Последние материалы и новости об AI

Статьи: 10

OpenAI Promises Zero Data Retention as Anthropic Lets Customers Keep the Data

OpenAI and Anthropic are competing over enterprise AI privacy, but their zero-retention promises differ in who stores data and how safety monitoring w...

2026-08-31

OpenAI Promises Zero Data Retention as Anthropic Lets Customers Keep the Data

PottsMPNN aims to expand AI-designed protein possibilities

MIT researchers developed PottsMPNN, a machine-learning framework that designs protein sequences beyond natural examples while improving structural an...

2026-08-30

PottsMPNN aims to expand AI-designed protein possibilities

Anthropic and OpenAI Join the AI Stage at TechCrunch Disrupt 2026

TechCrunch Disrupt 2026’s AI Stage will examine enterprise AI deployment, agent security, AI-native growth, video intelligence and evolving SaaS busin...

2026-08-28

Anthropic and OpenAI Join the AI Stage at TechCrunch Disrupt 2026

CUDA Python 1.0 Brings Stable APIs and Full CUDA Access to Python

NVIDIA’s CUDA Python 1.0, released with CUDA 13.3, unifies Python GPU development through stable APIs, shared CUDA objects, and access to the full pla...

2026-08-26

CUDA Python 1.0 Brings Stable APIs and Full CUDA Access to Python

NVIDIA Dynamo Restores LLM Inference Capacity in Seconds with Shadow Engine Recovery

NVIDIA Dynamo’s preview shadow engine recovery uses shared GPU memory and preinitialized standby engines to reduce GLM-5.2 failover from 283 seconds t...

2026-08-26

NVIDIA Dynamo Restores LLM Inference Capacity in Seconds with Shadow Engine Recovery

Granite 4.2 LLMs: How They're Built

IBM Granite’s Hugging Face post presents Granite 4.2 language models for efficient reasoning, multilingual generation, coding, and AI assistant workfl...

2026-08-26

OpenAI’s Jalapeño Chip Reportedly Outperforms Nvidia’s Blackwell and Rubin

OpenAI says its Jalapeño inference chip delivers higher performance per watt and lower latency than Nvidia’s Blackwell and Rubin accelerators.

2026-08-26

OpenAI’s Jalapeño Chip Reportedly Outperforms Nvidia’s Blackwell and Rubin

DoorDash’s SafeChat Uses Hybrid AI Moderation to Reduce Safety Incidents

DoorDash’s SafeChat combines fast internal models, LLM scoring, graduated actions, and no-code workflows to moderate marketplace content at scale.

2026-08-22

DoorDash’s SafeChat Uses Hybrid AI Moderation to Reduce Safety Incidents

World Models Need to Track Human Beliefs to Predict Actions, Study Finds

A new Mental World Modeling framework shows that AI agents predict human actions more accurately when they model beliefs, goals, emotions, and social ...

2026-08-22

World Models Need to Track Human Beliefs to Predict Actions, Study Finds

Study explains why AI agents benefit from “skills” and when they fail

A Princeton and UC San Diego study of 8,135 test runs finds that AI skills mainly improve procedures, while retrieval becomes difficult as libraries g...

2026-08-22

Study explains why AI agents benefit from “skills” and when they fail