Search the full wire by company, model, lab, or keyword. Every story we have ever aggregated.
Release: datasette-agent-micropython 0.1a0 I want Datasette Agent to be able to generate and execute Python code safely. This alpha is looking very promising so far. GPT-5.5 has so far failed to break out of the sandbox! Tags: python , sandboxing , datasette , webassembly , datasette-agent
Microsoft just kicked off Build 2026 with a keynote from CEO Satya Nadella and other company leaders. As expected, it was filled with announcements, ranging from new Surface hardware to an always-on personal assistant and updates across Microsoft's in-house AI models. If you didn't watch the event live, you can catch up on all the latest news in the roundup below. A mini Surface PC designed for AI development The Surface RTX Spark Dev Box is geared toward developers who want to run local AI models on their device, serving as a substitute for Qualcomm's canceled dev kit. It comes equipped with...
Release: micropython-wasm 0.1a1 Fixes for some limitations that emerged while I was trying to use this to build datasette-agent-micropython . Tags: python , sandboxing , webassembly
Uber's cutback has occurred after the company had reportedly encouraged staff to use AI as much as possible.
Microsoft on Tuesday took the wraps off Adaptive Spec-driven Scoring for Evaluation and Regression Testing, an open-source framework for spinning up AI evaluations.
AI agents are changing how you interact with your PC. Creators, developers, and AI enthusiasts are already using these agents extensively to assist with... AI agents are changing how you interact with your PC. Creators, developers, and AI enthusiasts are already using these agents extensively to assist with day-to-day tasks such as coding, video editing, and content management. NVIDIA and Microsoft are teaming up to enable the next generation of developers to build on-device agents on the Windows platform, with easier setup, native security… Source
President Donald Trump signed an executive order Tuesday creating a "voluntary framework" for AI companies to share their frontier models with the federal government before they're released "to promote secure innovation and strengthen the cybersecurity of critical infrastructure." The order says the US AI industry has succeeded in part "because we refuse to stifle this innovation with overly burdensome regulation," but that it also recognizes new AI capabilities come with security risks. Accordingly, it directs several federal agencies to come up with a framework to "assess the advanced cyber...
International Mathematical Union endorses warning about tech industry influence.
California Brown Pelican, in Fort Mason, CA, US I'm at the Microsoft Build conference today, held at Fort Mason in San Francisco. There are California Brown Pelicans diving into the water directly behind venue! Tags: microsoft , ai , generative-ai , llms , llm-release
The caveat is that one of the world's most famous living directors is using the tech solely for storyboarding.
Microsoft announced a bunch of new in-house AI models at Build 2026, including a new "flagship" model: MAI-Thinking-1. It's an ambitious step into model development for Microsoft, which introduced its initial in-house models last year - before then, it had relied on OpenAI's models. The two companies recently renegotiated their deal to loosen ties. According to Microsoft, MAI-Thinking-1 is a "medium-sized model" that "matches leading models" on "key" software engineering benchmarks. Microsoft says the company "trained it from the ground up on clean data, without distillation from third-party ...
Google's June Android feature drop includes more scam detection, more AirDrop, and yes, more AI.
The specification lets developer, compliance and security teams define their own policies for agents to follow in portable policy files.
Google is launching a new feature for its Phone app that aims to protect you from AI impersonation scams. Now, when you receive a call from a scammer that appears to be coming from the same number as one of your contacts, Phone by Google will flag the call as suspicious so you can hang up. In a post explaining the update, Google describes impersonation scams as a growing threat, with the FBI reporting that Americans lost over $893 million to scams using AI in 2025. To carry out the attack, scammers spoof one of your contacts' phone numbers and then use AI-powered tech to make their voice soun...
Much like Google, Microsoft is launching its own version of OpenClaw. Microsoft Scout is an always-on assistant that integrates into Microsoft 365 apps like Outlook, OneDrive, and Microsoft Teams, allowing businesses to assign a virtual assistant to employees to help with organizing calendars, expense reporting, email drafts, and much more. Unlike Copilot that lives inside Microsoft 365 apps, Microsoft Scout can see and do a lot more. "This is a personal assistant, it's the first real personal assistant we've offered customers," explains Omar Shahine, corporate vice president of Microsoft Sco...
As people increasingly refuse to answer calls from unknown numbers, scammers are shifting their tactics by spoofing trusted phone numbers and using AI deepfake technology to sound like authority figures, family members, or employers.
We investigate whether neuron populations within neural networks evolve predictably with scale, extending scaling laws beyond macroscopic observables such as loss. To probe this question, we study Rosetta Neurons, a previously characterized class of neurons whose activation patterns are similar across independently trained models (Dravid et al., 2023). In separate analyses of language models up to 30B parameters and vision models up to 5B parameters, we observe that the population of Rosetta Neurons follows a sublinear power law in model size, growing in absolute number but occupying a shrink...
Vision language models (VLMs) excel at many tasks but still struggle with spatial reasoning when critical information is not directly observable. Many such problems require imaginative perception: inferring what would be seen from an unseen viewpoint, tracing paths through occluded spaces, or integrating partial observations into a coherent spatial representation. We introduce Imaginative Perception Tokens (IPT), intermediate perceptual representations that externalize what a VLM would perceive under alternative spatial configurations while remaining consistent with the observed input. To stu...
We introduce Humanoid-GPT, a GPT-style Transformer with causal attention trained on a billion-scale motion corpus for whole-body control. Unlike prior shallow MLP trackers constrained by scarce data and an agility-generalization trade-off, Humanoid-GPT is pre-trained on a 2B-frame retargeted corpus that unifies all major mocap datasets with large-scale in-house recordings. Scaling both data and model capacity yields a single generative Transformer that tracks highly dynamic behaviors while achieving unprecedented zero-shot generalization to unseen motions and control tasks. Extensive experime...
Quantities with measurement units, such as 110 cm and 1.2 m, require language models (LMs) to combine a numeral with a symbolic unit scale. Here, we study how LMs compare such quantities in controlled settings spanning several unit systems. We find that accuracy degrades near the comparison boundary, where small changes in value determine the correct answer. The resulting errors are systematic: linear surrogate models predict LM preferences from numerical-difference and unit-scale-difference cues, and causal interventions on subspaces aligned with these variables shift model's output. The res...
Reward models (RMs) provide critical feedback signals for LLM post-training, notably in reinforced fine-tuning (RFT) and reinforcement learning (RL) pipelines. However, current reward evaluation relies on heterogeneous criteria such as rule-based verifiers, ground-truth references, procedural checklists, and complex rubrics, where a unified mechanism to integrate all types of evidence remains unexplored. To this end, we propose Skill Reward Model (Skill-RM), a unified framework that reformulates reward modeling as the execution of a reusable Reward-Evaluation Skill. By treating reward computa...
The past few decades have witnessed significant advances in the design of machine learning algorithms, from early studies on task-specific shallow models to more general deep Large Language Models (LLMs). Despite showing promising results in tasks that require instant prediction or in-context learning, existing models lack the ability to continually learn and effectively transfer their temporal in-context knowledge to their long-term parameters. Inspired by human learning process, we introduce a ''Sleep'' paradigm that allows the models to continually learn, distill their short-term fragile m...
Representations of the world, arguably, contain information about features (e.g. something is blue, something is a circle) but also information about which features are part of the same object (e.g. the circle is blue), which we call binding information. Any system with the ability to understand scenes with multiple objects must be able to solve the binding problem: it needs to know which features belong together. However, despite work showing that Vision Transformers (ViTs) know which patches belong together, it is not known whether current deep learning models learn to exhibit binding infor...
Reliable uncertainty communication is critical to the trustworthiness of LLMs, yet faithful calibration (FC)--the alignment between models' intrinsic and (linguistically) expressed confidence--is a persistent failure mode. This challenge is key for large reasoning models (LRMs), whose extended reasoning traces are often interpreted by users as evidence of deliberation, competence, and confidence. Despite the importance of FC and wide usage of LRMs, the extent to which LRMs can faithfully express their confidence remains poorly understood. Moreover, the prevailing paradigm to measure FC does n...
Rubric-based RL is a promising route for extending reinforcement learning beyond verifiable rewards, yet existing methods optimize rubrics while treating the query distribution as fixed. We identify a structural bottleneck: rubric quality is constrained by query structure. Open-ended queries yield vague rubrics; naively narrowing them introduces fabricated references that no model can verify, so all responses fail and training receives no reward signal. We present QUBRIC, a framework that co-designs queries and rubrics. Teacher-derived key points ground the rewriting of open-ended queries int...
We describe AlignAtt4LLM, an IWSLT 2026 simultaneous speech translation system for English to German, Italian, and Chinese. The system is a synchronous cascade: Qwen3-ASR with forced alignment produces an incrementally updated source transcript, and Gemma-4 E4B-it translates that prefix under an MT-side AlignAtt policy. To our knowledge, this is the first application of AlignAtt to a decoder-only LLM, where the encoder-decoder cross-attention used by earlier AlignAtt systems is absent. We recover a usable policy by proposing (1) an explicit source span in the prompt, (2) offline selection of ...
Large language models improve final-answer accuracy through extended chain-of-thought reasoning, but often spend tokens inefficiently and offer little inference-time control. Existing efficient reasoning methods control thinking length by shortening, early-stopping, or compressing traces, leaving how the model thinks implicit. In this paper, we propose Agentic Chain-of-Thought Steering (ACTS), which formulates reasoning steering as a Markov decision process where a controller agent adaptively steers a frozen reasoner during inference. At each step, the controller observes the reasoning trace ...
Deep reinforcement learning has shown strong potential for enabling autonomous robots to learn complex navigational tasks. However, its practical use still depends heavily on human designed reward functions and repeated manual fine tuning, which is time consuming and does not guarantee high success in the desired task. This paper presents AgenticRL, agent guided reinforcement learning framework that increases autonomy in reward design, policy refinement, and real world deployment for unmanned aerial vehicles (UAV) navigation tasks. AgenticRL uses a multimodal generative pre-trained tansformer...
Classical reinforcement learning (RL) typically seeks a deterministic policy that maximizes the expected sum of a scalar reward. Yet, modern applications such as language model fine-tuning or scientific discovery demand diversity. Existing remedies such as entropy regularization or diversity bonuses often require fragile trade-offs that sacrifice performance for stochasticity or rely on heuristic metrics that can misalign policy rankings. We argue that diversity is more naturally understood as the rational response to uncertainty in the reward. When the reward function is not perfectly known-...
The class action lawsuit, filed in Seattle by Virginia resident Charles Sigwalt, claims that Ring's Familiar Faces feature stores images of passersby without consent.