Perplexity
Perplexity is an AI answer engine that combines live web search with LLM-generated, cited answers — a direct challenge to Google Search. Its Comet browser and aggressive growth made it one of the fastest-rising AI startups. All our Perplexity coverage lives on this page.
TG: @tekhnocafe (Технокафе / Вайбкодинг / AI-агенты)Claude SEO Plugin Runs Up to 17 Agents and Audits a Site with One Command
Habr AIIn 94% of B2B shortlists, AI summaries barely show the companies' own websites
Latent Space (swyx)Perplexity hits $1 billion in annualized revenue: 4x growth in 2026
Habr AIHow LLM Applications Get Broken: From the "Grandma Exploit" to LLM-vs-LLM Attacks
The DecoderGoogle Opens Gemini 4 Carbon: $7.76 per Solved Task and 0% Breaches in an Injection Test
Habr AIGoogle Opens Gemini 4 Carbon: An Agentic Model 2–6x Cheaper Than Fable 5.2
Habr AIHow to find ChatGPT traffic in Metrica and GA4: up to 6,3% may be hiding in direct visits
TG: @ai_machinelearning_big_data (Machinelearning)ChatGPT Took 81.4% of Traffic Across Six AI Chatbots in Graphite Study
Google Opens Gemini 4 Carbon: $7.76 per Solved Task and 0% Breaches in an Injection Test
Google has made Gemini 4 Carbon generally available: it handles complex code and debugging better than Argon, but burns more tokens. It also ships a /swarm mode in…

Google Opens Gemini 4 Carbon: An Agentic Model 2–6x Cheaper Than Fable 5.2
Google has released Carbon, the flagship Gemini 4 model for long-running agentic tasks: close to the leader on quality, but a task costs several times less. We look…

How to find ChatGPT traffic in Metrica and GA4: up to 6,3% may be hiding in direct visits
Clicks from ChatGPT answers often appear in Metrica as visits with no source. Here’s how to build a referrer-and-UTM segment and estimate hidden AI traffic without…

Perplexity puts Teammates in charge of 5 workflows without a new prompt

Perplexity launches Teammates: AI agents that work around the clock for $200

Claude Pulls Its Links from Brave: 89 out of 100 Citations Sit in the Top 20 of That Search Engine

SIBUR added an AI critic to its multi-agent system and cut idea evaluation from 88 days to two weeks

GPT‑6 Cyber Wrote Exploits for 948 Linux Kernel Bugs and Stayed in Scope in 99.93% of Tests
OpenAI's GPT‑6 Cyber, an AI agent for vulnerability hunting, solves 48.5% of ExploitGym tasks versus 32.1% for its predecessor, and rarely strays outside its…

Near-Zero Token Embeddings Wiped Out 20% of Fine-Tuning Steps in Gated Models on Telugu
An audit of 291 open repositories: 52.6% of untied gated checkpoints have tokens with near-zero embeddings, which silently zero out batch gradients on roughly 20%…

TJ Digital: 23% of AI-cited pages open with a direct answer, versus 14% of uncited ones

Claude in Houdini on Windows: Two Bugs in the Agent Panel and a 31-Node Geode

NVIDIA Releases 120B Nemotron 3.5 Super VL and Bolt Models for Local Agents

Modal Raises $750M at a $15.75B Valuation and Reaches $1B in ARR

Lovable builds GPT-Live, Claude and Gemini Omni into its chat: 9 AI tools in one place
Lovable has added voice-based GPT-Live, Claude models, Gemini 4 Argon, Omni and Lyria, plus Perplexity, Firecrawl, Parallel and Jev. Everything is available…

DecisionPivot Tests 40 Solver Models: Up to a 26-Point Gap at Equal Accuracy
DecisionPivot's 40,033-item benchmark shows two 27B models 2.6 points apart on the reference track diverge by 25.9 points on contrastive groups.









