2026
7,718 articles · 95 days · 10 categories · avg 81 per day

NVIDIA Shows the Difference Between Evaluating Models and Evaluating AI Agents
Model benchmarks and agent evaluation solve different tasks: the former tests language understanding, the latter—the real-world behavior of

DeepMind Ready to Reimagine Drug Discovery and Defeat Diseases
At Google I/O, DeepMind's CEO announced a goal to reimagine drug discovery with AI—but experts recommend staying realistic.

OpenAI invests $234 million in new AI lab in Singapore

RAG systems break on real data: the culprit is retrieval, not the model

NVIDIA developed a skills verification system for managing AI agents

Alibaba Introduces Its Own AI Chip for Neural Network Training and Inference

Zo Computer Improved AI Agent Reliability 20x with Vercel
Zo Computer reduced AI request errors from 7.5% to 0.34% using Vercel AI Gateway and SDK. Chat session success rate grew to 99.93%, latency

Google Unveils Gemini Spark — An AI Agent Working for Users 24/7
Google unveiled Gemini Spark at I/O 2026 — a personal AI agent based on Gemini 3.5. It operates in the background around the clock, position

Alibaba unveils Zhenwu M890 — an accelerator optimized for AI agents

Vercel added Zero Data Retention to protect data from AI model training

Codex and GitLab: From Code to Production in Three Steps

How Ramp Engineers Accelerate Code Review with GPT-5.5 and Codex

🎧 Weekly Summary · 2026-W21

🎧 Language Models: The Week in Review

Sigma Computing doubled its valuation to $3 billion in Series E round amid agentic analytics race




























