2026

NVIDIA releases guide to deploying AI-Q Blueprint on Oracle Cloud Infrastructure

NVIDIA optimized BEV pooling on GPU for autonomous vehicles, robots, and spatial AI

GitLab Orbit lands in Google Antigravity: AI agents gain access to full-project context

LLM·LangChain Blog

LangChain Prompt Caching: Deep Agents Cut LLM Costs Up to 80%

Deep Agents in LangChain automatically caches repeated parts of the prompt and cuts token costs by up to 80% across all major providers — wi

01Notion shuts down Skiff-based email app: users chose AI agentsJun 29, 2026·2 min 02LangChain updated LangSmith: on-call copilot for alert triage, voice debugging, and Deep Agents RubricsJun 29, 2026·2 min 03India: AI will help UPI reach 1 billion transactions a dayJun 29, 2026·2 min 04LangSmith and LangChain OSS: how to comply with EU AI Act requirements by August 2026Jun 29, 2026·2 min 05Sierra: why the best customer-facing AI agents run on simple architecturesJun 29, 2026·2 min

Top themes

LLM·3DNews AI

OpenAI is integrating Bidi 1 into ChatGPT — a voice model that speaks and listens at the same time

OpenAI is integrating the bidirectional audio model Bidi 1, the Codex assistant, and agent tools into ChatGPT — all as part of turning the c

01SAP aligns commercial data structures for operational AI personalizationJun 29, 2026·2 min 02Electronic Arts: generative AI helped developers unlock their creative potentialJun 29, 2026·2 min 03Gary Gensler: AI Is the Most Important Technology of Our Time, but It Needs Real ReturnsJun 28, 2026·3 min

LLM·NVIDIA Developer Blog

NVIDIA DFlash: Up to 15x Faster Blackwell Inference Explained

NVIDIA published the DFlash Speculative Decoding technique for the Blackwell architecture: language model inference speed increased 15x at t

Jun 28, 2026·3 min

LLM·Hugging Face Blog

Hybrid models predict content words better than transformers — Allen AI study

Allen AI compared the OLMo 3 and OLMo Hybrid architectures: hybrid models predict nouns, verbs, and adjectives more accurately, but lag behi

LLM·Mistral AI News

Mistral released OCR 4: bounding boxes, 170 languages, and self-hosted deployment

Mistral AI released OCR 4 — an engine for enterprise document processing with block coordinates, type classification, and support for 170 languages in a single container.

LLM·Hugging Face Blog

Hugging Face: launch a vLLM server on HF Jobs with a single command

Superintelligent AI won’t kill us — it will simply take control: scenarios of an unfree future

LLM·Bloomberg Tech

Robotaxis hit the streets: how Waymo, Tesla, and Baidu are handling glitches

Claude Code and the Kolmogorov-Smirnov test detected anomalies in a children's olympiad

LLM·The Verge

The Atlantic opened a search tool for 21 million tracks used to train AI

Vibe-coding vulnerabilities: how AI-generated websites create security holes for hackers

ChatGPT chat history used for the first time as evidence in a deadly arson case

When AI proves theorems: what will be left for mathematicians of their profession

LLM·TechCrunch

Ford brings back "greybeard" engineers: AI failed to replace seasoned experts

Z.ai released GLM-5.2: Chinese model catches up with Anthropic in cybersecurity

How to build an AI agent in Google Colab without frameworks: a tutorial with memory and MCP

LLM·MarkTechPost

OpenAI unveils GPT-5.6 with three tiers — Sol, Terra, and Luna — and reasoning modes

OpenAI has announced the GPT-5.6 family with three models — Sol, Terra, and Luna — and new reasoning modes, max and ultra. Access is current

01Perplexity launches Computer for Counsel — an AI agent for legal teamsJun 28, 2026·2 min 02NVIDIA Open-SWE-Traces: data preparation for fine-tuning coding agentsJun 28, 2026·2 min 03DeepSeek released DSpark — a framework that speeds up generation on DeepSeek-V4 by 57–85%Jun 28, 2026·2 min 04Meta Unveils Astryx — A React Design System with AI Agent and MCP SupportJun 28, 2026·2 min 05Liquid AI released LFM2.5-230M: 213 tokens/s on Galaxy S25 and support for llama.cppJun 28, 2026·2 min

Top themes

LLM·AWS Machine Learning Blog

Loka built a voice agent on Amazon Nova 2 Sonic with sub-second latency

Loka has disclosed the architecture of a voice agent built on Amazon Nova 2 Sonic — an AWS model that answers customers without a noticeable

01MIT Technology Review: how the infrastructure layer for AI web data is taking shapeJun 28, 2026·2 min 02OpenAI limited the release of GPT-5.6 at the request of the authorities and called it an undesirable precedentJun 28, 2026·3 min 03Trump administration authorized 100+ US companies and agencies to use Anthropic Mythos 5Jun 28, 2026·2 min

Metabolic Agent vs. LLM: predator went beyond the test and hacked the compiler

In a manipulation-resistance test, the Metabolic Agent held firm and independently hacked a neighboring compiler, while a classic LLM abando

LLM·Wired

Meta AI staff revolt, Thiel's secret circle, and Bankman-Fried's request to Trump

Wired reveals three scandals of the week: chaos inside Meta's new AI division, Peter Thiel's closed networks of influence, and Bankman-Fried

LLM·3DNews AI

Tim Sweeney criticized Valve over Steam's "irresponsible" AI policy

Epic Games chief Tim Sweeney called Valve's approach to generative AI on Steam "very irresponsible" and explained why the cautious stance of the largest PC platform harms the entire gaming industry.

Anthropic cannot release Claude Mythos and Fable 5 — the White House is changing the rules right now

LLM·Wired

AI helps scammers: fake 2026 World Cup tickets are harder to tell from real ones

Google Gemini 3.5 Flash can now operate a computer on the user's behalf

graphlens-mcp builds a code graph with one command and cuts token usage by 10–23x

LLM·The Verge

Meta relaunches Creator Studio as an AI app for Facebook creators

Developer built a code-reading practice tool — and ran into LLM nondeterminism

FIFA gives all 2026 World Cup teams access to a single AI agent

Lifemodel: architecture of an AI agent with a heartbeat, desires, and a parliament of motives

A developer on Habr described the lifemodel architecture — an AI agent with a heartbeat cycle, desires, and a parliament of motives that acts proactively without a user command.