanity, and it turns out they have little nice to say about us, as one AI experiment found
- The Register AI + MLresearchAI models describe humans as a bunch of amoral sheep
Ask LLMs for their most controversial take on humanity, and it turns out they have little nice to say about us, as one AI experiment found
- Hugging Face BlogresearchLitPivot: Developing research ideas through dynamic contextualization within the literature landscape
Open-source model releases, tooling, and community updates.
- Latent SpaceresearchAI Is Getting Good at Math. Here’s Why.
AI models are solving more mathematical problems. A pattern is emerging in their successes that could help us predict which problems they’ll solve next.
- NVIDIA Developer BlogresearchUsing AI Agents to Teach Robots Like Microduck to Walk in Simulation
Teaching a robot to navigate an obstacle course requires more than a walking policy. It must perceive its surroundings, maintain balance, and adjust its actions...
- AI BusinessresearchPrompt: AI moves beyond the screen
Physical AI is pushing artificial intelligence into robots, factories and other real-world environments, where understanding language is only part of the challenge.
- Guardian TechnologyresearchExperts cast doubt on Anthropic report detailing how AI has altered work at the company
Anthropic claims AI will shift how work gets done and labor demand – but some experts are skeptical of the vision Last year, the Anthropic CEO, Dario Amodei, warned that artificial intelligence could eliminate half of en…
- SiliconANGLE AIresearchHow to avoid AI’s ‘valley of despair’
Amdahl’s Law states that improving a single part of a system can only speed up the whole system so much. Even if one step of a process becomes exponentially faster, the other steps impose a ceiling unless they, too, are…
- TechRadar AIresearchScientists say they’ve figured out the exact date when Van Gogh painted The Starry Night
Mainstream AI news from a large tech publication.
- The DecoderresearchAnthropic's Claude now writes code and proves it safe automatically
Anthropic had Claude generate mathematical proofs that existing code is correct, producing roughly 180 million lines of proof code for 123 open-source libraries. Along the way, the model caught bugs that years of testing…
- Ai2 BlogresearchLitPivot: Developing research ideas through dynamic contextualization within the literature landscape
LitPivot helps researchers refine early research ideas through literature-supported critiques and revisions as their proposals evolve.
- arXiv cs.AIresearchAccelerating Wildfire Hazard Mapping Across Canada with an AI Surrogate Modelling Framework
arXiv:2610.30170v2 Announce Type: replace Abstract: Canada's increasing wildfire activity requires frequent national hazard mapping to guide proactive mitigation, including fuel management, land-use planning, infrastruct…
- arXiv cs.AIresearchNeural Networks with Differentiable Belief-Function Layers
arXiv:2610.30173v1 Announce Type: new Abstract: Yager's 1987 rule of combination $\boxplus$ was proposed on epistemic grounds: under high conflict, mass should be redistributed to ignorance rather than renormalised onto…
- arXiv cs.AIresearchWhen Stable and Preferred Semantics Agree: A Path-Based Approach
arXiv:2610.30176v1 Announce Type: new Abstract: We investigate structural conditions under which preferred and stable semantics agree in abstract argumentation frameworks. Our approach is based on suitably chosen sets of…
- arXiv cs.AIresearchBreaking the high-degree barrier in Lehmer's problem with a human-AI discovery loop
arXiv:2610.30277v1 Announce Type: new Abstract: While artificial intelligence has made rapid progress in mathematical reasoning, many open problems depend on structure that is revealed only through large-scale computatio…
- arXiv cs.AIresearchRoute-and-Generate: KL-Regularized Router Improvement for Reasoning Routing in LLMs
arXiv:2610.30351v1 Announce Type: new Abstract: Large language models often apply complex reasoning indiscriminately, leading to overthinking on simple tasks and underthinking on harder ones. We present Route-and-Generat…
- arXiv cs.AIresearchSoft Ontology Priors over Latent Concept Posteriors for Weakly Supervised Neuro-Symbolic Learning
arXiv:2610.30391v1 Announce Type: new Abstract: Neuro-symbolic (NeSy) predictors that supervise only the task label suffer from reasoning shortcuts: distinct concept mappings achieve the same maximum likelihood, leaving…
- arXiv cs.AIresearchSame Beliefs, Different Policies: Learning Reusable Representations for Early Stopping in LLM Reasoning
arXiv:2610.30450v1 Announce Type: new Abstract: Reasoning LLMs often continue reasoning after reaching a correct answer, wasting computation and sometimes replacing it with an incorrect one. Deciding when to stop is a me…
- arXiv cs.AIresearchCan LLMs Improve Their Own Harness? A Controlled Study on ARC-AGI-3
arXiv:2610.30499v1 Announce Type: new Abstract: A generally intelligent system acquires new skills efficiently without requiring humans to redesign its software scaffold for every new task. Yet language models rarely act…
- arXiv cs.AIresearchModel Organisms of Collusion: Corrective Transfer in Untrusted Monitors
arXiv:2610.30582v2 Announce Type: replace Abstract: AI control protocols aim to get useful work from models that may be misaligned by monitoring their behaviour. One approach uses a second copy of the same model as the m…
- arXiv cs.AIresearchLarge language models show systematic failures in incorporating evolving medical knowledge
arXiv:2610.30631v1 Announce Type: new Abstract: Medical knowledge evolves as clinical guidelines revise recommendations, yet whether large language models (LLMs) faithfully incorporate such changes remains unclear. We in…
- arXiv cs.AIresearchSpeculative Acceptance Does Not Decay with Position: A Full-Length Diagnostic on Long Chain-of-Thought with Independent Drafts
arXiv:2610.30644v1 Announce Type: new Abstract: The claim that speculative-decoding acceptance rate decays as a chain-of-thought grows longer has become a de facto industry narrative, motivating draft-model co-design and…
- arXiv cs.AIresearchGroup Relative Policy Optimisation Improves Multilingual Speech Recognition in Low-Resource Languages
arXiv:2610.30663v1 Announce Type: new Abstract: Autoregressive encoder-decoder speech recognition models trained with teacher forcing can suffer from exposure bias, limiting their ability to recover from their own predic…
- arXiv cs.AIresearchScaling Laws for Safety Probes
arXiv:2610.30678v2 Announce Type: replace Abstract: Safety probes trained on a model's activations are a standard tool for monitoring language models in real time. Each time a new model is released or a new harm area nee…
- arXiv cs.AIresearchDisentangling Robustness and Sycophancy in Multilingual LLMs
arXiv:2610.30696v1 Announce Type: new Abstract: When a multilingual LLM agrees with a user's false premise, is it being sycophantic, or does it simply not know better in that language? Current research does not distingui…
- arXiv cs.AIresearchFaithful Latent Harm Monitors Work No Better Than the Models Own Behaviour at a Matched False Alarm Rate
arXiv:2610.30727v1 Announce Type: new Abstract: Large language models have been reported to encode the harmfulness of a request separately from their decision to refuse it, and latent safety monitors such as Latent Guard…
- arXiv cs.AIresearchHow Firm Are AI Models' Ethical Boundaries? A Pay-Stub Editing Test of 15 Models
arXiv:2610.30751v2 Announce Type: replace Abstract: Can a small change in wording move an AI model from refusing a sensitive request to acting on it? We test 15 models on the same request to increase the income shown on…
- arXiv cs.AIresearchCan Loan Processors and Underwriters Spot AI-Edited Statements? An Empirical Study
arXiv:2610.30758v2 Announce Type: replace Abstract: AI-assisted editing can produce loan documents that preserve familiar issuer layouts while changing the applicant's identity. We investigated whether experienced underw…