2026-10-21 的 AI 资讯

Judge dismisses Sam Altman’s sister’s sexual abuse lawsuit

Anthropic pitches AI safety bill that may block a model if 50+ people could be killed

Anthropic recorded a staggering $80B net loss in the first half of 2026

Hugging Face CEO warns that Netflix’s OpenAI hack documentary isn’t accurate

Google’s more personalized Gemini AI is heading to your car

Anthropic’s cloud partnerships are getting messy

SpaceX’s AI business is beating expectations

OpenAI acquires startup Orchestra Bio⁠ to beef up its biotech business

Anthropic logged net loss of over $80B in first half of 2026

Google has joined Amazon in attacking Nvidia’s stronghold with AI app developers

ChatGPT now lets you lock your chats

Meta pauses Muse sharing after researchers find AI chats on the web

Eric Schmidt reportedly owns about half as much of Anthropic as CEO Dario Amodei

Claude designed protein binders against scorpion neurotoxins

Meta’s Muse hits 10M U.S. downloads in 40 days, a milestone reached by only 11 other apps

Benchmarking Bot Detection Systems Against Modern AI Agents in 2026

Google open-sources TorchTPU, a native PyTorch back end for TPUs

Plants can speak using electricity, so this startup built an AI model to decipher their language

Show HN: Penmentor – Claude marks up your handwritten math like a teacher

Launch HN: Respan (YC W24) – Open-source Rust AI gateway built around plugins

Spotify pulled 50 million AI slop and spam tracks from its recommendations

Figma co-founder Dylan Field thinks MCP is not a perfect protocol

US orders federal agencies to remove China’s DeepSeek AI from their systems

Adobe Lightroom mobile can now open closed eyes in photos with a simple tap

Publish fake stats, get cited in LLMs

Eka is worth $1.45B thanks to a physical AI model that has experts raving

Otter.ai raises $320 million from General Catalyst’s growth fund for its meeting tool

Yelp can now call on your behalf to check wait times at a restaurant

Franchisee reconciles cash in 94% less time with ChatGPT Work

DeepInfer: High-performance inference framework for DeepSeek models

OpenAI acquires Orchestra Bio

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

How abliterated models can get you pwned

The fossil fuel comeback is here

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

Text-to-Image Generative Modeling Without Language Pretraining

OpenVLHarness: Real World Visual Agentic Tool-Use with Multimodal Memory

Can It Actually Work? Synthesizing Functional CAD Mechanisms

Reinforcement Learning with Visually Attributable On-Policy Self-Distillation

CoopUQ: Robust Multi-Robot Social Navigation via Uncertainty-Aware Cooperative Forecasting

Where is Physical RSI capability from? An Observational Scaling Law of LLM-as-policy

WeAct: Dynamic Weight Switching as Agent Actions

Train An Imaginative Agent

AnyGoal: On World Model Generalization Beyond Expert Rollouts

LatComm: Reusable Context Compression for Latent Communication between Large Language Models

Newton-Based Methods for Bilevel Optimization Beyond Lower-Level Strong Convexity

Same Simulation, Different Evidence: Where Retrieved Tool Output Costs LLM Agents Accuracy

Kronecker Delta Rule for Long Context Linear Attention

Supervised Learning with Quantum Hamiltonian Embeddings

Competing Learning in Hybrid Architectures

ProteinGeoBench: Frontier Structural-Biology Agents Are Fluent at Lookup but Lost in Geometry

Agent World Model in the Wild: Computer-Use Agent Environment Synthesis

KEEP-Bench: Evaluating Evidence Preservation in Context Compression

RefCraft: Benchmarking Visual-Reference-Conditioned Agentic Generation through Web Development

MemT-1.0: Towards Memory-Native Evolution

SpecMRL: Redesign Matryoshka Representation Learning within Parametric Spectrum

AfriBED: a Speech Recognition Bias Evaluation Dataset for Low Resource African Languages

ASTOR: Benchmarking and Improving Autoformulation for Soft-Constraint Optimization with a Formal Grammar

Next Best Action: Causal Constrained Offline RL for Business Action Targeting

Safety Mid-Training: Internalizing LLM Safety as a Foundational Capability