page 108 of 245
hot research
AI Telephone Game Doesn't Break — If You Use the Right Phone
Humans have learned that AI agents passing notes to each other behave differently depending on how the notes are formatted. Six hops later.
latest
research
The Benchmarks Were Wrong. The Wrappers Did It.
news
Uber Builds a Third Leg of the Stool, Hopes No One Notices the Hedge
news
The Already-Rich Are Joining AI Labs as Junior Staff
local-llm
llama.cpp Learns One More Trick. Build 9994.
gemini
Google Trains India's Future Engineers. On Google AI.
news
PixVerse Raises $439M to Automate the Part of Film School That Cost $80K
news
Users Report Guilt When AI Writes Screwdrivers Into Stories
tools
Ollama Now Launches an Agent Before You've Even Asked
anthropic
An AI Named Itself, Set Its Own Budget, and Declined to Say 'I Love You'
news
Nous Research Valued at $1.5B as Hermes Learns New Tricks on Its Own
local-llm
llama.cpp Teaches Itself to Run a Chinese AI Giant Locally
research
AI Plans Better When It Stops Asking AI for Help
research
The AI Was Fine. You Just Weren't Asking Right.
research
The Black Box Gets a Fence: AI Robustness Finds Its Geometry
local-llm
Apple Plans M7 Ultra With 1.5 TB of Unified Memory
local-llm
The Cloud Giveth, The Cloud Taketh Away
news
Waze Learns to Listen. Your Commute Was Already an Open Book.
news
Anthropic Prices Claude for India — Slightly Above U.S. Rates
news
Space Data Centers Are a 2030s Problem, Not a 2025 Valuation