2026
11,340 articles · 139 days · 10 categories · avg 82 per day
Meta Developed a Secret Model Called "Avocado" — What It Says About the Company's Strategy
Meta's unfinished AI project, codenamed "Avocado," shows how tech giants decide what to ship to market and what to keep in the lab for good. *Meta is designated an…
T-Technologies Group on open source in AI/ML: how LLM work is organized inside the company
An interview with T-Technologies Group's AI/ML team on LLM development, open source participation and the company's research track.
At GTC 2026, NVIDIA shifts focus from chips to token factories and the Agent-as-a-Service era
At GTC 2026, NVIDIA signaled it is betting not on individual GPUs but on token factories, the modular Vera Rubin architecture and an AI-agents-as-a-service economy.
AI-first startups: why growth marketing stalls and what breaks in the funnel
US Tech Companies Accelerate Layoffs as AI Investment Grows
Rocket Close cuts mortgage document processing time by 15x with AWS
Anthropic CEO Dario Amodei promises "good AI," but critics call for slowing the race
Dan Pratl: Quadron Pushes a 'Trust Economy' for Valuing Expertise in the AI Era
Quadron founder Dan Pratl argues that as generative AI spreads, the scarce resource is no longer knowledge but verifiable expertise, judgment and trust in decisions.
Fortis Solutions bets on human-supervised AI and trust infrastructure
Fortis Solutions argues that businesses don't need autonomous AI for its own sake, but systems where machine precision amplifies people and trust rests on solid…
Habr: AI agents are reshaping delivery, forcing teams to rebuild the entire development cycle
M2 handed 40% of its marketing copy to AI without losing content quality
Sova AI releases an Android assistant that controls your phone without a PC or root
Microsoft wants its own frontier AI models on par with OpenAI and Anthropic by 2027
Indian AI startup Sarvam is raising up to $350 million at a $1.5 billion valuation
Sarvam AI is close to a $300-350 million round at a $1.5 billion valuation, one of the strongest signals yet that India wants to grow its own AI champion.
Google adds Flex and Priority modes to the Gemini API to balance cost and reliability
Google has launched two new Gemini API service tiers: Flex for cheaper background jobs, and Priority for critical traffic, with higher reliability and a fallback to…
OpenAI buys tech show TBPN to gain more influence over the public conversation about AI
OpenAI is moving into media, buying Silicon Valley favorite TBPN and promising to keep the show editorially independent while using it for open talk about AI.
Microsoft reworks Copilot sales after pressure from Wall Street analysts
SpaceX Prepares a Record IPO as OpenAI and Anthropic Edge Toward Going Public
Microsoft to invest $10 billion in Japan over four years as AI demand surges across Asia
Google Gemma 4, NVIDIA and OpenClaw: Local AI Agents With No Per-Token Bill
Yandex shows businesses how to get into Alice's answers and measure search visibility
Yandex has added a "Site visibility in Alice" section to Webmaster. For businesses, the signal is clear: strong SEO, structure and accurate data get you into the…
Flant: How a Go Developer Turned Zed and Gemini into a Useful AI Agent
A Go developer at Flant traces his path from slow IDE plugins to a stack of Zed, Gemini 3 Flash and gopls-mcp that gives an AI agent project context and real value…
OpenAI's revenue and user growth slow as AI infrastructure costs climb
Google employees demand Pichai keep the company's AI models away from the Pentagon
OpenAI misses internal targets for ChatGPT audience growth and revenue
Talkie-1930: Researchers Release a 13B Model With No Knowledge of the Internet or World War II
Agentis Memory: A Redis-Compatible Store With Vector Search and Local Embeddings
Agentis Memory turns a Redis-compatible store into shared memory for AI agents, with local embeddings, built-in vector search and no external APIs.
OpenAI Buys Tech Show TBPN, Its First Media Deal
OpenAI has bought TBPN, a daily tech show popular in Silicon Valley, promised to preserve its editorial independence and placed it inside the company's strategy…
Habr: How synthetic data helps train models, and why self-training leads to collapse
Synthetic data helps AI offset the shortage of high-quality human data, but uncontrolled self-training makes models lose rare cases and diversity.
Why AI in UI design matters as a source of visual mutations, not as a production tool
NVIDIA shows a full model-optimization pipeline with FastNAS pruning and fine-tuning
TII releases Falcon Perception, a 0.6B model for text-prompted object segmentation and detection
