Top now
6 storiesWhat is climbing fastest across every source
Why Greenpeace Interrupted Jensen Huang’s Keynote
From Data Center Knowledge.
IFM/PAN2
IFM/PAN2 is an image-text-to-video diffuser model featuring 29 billion parameters available for download and exploration on Hugging Face. ·
LVMH’s Media Division announces strategic partnership with OpenAI
LVMH Media Division partners with OpenAI to support journalism with AI and improve publication accessibility via ChatGPT. ·
"Who is Hiring?" postings stuck near 12-year lows as AI reaches half of all jobs
Hacker News reports that job postings remain near twelve-year lows while artificial intelligence appears in more than half of all listings. ·
"Who is Hiring?" postings stuck near 12-year lows as AI reaches half of all jobs
Hacker News reports that job postings remain near twelve-year lows while artificial intelligence appears in more than half of all listings. ·
68 points and 36 comments on Hacker News, linking hntrends.com.
Envisioning Agentic Robotics: A New Era of Robotic Systems
GPT-6 Astra introduces agentic robotics, where foundation models control robots in closed loops to perform open-world tasks. ·
IFM/PAN2
IFM/PAN2 is an image-text-to-video diffuser model featuring 29 billion parameters available for download and exploration on Hugging Face. ·
Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game Agent
Gamer is a generalist vision-language-action foundation model for long-horizon real-time game agents that combines perception, memory, and reasoning. ·
What's Next for Nvidia?
Nvidia board announces Jensen Huang resigned 33 years after co-founding the company, raising questions about future innovation. ·
Understanding Positional Embeddings in Modern LLMs
The item explains methods behind how modern large language models represent token positions through positional embeddings. ·
232 points and 27 comments on Hacker News, linking magazine.sebastianraschka.com.
Envisioning Agentic Robotics: A New Era of Robotic Systems
GPT-6 Astra introduces agentic robotics, where foundation models control robots in closed loops to perform open-world tasks. ·
VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender
VibePrevisBench evaluates agentic coding for 4D scene generation in Blender using 242 tasks and two evaluation dimensions. ·
What's Next for Nvidia?
Nvidia board announces Jensen Huang resigned 33 years after co-founding the company, raising questions about future innovation. ·
YellowLaborerMarket/Gemini-Config-Checker-
YellowLaborerMarket releases Gemini-Config-Checker, a chain region scanner that finds proxy configurations for accessing Google AI Studio and Gemini across multiple operating systems. ·
WyrmClimberManacle/Gemini-Config-Checker-
WyrmClimberManacle releases Gemini-Config-Checker, a chain region scanner that finds proxy configurations for accessing Google AI Studio and Gemini across multiple operating systems. ·
Trending topics today: heat
Table
| NVIDIA | 1.27 |
| Qwen 4 Max | 0.85 |
| Jensen Huang | 0.77 |
| OpenAI | 0.43 |
| Agents | 0.39 |
| Context | 0 |
Heat = summed hotness of today's stories naming the topic.
Models
5 storiesReleases, papers, trending weights and new listings
Top 6 trending on Hugging Face: likes
Table
| Qwen/Qwen3.8-27B | 20186 |
| Lightricks/LTX-2.5 | 12143 |
| xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic | 7640 |
| MiniMaxAI/MiniMax-H3-Image | 1244 |
| Comfy-Org/MiniMax-H3-Image | 439 |
| Zyphra/ZONOS3-Mini ▲1 | 426 |
PirateBench: score out of 100
Table
| gpt-6.1-sol | 98.9 |
| opus-5-5 | 96.46 |
| gpt-6-sol | 91.6 |
| gpt-6-luna | 65.25 |
| haiku-4-5 | 4.28 |
Agents build our own projects from a brief; every run, failures included: https://cowerx.dev/daily/piratebench/
Envisioning Agentic Robotics: A New Era of Robotic Systems
GPT-6 Astra introduces agentic robotics, where foundation models control robots in closed loops to perform open-world tasks. ·
83 upvotes on Hugging Face Daily Papers.
Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game Agent
Gamer is a generalist vision-language-action foundation model for long-horizon real-time game agents that combines perception, memory, and reasoning. ·
AI Reviews Are Biased Towards the Appendix
GPT and Claude models derive over half their review weaknesses from paper appendices, while humans cite only 8.4% from that section. ·
VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender
VibePrevisBench evaluates agentic coding for 4D scene generation in Blender using 242 tasks and two evaluation dimensions. ·
G-CAST: Governed Continual Self-Training for E-Commerce Agents under Evolving Demands
Researchers present G-CAST, a governed continual self-training system that turns daily production evidence into training data for e-commerce shopping agents as user demands and agent tools change. ·
Envisioning Agentic Robotics: A New Era of Robotic Systems
GPT-6 Astra introduces agentic robotics, where foundation models control robots in closed loops to perform open-world tasks. ·
83 upvotes on Hugging Face Daily Papers.
VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender
VibePrevisBench evaluates agentic coding for 4D scene generation in Blender using 242 tasks and two evaluation dimensions. ·
Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game Agent
Gamer is a generalist vision-language-action foundation model for long-horizon real-time game agents that combines perception, memory, and reasoning. ·
MMFlash: Faster Speculative Decoding for Vision-Language Models
MMFlash accelerates vision-language model generation by upweighting visually dependent tokens and using sparse attention for long videos. ·
V-JEPA-LM: Video-Language Models That Read Dynamics Beyond Static, Text-Aligned Representations
V-JEPA-LM models use counterfactual supervision to enable language models to read video dynamics from frozen encoders trained on 4.53 million samples. ·
X voices
5 storiesX posts other sources are linking to, then named voices on Bluesky and Mastodon
Running a 6M parameter model on a Game Boy
2 points and 0 comments on Hacker News, linking an X post.
I'm Glad That Literary Awards Are Going to AI
2 points and 1 comments on Hacker News, linking an X post.
First pre-2020 long-form writing detected as 100% AI by Pangram
5 points and 0 comments on Hacker News, linking an X post.
I run multiple Claude Code accounts on one Mac
8 points and 3 comments on Hacker News, linking an X post.
DeepSeek releases API for its apps along with open source front end
2 points and 0 comments on Hacker News, linking an X post.
I run multiple Claude Code accounts on one Mac
8 points and 3 comments on Hacker News, linking an X post.
First pre-2020 long-form writing detected as 100% AI by Pangram
5 points and 0 comments on Hacker News, linking an X post.
I'm Glad That Literary Awards Are Going to AI
2 points and 1 comments on Hacker News, linking an X post.
Running a 6M parameter model on a Game Boy
2 points and 0 comments on Hacker News, linking an X post.
The AI revolution is leaving younger tech workers behind
3 points and 0 comments on Hacker News, linking an X post.
GitHub
5 storiesNew repos, releases and changelog
New AI repos this week: stars
Table
| hitkeeen/blender-ai-mcp-addon | 1922 |
| FrigadeHQ/yogurt | 1763 |
| HybridRobotics/supermimic | 1641 |
| Noniv/abyssal-garden | 1361 |
| duwqijlk/muse-protocol | 1320 |
| ZiChuanLan/freebie-radar | 1008 |
Show HN: db-branch – database branches in seconds for AI agents and CI, via ZFS
db-branch creates database branches in seconds for AI agents and CI using ZFS clones and ProxySQL routing. ·
24 points and 6 comments on Hacker News, linking github.com.
Show HN: Turbo-Browser-MCP – Batch forms, DOM deltas and submit guards for MCP
Tthanh2k6 releases Turbo-Browser-MCP, a high-speed Playwright server for MCP that supports multi-tab batching, DOM deltas, and form submission guards. ·
langchain-mistralai==1.2.0
LangChain releases version 1.2.0 of its Mistral AI integration, which fixes streaming behavior for reasoning models and updates dependencies. ·
langchain-xai==1.4.0
LangChain releases version 1.4.0 of its xAI integration package, which routes ChatXAI through the LangSmith gateway and updates several dependencies. ·
n4ze3m/tiny-voca
A new repository written in Python with 501 stars in its first week.
YellowLaborerMarket/Gemini-Config-Checker-
YellowLaborerMarket releases Gemini-Config-Checker, a chain region scanner that finds proxy configurations for accessing Google AI Studio and Gemini across multiple operating systems. ·
A new repository written in Makefile with 474 stars in its first week.
WyrmClimberManacle/Gemini-Config-Checker-
WyrmClimberManacle releases Gemini-Config-Checker, a chain region scanner that finds proxy configurations for accessing Google AI Studio and Gemini across multiple operating systems. ·
wisemeadowlarkspike/Gemini-Config-Checker-
wisemeadowlarkspike releases Gemini-Config-Checker, a chain region scanner that finds proxy configurations for Google AI Studio and Gemini across multiple operating systems. ·
WireProvostPucker/Gemini-Config-Checker-
WireProvostPucker releases Gemini-Config-Checker, a chain region scanner that finds proxy configurations for Google AI Studio and Gemini across multiple operating systems. ·
WindowStaffGate11/Gemini-Config-Checker-
WindowStaffGate11 releases Gemini-Config-Checker, a chain region scanner that finds proxy configurations for accessing Google AI Studio and Gemini across multiple operating systems. ·
Infrastructure
5 storiesData centers, power, chips and what tokens cost
Data-center and energy stories per day
Table
| Tue 10 | 32 |
| Wed 11 | 31 |
| Thu 12 | 29 |
| Fri 13 | 29 |
| Sat 14 | 7 |
| Sun 15 | 1 |
| Mon 16 | 14 |
Stories from DCD, Data Center Knowledge, EIA and Hacker News, by publication day.
Median output price of models added in the last 30 days, $ per million tokens
One point per OpenRouter reading, last 8 days. Data by OpenRouter, CC BY 4.0.
Output price, $ per million tokens (newest models)
Table
| qwen4-max | 4.5 |
| gemini-4-hydrogen | 7.5 |
| gemini-4-hydrogen:batch | 3.75 |
| ernie-light-dt-1.0-intl | 3.94 |
| gemini-4-carbon | 10 |
| median, all paid models | 2.08 |
441 paid models listed; median of the 27 added in 30 days: $3.75. Rankings data by OpenRouter, CC BY 4.0.
Engie could supply power to planned Google data center in Châteauroux, France
Engie may supply power to a planned Google data center in Châteauroux, France, under a potential deal including renewable energy investment. ·
From DCD.
Mount Thor raises $45m to fund new AI cloud offering Apple chips and software
Mount Thor raises $45 million to launch an Apple-based cloud platform featuring Apple chips and software. ·
Nvidia deliberately kept GPU prices high through circular financing deals
Nvidia kept GPU prices high through circular financing deals that helped finance its own customers. ·
447 points and 236 comments on Hacker News, linking techspot.com.
Nvidia allegedly engineered scarcity by funding AI startups to buy its GPUs
Nvidia allegedly engineered scarcity by funding AI startups to buy its GPUs. The Wall Street Journal reports this strategy inflated the ongoing GPU and memory shortage. ·
Will we put data centers in space?
Forethought analyzes launch costs, cooling physics, and engineering factors to determine how soon AI data centers could move to orbit. ·
Engie could supply power to planned Google data center in Châteauroux, France
Engie may supply power to a planned Google data center in Châteauroux, France, under a potential deal including renewable energy investment. ·