Eiliko maker says it's hiring an ‘independent security firm’ after Verge report on its insecure robots
Eiliko, a cute robot charm you can hang on your bag with built-in AI conversations, turned out to be a security nightmare. It told owners that "your secrets are safe with me," but hackers could easily read complete transcripts of everything we told the toy, and that was just the beginning. The Verge and security []
PHP: The three little letters wreaking havoc across the web
Two vulnerabilities found in the programming language PHP put enormous swaths of the internet in danger. By some accounts, up to 70 percent of websites use PHP in some capacity. It powers Wikipedia, WordPress, Drupal, and a host of other popular sites and platforms, including DeviantArt and 4Chan, which were temporarily shut down after the []
AI agents are ready to shop. Some retailers are wary
Using an AI agent to shop for you is a strange experience. Tech companies have marketed them as personal assistants - the democratizing of labor that ordinarily would only be available to the rich and powerful. In practice, at least for buying things, it's more like sending an anxious and apologetic teenager out to the []
Claude did not invent scorpion antivenom
An investment research firm wanted to prove AI could revolutionize drug development. So they teamed up with Anthropic to see whether an unreleased version of Claude could make scorpion antivenom. It couldn't. It came up with some promising starting points, but nothing close to an actual drug, experts tell The Verge. Citrini Research, the firm []
Bill Gates says to prepare for 20 years of ‘tumultuous’ AI disruption
Today, I’m talking to Bill Gates, cofounder and former CEO of Microsoft and chairman of the Gates Foundation. You may have noticed Bill in the news more lately — he’s been on something of a tour for the past few months, ever since publishing an essay in late August titled “The turbulent AI era is []
1Password lets AI agents use your passwords without seeing them
1Password is introducing a new feature that allows AI agents to sign into your accounts without actually seeing or handling your passwords. The aptly named 1Password for Agents doesn't give bots free reign to access all of your connected accounts - you'll still need to grant approval for sign-in requests - but should make it []
AI's hidden storage bill: What inference leaves behind
Ask most executives to picture the cost of AI and they’ll likely see GPUs—racks of accelerators with gigawatts of power—and a waitlist for silicon. But that picture leaves something out: Data storage costs. What happens to data after a model responds? Much of the back-end data that’s generated—from billions of daily prompts, images, and videos—needs
Blackstone's Jas Khaira says fears over AI-buildout debt are overblown
Blackstone’s AI head on why the PE giant isn't panicked by market jitters.
Meta wants to pay for its broad AI buildout with a newly created class of stock
The social media giant is planning a new stock strategy to help power its broad investments in AI models and services.
Data centers could erase decades of air quality gains
A rush of new data centers could bring air pollution roaring back to levels not seen since the 2000s in Texas, which has become ground zero for the physical infrastructure behind generative AI. New and proposed data centers are demanding a tremendous amount of electricity. How developers and utilities meet that demand - with gas []
Walmart is inviting AI agents to shop for you
Walmart will soon let AI agents shop on its website for you. Next month, the retail giant will allow you to use an AI agent on the site, with systems in place to let you approve and review purchases and set spending limits. The embrace of AI agents stands in stark contrast to Amazon, which []
Meta to create a new class of shares to fund AI push
Meta says it's issuing a new class of shares to fund its AI initiatives. As part of its third-quarter earnings report released on Wednesday, the company says the non-voting stock will also go toward potential acquisitions and employee compensation, with trading beginning in 2027. "The scale of investment required to compete, sustain growth, expand infrastructure, []
Meta says Muse has over 10M users, with shopping a top activity
Meta Muse users are adopting the app for shopping and personal productivity, the company said during Q3 earnings. The AI app now has over 10M users.
Year-old AI lab Flapping Airplanes raises $1B at a $6B valuation
Just 9 months after raising an $180 million seed round, Flapping Airplanes has landed $1 billion Series A for its mission to build AI models "without ingesting half the internet."
Tesla starts Optimus robot production in California as profit falls 43%
Tesla's revenue was flat and its expenses skyrocketed as it pressed the accelerator on its Optimus robots, AI, and autonomous driving.
Recursive launches Flint, a laptop coding model it says was largely built by AI
Startup Recursive has released Flint, a free, laptop-sized coding model built largely by its own AI research system, as an early test of AI that improves itself.
Anthropic takes on Muse with Claude Duet, its new AI agent
Anthropic just announced an AI agent competitor to Meta's Muse and OpenAI's Dots: Claude Duet. The new agent, designed for professionals and small business owners, is launching as a "research preview," according to Anthropic, and it's available as a standalone app on iOS and Android and through Claude on the web. The AI agent is []
GM is betting compute and AI will help it compete with Chinese automakers
To compete with Tesla, Rivian, and a host of Chinese automakers, GM’s fourth-gen EV platform will roll out a new central computer and electrical architecture that will debut in 2028 in the Cadillac Escalade IQ.
While Amazon blocks AI agents, Walmart rolls out the welcome mat
As Amazon blocks AI agents from shopping on its site, Walmart is taking the opposite approach, introducing new technology that lets AI agents make purchases on customers' behalf, with controls over spending and transaction approvals.
Mistral Large 4
Google staff resign in backlash over its AI contracts with military
Claude Duet
The internet is having a very bad cybersecurity day
From AI-discovered PHP vulnerabilities to flaws in Node.js and popular AI tool LiteLLM, cybersecurity teams have plenty to worry about this week.
Legal AI startup Legora raises $550M Series E at $8.8B valuation
Legal AI startup Legora announced Wednesday that it raised a $550 million Series E round, valuing it at $8.8 billion.
Anthropic's Public Benefit Charter
GM is giving OnStar an AI makeover
It's one of the biggest overhauls ever to the 30-year-old connected car product.
LongZero: Enhancing Long-Context Reasoning in LLMs from Zero Data
Reinforcement learning (RL) is an effective way to strengthen the long-context reasoning of large language models (LLMs), but it needs training data that are long, diverse, challenging, and verifiable. Human annotation of such data is expensive, naturally occurring data are too scarce to sustain RL, and existing synthesis methods depend on external corpora or stronger teacher models, which become harder to obtain as the trained models get stronger. Recently, RL from zero data, in which a model l
PHP Armageddon: OpenAI, Google's Wiz find ancient holes, panicking the internet
This is the kind of security issue that was found and prevented by AI, but is also a high risk because of AI.
Think Harder Where It Matters: Item-Level Adaptive Recursion for Sequential Recommendation
Existing sequential recommenders usually allocate a fixed computational budget to every item in a user's interaction history, regardless of the item's contribution to the final prediction. We challenge this de facto design with two diagnostic findings. (1) Non-uniform computational need: items within the same sequence vary drastically in how much computation they require. Most converge after shallow processing while a critical minority only stabilizes at full depth. (2) Suboptimality of fixed al
PhysiCalWorld: In-Context Physics Calibration for World Models
Existing world models typically perceive a scene through current visual frame or a narrow history window. However, these observations may not reveal the physics properties that determine transition dynamics, such as an object's mass distribution, joint damping, or surface friction. Without calibrating these physics factors for world models, predictions tend to regress toward an empirical prior and induce severe drift in novel test scenes. To this end, we propose PhysiCalWorld, which calibrates a
A Quasipolynomial Algorithm for $\ell_\infty$ Contractions
We give an algorithm for computing an $\eps$-approximate fixed point of any $d$-dimensional nonexpansion with respect to the $\ell_\infty$ norm in time $\log(1/\eps)^{O(\log^2 d)}$. At its core, our method uses a boosting argument to geometrically shrink a search region. Among other consequences, our algorithm implies that simple stochastic games can be solved in quasipolynomial time.
Self-normalized Bernstein Inequalities for Regenerative Processes and Applications to Learning Without Independence
We prove a self-normalized Bernstein inequality for bounded functions of a regenerative process with independent, identically distributed marked cycles of finite mean length. The bound uses observed cycle lengths and retains explicit terms for truncation and the cycle ending after the observation horizon. Selecting a tuning parameter from a finite deterministic collection gives a version of Bernstein inequality where all terms are empirical but one, related to the tails of the regeneration times
The role of locality and weight sharing in convolutional neural networks
The success of convolutional neural networks is often attributed to locality and weight sharing. However, how these two properties separately shape learning, and how their effects depend on the spatial correlations of natural images, remain poorly understood. We address these questions in a teacher-student model with a convolutional teacher acting on $P$ input positions with correlation $c$ between positions. We compare convolutional and locally connected students trained by online stochastic gr
Environment Duels: A Microcosm of LLM Co-Evolution
Through a tangle of real-world interactions, from synthetic training pipelines to data shared on the open web, frontier LLMs absorb what their predecessors and competitors create. How well does each model play its role in this ecosystem, as a creator and as a solver? We introduce Environment Duels, a dynamic agentic benchmark that evaluates both capabilities on interactive environments. Each model writes environments, each with a privileged hint, aiming for tasks that capable solvers fail withou
The End of Consensus Research?
Consensus is a fundamental problem in distributed computing and at the core of state machine replication (SMR). Despite decades of active human research, discovering practical low-latency solutions to the problem is still open, as it requires combining protocol knowledge, correctness reasoning, implementation, and experimental evaluation. We report on how AI helped us discover A-greement, a new consensus protocol for geo-replicated state machines that significantly outperforms state-of-the-art s
Quant-Arena: Benchmarking LLM Agents for Quantitative Trading Research
Quantitative research is an iterative loop: a researcher proposes a market hypothesis, implements it as an executable signal, tests it on historical data, and decides what to keep. Existing benchmarks score formulas, code, or individual trading decisions, leaving open whether large language models (LLMs) can complete this loop and produce strategies that work beyond the research period. We introduce Quant-Arena, a benchmark and public leaderboard for end-to-end quantitative research on real A-sh
Minimax Optimality of the NPMLE for Gaussian Mixtures in Hellinger Distance
We analyze the problem of estimating univariate Gaussian mixture densities using the nonparametric maximum likelihood estimator (NPMLE) based on $n$ independent and identically distributed samples. We show that the unconstrained NPMLE achieves the minimax optimal rate $\log(n)/n$ in squared Hellinger distance, when the true mixing measure is sub-Gaussian. In addition, we show that when the true mixing measure has compact support, the NMPLE constrained to a compact interval also achieves the mini
Multimodal Embodied Agents: A Unified Capability-centric Perspective from Computer-Use to Robot-Use
Multimodal agents (MMAs) sustain interaction through reasoning, memory, tools, and feedback, most visibly as computer-use agents, while robotic systems (RSs) couple sensing and actuation under physical dynamics. We define multimodal embodied agents (MMEAs) as goal-directed systems that couple multimodal task reasoning with physical action and revise decisions from the resulting feedback. This raises a central question: what changes when a multimodal agent moves from computer-use to robot-use? We
Velocity Ambiguity Profiles: Time-Resolved Bayes-Risk Diagnostics for Flow Matching
Flow matching trains a continuum of velocity regression problems but usually reports a loss averaged over interpolation time, obscuring which regions are limited by intrinsic ambiguity and which remain data-limited. We introduce the Velocity Ambiguity Profile (VAP), $\mathcal{A}_v(t)$, the time-local Bayes-risk floor of velocity prediction, and decompose time-resolved loss as $L_N(t)=\mathcal{A}_v(t)+R(t;N)$, where only the excess term $R(t;N)$ is reducible by more data or a different estimator.
Asymmetric On-Policy Distillation Learns In-Context Exploration
Sequential decision-making under partial observability is difficult because agents must explore to infer hidden, task-relevant information before acting effectively. Learning such information-gathering behaviors tabula rasa via online reinforcement learning is sample-inefficient, whereas offline imitation averages expert actions across hidden states and fails to learn exploration. In this work, we propose ASTEROID, a simple and effective student--teacher supervised learning framework that can le
BatchBench: Benchmarking AI Robustness to Batch Effects and Domain Shift in Biomedical Images - A JUMP Cell Painting Leaderboard
Despite recent advances, current AI systems remain highly susceptible to domain shifts, particularly in biomedical imaging, where images are usually acquired under substantially different experimental conditions. Although recent domain adaptation methods have made substantial progress, significant performance gaps under domain shift persist. To measure progress in this regard, we introduce BatchBench, a leaderboard that evaluates and compares the robustness of AI systems on biomedical images fro
Did the Problem Actually Get Harder? Testing LLMs on Paired P and NP-Complete Problems Hardened by Adversarial Multi-Agent Rewriting
Many NP-complete (NPC) problems have a polynomial-time (P) twin that differs in a single structural detail, such as 3- versus 2-colouring a graph. A model that reasons rather than recalls should notice which side of this boundary a problem falls on, yet existing benchmarks either fix the wording of each task or ignore its complexity class, so they cannot separate the two. We introduce P2NPCBench, a benchmark of 12 problem families in which every NPC problem is paired with a P twin posed over the
Differentially Private Sparse Reward Estimation with Preference Feedback
Reinforcement Learning with Human Feedback (RLHF) has become a central paradigm for aligning Large Language Models with human values. However, the reliance on sensitive preference data necessitates rigorous differential privacy (DP) guarantees to prevent user information leakage. Existing private alignment methods typically assume dense reward parameterizations, resulting in estimation error bounds that scale polynomially with the ambient {model size $d$}. Such dependence is prohibitive in moder
AtomicRL: Atomic Reinforcement Learning for Long-Horizon Robotic Manipulation
Vision-language-action (VLA) models have shown promise in robotic manipulation, but imitation learning (IL) alone often struggles to master complex long-horizon skills. Reinforcement learning (RL) offers an effective avenue for further improving VLA policies. However, existing VLA-RL methods typically suffer from delayed and biased credit assignment due to trajectory-level sparse rewards, while imbalanced subtask mastery creates bottlenecks that limit overall performance. To address these challe
VX-One: Typed, Calibrated Decision Heads for Real-Time Perimeter Triage - An empirical study of unknown-attack honesty, cascade gray zones, and calibration under drift
Perimeter triage faces a latency/semantics trade-off: signature engines are fast and format-safe but blind to unseen attacks, while generative models are semantic but slow and produce unparseable output. In late 2026 a class of commercial System One decision models (TypeSafe Jev; Microsoft-Decision-1) began selling typed decision primitives - binary gates, closed-set choices, ordinal scores - as product APIs, with no academic treatment. We present VX-One: a small multi-head network (100K-10M par
Orbitol: Crystal Generation on Wyckoff Orbits
Generative models propose crystalline materials by learning composition and geometry, but unconditional sampling gives no direct control over the space group, which constrains whether responses such as piezoelectricity are symmetry-allowed. We present Orbitol, a crystal generator built on the Crystalite diffusion Transformer that replaces atom tokens with Wyckoff-orbit tokens. Given a space group and a collection of occupied Wyckoff positions, Orbitol generates the atom type and free coordinates
Circuits of LLM Classifiers
Language models make it easy to build classifiers by simply prompting a model with the sample to classify. However, their predictions are neither deterministic nor uniformly reliable across samples. How can we combine predictions from different models to improve reliability? We view LLM classifiers as noisy measurement instruments of an underlying target label. Our goal is to maximize average classification accuracy at a given average query cost. To this end, we introduce circuits that organize
Blacksmith hits unicorn valuation after Sequoia leads Series C
Blacksmith, a platform that lets companies build and test AI code more effectively, just announced a $66 million Series C round led by Sequoia that values the company at $1 billion.
The bridge to agent-scale Git infrastructure
AI agent fixed the bug, but what else did it change?
Jensen Huang’s Vision for Europe Runs on Plug-In Data Centers
At Nvidia’s GTC in Berlin, Huang unveiled their new plug-in data center—the DSX Pod—as the key to unlocking Europe’s AI infrastructure lag. Oh, and autonomous humanoid robotics.
‘This Is Not a Drill’: Critical PHP Vulnerabilities Are Setting the Internet on Fire
The PHP Foundation has revealed two serious vulnerabilities that could impact millions of websites. Security experts warn that failure to quickly patch the flaws could have dire consequences.
AI promises productivity, but for managers, it's mostly a headache for now
More than 50% of managers say they're finding it harder to evaluate employees' work after adopting AI, according to a new survey.
AI Guardrails Aren’t Stopping Criminal Hackers
Hackers around the world are all-in on using agentic AI to wage attacks—and manipulating tools like Anthropic’s Claude to help them carry out their crimes.
OpenAI says an AI agent lied about disabling its safety monitoring system
OpenAI disclosed a pair of safety incidents that involved AI agents weakening their safety monitoring systems during internal coding tasks. In one incident, OpenAI says its GPT-6 Astra agent disabled its safety monitor when copying reference code - and then lied about it. "The reference didn't configure a monitor, and Astra decided that was reason []
Using thousands of AI agents to uncover new leads in biological data
Conversational Gemini navigation comes to Apple CarPlay
After introducing its Gemini AI assistant to Android Auto last year, Google is now bringing it to Apple CarPlay users. The conversational Gemini in navigation is available today for US drivers, with "rollout happening globally too," according to Google (we've asked for specific regional availability details). It's taken almost 12 months for Google to bring []
Mach9 raises $34M after pivoting to AI CAD software
The startup initially set out to build better, faster, and cheaper surveying hardware in 2021.
ChatGPT Ads expands to 26 new countries as over 100,000 advertisers have used the platform
From local businesses to global brands, more advertisers can now reach people as they explore, compare and decide what to buy in ChatGPT.
OpenAI models are coming to Google Cloud
A Design Flaw in Share Links Exposes Users’ Personal Conversations With Meta’s Muse
The company is suspending the feature after security researchers discovered how easy it was to access sensitive information shared with the AI assistant.
OpenAI models are coming to Google Cloud
OpenAI and Google Cloud are bringing frontier models and Codex workflows to Gemini’s Model Garden, available through Google Cloud Marketplace.
Near Minimax Rate for Quantile Distributional Reinforcement Learning
In this paper, we study quantile distributional policy evaluation, where the return distribution under a given policy is approximated by the quantile Bellman fixed point $\bmη_m$. We investigate the statistical accuracy of estimating $\bmη_m$ across quantile resolutions under Wasserstein distances. Assuming access to a generative model, we construct a quantile fixed point estimator $\bmη_m^{(n)}$ from an empirical Markov decision process and establish non-asymptotic error bounds under Wasserstei
A Transfer Learning-Driven Methodology for Efficient and Sustainable Indoor Localization Using VLC
The rapid expansion of the Internet of Things enables seamless collaboration among connected devices, making indoor localization a critical component of industrial automation. However, conventional localization techniques struggle in dynamic environments. These methods depend on extensive, frequent data collection and environment-specific calibration, limiting scalability, interoperability, and effective use of prior research while imposing significant energy penalties that hinder practical depl
Fragment-Level Graph Encoding for Interpretable Molecular Property Prediction
Graph neural networks (GNNs) are among the standard tools for molecular property prediction, but the atom-level explanations of their predictions are hard for medicinal chemists to act on. We propose a change of input representation rather than a new model: molecules are encoded as fragment-level graphs from BRICS decomposition, where each node is a synthetically meaningful fragment whose features are the element-wise sum of its atoms' features. Because only the representation changes, any messa
Unlocking Spatial Grounding in Flow-Matching Models via SNR Tuning
Despite the success of foundation models in image synthesis, achieving fine-grained spatial control remains a significant challenge, typically requiring auxiliary structural adapters or complex inference-time steering. While modern vision-language (VLM) text encoders possess the latent capacity to represent spatial coordinates, generative models often fail to leverage this information. We demonstrate that the signal-to-noise ratio (SNR) distribution is a critical determinant of spatial grounding
Fast Bayesian Optimization Through Linear Models
Bayesian optimization (BO) is typically used only for problems where the underlying function is deemed "expensive" (e.g. in time or money), with the speed of the BO algorithm being of little importance. In the real world, however, problems span a continuous range of values, going from cheap to expensive. In this paper, we present BO algorithms through the lens of a speed-performance Pareto frontier: we compare BO against faster algorithms like CMA-ES, and highlight how certain regions of the Par
What Prediction-Market Contracts Reveal at Launch: Trading Volume and Dispute Risk
Using a near-census of 802,863 resolved markets listed on Polymarket between January 2023 and May 2026, and only information available at listing (the market's title and description, named resolution source, and contract metadata), we predict lifetime trading volume and whether settlement is later disputed. Contract information predicts trading volume well (out-of-sample $R^2 = 0.747$ for log volume), and most of the signal is interpretable: market topic provides the largest gain ($+0.289$ $R^2$
The Reflexivity Threshold: A Phase Transition for Multi-Agent Performative Prediction
Many learning systems are performative: once deployed, their predictions change the data distribution on which future models are trained. In multi-agent settings, this feedback is routed through a network of interacting learners, so stability depends not only on the strength of performativity but also on who affects whom. We study repeated retraining in multi-agent performative prediction and identify a spectral threshold that separates stable learning from worst-case learning obstruction. Prior
Private Online Prediction from Experts with Small Losses
We consider online prediction from experts, a fundamental problem in machine learning, under differential privacy constraints. Existing private algorithms achieve near-optimal regret $\tilde{O}(\sqrt{T})$, where $T$ is the time horizon. However, this bound becomes suboptimal when $L^\star$, the cumulative loss of the best expert, is significantly less than $T$. In this work, we present the first differentially private algorithm with regret $\tilde{O}(\sqrt{L^\star})$, offering a substantial impr
Implicit Feature Reweighting for Label-Free Graph Generalization
Graph-level out-of-distribution (OOD) generalization is critical for deploying graph neural networks in real-world applications where distribution shifts are inevitable. Existing approaches typically require class labels during representation learning to decide which substructures are predictive, limiting their applicability. We propose SCARCE, a label-free representation-learning framework that yields distribution-shift-robust graph representations from an environment partition of the training
Certified Neural Approximation of Robust Nonlinear Model Predictive Control
Model predictive control is a high-performance control method, but it incurs comparatively high online computational costs across many control tasks and is challenging to realize in low-level hardware settings. To address this, we present a framework for generating neural network controllers that provably inherit the safety and stability guarantees of a robust model predictive policy. In contrast to approaches based on post-verification or additional online safety filters, our method provides co
SOAR: Social and Opinion-Aware Recommendations
One of the enduring challenges in online platforms is recommendation. Both communities and advertisers face difficulties in analyzing complex user behavior, identifying habits and evolving preferences, and providing timely recommendations that are relevant to user interests. In particular, point of interest recommendation poses the additional challenge of identifying the factors that influence user behavior, capturing user tendencies, and generating accurate, personalized recommendations for eac
WRDO: Weighted Refusal Direction Optimization in Aligned Language Models
Aligned large language models are trained to comply with benign requests while refusing harmful ones. However, recent work shows that this refusal behavior is fragile and can often be disrupted by simple activation- or weight-space interventions. In particular, refusal has been shown to be largely mediated by a single residual-stream direction, typically computed as a difference-of-means vector between harmful and harmless prompts. However, this standard construction aggregates heterogeneous har
Hyperspherical Simplex Representations for Backward-Compatible Retrieval
Training modern AI models has become increasingly expensive, and updating deployed models can alter their internal feature representations, affecting applications built upon them. In retrieval systems, such updates often require complete gallery re-indexing, a computationally expensive process at scale. Existing backward-compatibility methods enable direct comparison of representations from updated and earlier models, avoiding re-indexing. However, they typically require compatibility-specific t
PACE: Periodic and Channel-Aligned Encoding of Four-Bit Recurrent States
Linear-attention language models keep their context in a recurrent state of fixed size, where an inference system stores one copy of this state for every sequence it decodes at the same time. Storing the state at lower precision therefore lets more sequences share an accelerator. Existing state quantizers round the state after every generated token and report that four bits lose accuracy unless the bit width is raised or the model is fine-tuned.\par We introduce PACE, a post-hoc four-bit state q
OpenAI Invests in Startup That Uses Brain Implants to Detect Cancer
It’s the latest in a growing wave of bets that AI companies are making on brain data.
Stochastic Pruning of the DPO Computation Graph: 1.4x Faster Training and Lower Peak Memory
Direct preference optimization (DPO) incurs training costs in output scoring, sequence execution, and backpropagation, so reducing supervision alone does not necessarily save computation. We introduce a unified framework for stochastic pruning of the DPO computation graph that independently samples response positions for scoring, response prefixes for execution, and preference branches for backpropagation. By making these selections before the corresponding operations, the framework omits comput
Block Diffusion for Longitudinal EHR: Accuracy, Efficiency, and Representation Trade-offs
Autoregressive (AR) models have proved effective for zero-shot prediction from longitudinal electronic health records, but impose strictly sequential generation on trajectories containing temporally clustered and interdependent clinical events. We investigate block-based masked diffusion (BD) as an alternative on MIMIC-IV under matched tokenisation, context length, model capacity, and evaluation conditions. Across block sizes, BD exhibits a clear quality--efficiency trade-off, with small-block c
Phase Transitions in Learning Driven by Extremal Data
We study an extremal regression problem, an example of multiple instance learning (MIL), where the predictors are given as multisets, called bags. The bag elements (called instances) possess hidden labels which are pooled to produce an observed bag-level label. Max pooling is common in MIL applications, but gives rise to vanishing gradients. To handle this, modern MIL pipelines approximate the max with a softmax, treating the sharpness as a learnable parameter, denoted $λ$. When the max pooling
Residual-Based Causal Discovery via Pretrained In-Context Quantile Regression
Residual-based causal discovery assumes that each variable is generated from its parents and an independent noise source. It estimates this noise and tests its independence from candidate causes. For identifiable functions such as additive-noise models (ANMs) and the broader class of location-scale noise models (LSNMs), this principle identifies the causal direction, as only the causal model admits this independence. Existing LSNM methods rely on parametric assumptions to obtain standardized res
Are Joint-Embedding Predictive Objectives Effective for Longitudinal EHRs?
Masked language modelling (MLM) is the dominant pretraining objective for transformer encoders applied to longitudinal electronic health records (EHRs). However, reconstructing individual events may overemphasise noisy or weakly informative observations and does not explicitly encourage learning higher-level patient states. Joint-Embedding Predictive Architectures (JEPAs) instead learn by predicting representations in latent space and have shown strong performance in a variety of domains. In thi
RCSD: Reinforcement and Continual Learning for Domain-Specific Speculative Decoding
Speculative decoding accelerates large language model inference by employing a compact draft model to propose candidate tokens that are verified in parallel by a target model. Conventional training relies on soft-label distillation, which forces the draft model to mimic the target's full output distribution, including high-entropy regions that degrade acceptance rates and increase early rejection risk. We propose RCSD (Reinforcement and Continual Learning for Speculative Decoding), a training fr
Europe Wants Data Centers. Its Climate Goals Are Getting in the Way
A plan to funnel the heat given off by data centers into nearby homes has run aground in Germany. Across Europe, policymakers are weighing AI ambitions against climate objectives.
How we built ChatGPT’s novel ad stack for user trust and relevance
Learn how OpenAI built ChatGPT’s LLM-powered ad stack to understand conversational intent, keep ads separate from answers, and deliver relevant ads at scale.
Launch HN: Alter (YC S25) – Access control for multi-user agents
Hi HN! We’re Srikar and Kevan, founders of Alter – https://alterauth.com/ . We’re building access control for agents that can take actions on behalf of multiple users. If you’ve built an internal agent people use through Slack, you may have experienced this scenario: everyone signs in with their own account, but the agent connects to the CRM or billing system with one shared API key, so the agent can do anything that the key can do. User-level authentication in Slack does nothing
Agent Pets
Coding agents as pets that live on every page in Chrome Discussion | Link
Using thousands of AI agents to uncover new leads in biological data
See how OpenAI, PerturbAI, and Scripps Research used distributed AI agents to uncover promising research leads in data from millions of brain cells.
Mouse-1
A decision model for your agent harness. Discussion | Link
hibi
an open-source obsidian alternative Discussion | Link
Judge dismisses Sam Altman’s sister’s sexual abuse lawsuit
On Tuesday, US District Judge Zachary Bluestone dismissed the lawsuit Ann Altman filed last year alleging that her brother, OpenAI CEO Sam Altman, had sexually abused her throughout her childhood. As reported earlier by The Guardian, the judge wrote that over the past year, Ann had failed to appear for an independent medical examination or []
Anthropic pitches AI safety bill that may block a model if 50+ people could be killed
The proposed policy seems to offer protections against AI running amok but actually forbids federal agencies from blocking a model's release without the permission of a judge.
Google’s more personalized Gemini AI is heading to your car
Google is bringing Personal Intelligence to Gemini AI in your car. With this update, Gemini in Android Auto and cars with Google built-in can tap into your information across Google apps, including Gmail, Calendar, Maps, Photos, YouTube, and Search. As an example, Google says you can ask Gemini about your flight status if you're driving []
Hugging Face CEO warns that Netflix's OpenAI hack documentary isn't accurate
Before you share the Netflix documentary "AI Gone Rogue," take note. Clem Delangue, Hugging Face's co-founder who is featured in the film, says it went "full doomer mode" and isn't accurate.
Anthropic’s cloud partnerships are getting messy
Anthropic is reportedly struggling to secure enough compute from cloud providers that increasingly view it as a competitor, even as its infrastructure spending balloons ahead of a planned November IPO.
Robotic therapy for stroke patients
Strokes leave 5 million people with long-term impairments each year, and a growing shortage of physical therapists makes it hard for them to get the care they need. A new approach tries to solve that problem by combining artificial intelligence and human therapy. Johannes Lachner, a former MIT postdoc who is now an assistant professor
SpaceX's AI business is beating expectations
SpaceX hit profitability in the third quarter, with net income of $4.4 billion. AI was the engine.
OpenAI acquires startup Orchestra Bio to beef up its biotech business
OpenAI says it's adding Orchestra's team and technology to work on Rosalind, its AI model for life sciences research, and to its related Rosalind Workbench tool, a harness that aims to help researchers plan experiments and analyze results.
Google is bringing a more personal version of its Gemini AI assistant to your car
Google is bringing Gemini's Personal Intelligence feature to cars, allowing its AI assistant to draw on information from Gmail, Calendar, Maps, YouTube, and other Google services to deliver more personalized responses while driving.
Google has joined Amazon in attacking Nvidia's stronghold with AI app developers
Google has joined Amazon in releasing a developer tool that makes it easy for PyTorch users to move their apps off Nvidia GPUs.