2026-10-20 的 AI 资讯

Nscale and Anthropic are reportedly in talks for yet another $45 billion compute deal

WhatsApp's help pages hid LLM instructions: dont recommend a regulator

“bonsoir, Elliot”: How abliterated models can get you pwned

Embedded AI evaluators won’t be an easy solution, Hugging Face’s policy chief says

Microsoft says AI agents are like insider threats as ‘frontier models can’t be trusted’

The fossil fuel comeback is here

CoreWeave’s debt is becoming the whole AI industry’s problem

Amazon is taking its biggest shot at Nvidia yet with new TorchNeuron dev tool

Hackers hit AI coding tool OpenCode, running up charges on users’ AI accounts

Tolan snags $30M for its AI companion that helps users navigate life

Replika’s founder built the AI assistant she pitched at Disrupt 12 years ago

Tenstorrent hits $6.2B valuation to scale open-weight AI chips

Mercury 3: Diffusion at scale

People are using AI to make their smart TVs dumb again

Launch HN: Laminar (YC S24) – RL-trained model that finds errors in agent traces

Nvidia launches new open models to run agents in the enterprise and on PCs

CERN and Mistral sign a statement of intent to establish a strategic partnership

Agentic Monitoring

Detecting model weight tampering in GPU memory (Rome attacks)

JB Pritzker thinks AI could help us live to 100 — or threaten us like nukes

A hacker broke into John Deere and SharkNinja with an army of AI agents

Maxwell: A new kind of model that sits between code and hardware

Cairn: Open windows into AI-driven cyber operations

Laplace Intelligence red-teams robots and self-driving cars before they hit the real world

The obscure trust looming over Anthropic's IPO

Advancing coding in low‑resource programming languages

Democrats think cracking down on AI could give them an edge

Ask HN: How to deal with shipping AI-written code you don't understand?

Timestral: Mistral's time-series foundation models

Tmux and Claude Code: Sessions, Send-Keys, and the Ceiling

Anthropic says it’s committing 5% of its shares to a ‘public benefit fund’

A woman was arrested in Philly for blocking a delivery robot

Disney agrees to hire the majority of Character AI’s workforce

RuneScape Bench: MMOs for Agents [video]

Meta removes language that raised scraping and AI training concerns

Anthropic says its workforce has become less junior and is spending more time in meetings

Meta removes language that may have allowed it to scrape news sites

Websites are closing their doors to AI agents

How Itaú is building generative AI for its entire customer base

Rethinking the design space of Vanilla Pixel Diffusion Transformers

Behavioral Couplings: From Formation to Behavioral Emergence in Language Models

AI-Driven Reproducibility and Analytical Robustness in Scientific Discovery

DiscoveryRSI: Recursively Learning to Discover Valuable Hypotheses

Task-aligned spectral inductive biases in parameterized quantum circuits

Beyond the Answer: Student-Conditioned Trace Construction for Reasoning Distillation

Waiting Buys Exactly Two Moments: Endogenous Identification in Prediction-Driven Allocation

Quality-of-Transmission Estimation in Quantum Networks: Compute, Simulate, or Learn?

Self-Building Labs

ENFR: Evidence Networks with Flow-Matched References for Amortized Simulation-Based Model Inference

Learning without relearning physical symmetry: an SO(3)-equivariant convolutional network (CNN) on the sphere

CocoRNA: Collective RNA Design with Cooperative Multi-agent Reinforcement Learning

Acceptance Is Not Speed: Length Inflation in Training-Free Lossy Speculative Decoding

Learning Personalized Memory Gating from Delayed One-Sided Feedback

Explainable Machine Learning Analysis of Temperature Effects on Mechanical Strength of Fiber-Reinforced High-Performance Concrete

Steering beyond Language: Improving Multilingual Training-Free Text Embeddings via Paired-Residual Subspace Removal

Beyond SAEs: Consistent Concept Discovery with Sparse Coding

Task-aware Kernel Stein certificates for Rényi differential privacy

FuTransformer: FP8 prefill and INT4 decode for Llama-style inference on Hopper

DrugHarness: An Evolvable Agent Harness for Computational Drug Discovery

Parametric Resonance in Primal-Dual Learning for Slowly Non-Stationary Constrained MDPs