Llama.cpp on a jailbroken PS5: run any GGUF model natively on the console
8 points and 0 comments on Hacker News, linking github.com.
Top now
6 storiesWhat is climbing fastest across every source
Anthropic4.6xChinese7.1xnvidia6.7xA42B7.1xOpen-816B7.1xmeta-models7.1xClaude6.9xMuse Spark6.0xGemini 4 Argon6.1xWorld6.7x
Designing more efficient quantum circuits with Claude
19 points and 4 comments on Hacker News, linking anthropic.com.
nvidia/Qwen3.8-Flash-Next-FP4-Local
A image text to text model from nvidia with 128 likes and 771 downloads on Hugging Face.
[Industry Event] Advancing Data Center Construction East 2026
From Data Center Knowledge.
Grounding AI agents in biological reality: Building blocks for autonomous wet labs
From Google Research.
Closing the Skills Gap in Data Center Facilities Management
Data centers require skilled professionals, and collaboration between IT and facilities teams, along with proactive maintenance and remote monitoring, helps address workforce shortages. ·
90% of US working hours are spent on tasks not covered well by AI benchmarks
A new map shows that 90% of US working hours involve tasks poorly covered by current AI benchmarks, highlighting significant gaps in evaluation coverage. ·
175 points and 141 comments on Hacker News, linking mercor.com.
Designing more efficient quantum circuits with Claude
19 points and 4 comments on Hacker News, linking anthropic.com.
nvidia/Qwen3.8-Flash-Next-FP4-Local
A image text to text model from nvidia with 128 likes and 771 downloads on Hugging Face.
meta-models/Muse-Spark-Open-816B-A42B-NVFP4
A image text to text model from meta-models with 145 likes and 814 downloads on Hugging Face.
MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments
MAI-Gym is an open platform for evaluating and training GUI agents across mobile and computer environments. It supports 20,000 parallel instances and trains a 9B agent. ·
Waypoint-OPD: Transferring RL-Induced Policy Shifts via Gated On-Policy Distillation
Researchers introduce Waypoint-OPD, a distillation method that gates policy shifts using intermediate checkpoints to improve transfer between Qwen3 model sizes in reasoning tasks. ·
90% of US working hours are spent on tasks not covered well by AI benchmarks
A new map shows that 90% of US working hours involve tasks poorly covered by current AI benchmarks, highlighting significant gaps in evaluation coverage. ·
175 points and 141 comments on Hacker News, linking mercor.com.
MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments
MAI-Gym is an open platform for evaluating and training GUI agents across mobile and computer environments. It supports 20,000 parallel instances and trains a 9B agent. ·
Postgres 19 delayed after AI reviewers find more bugs than humans can fix
121 points and 38 comments on Hacker News, linking theregister.com.
Waypoint-OPD: Transferring RL-Induced Policy Shifts via Gated On-Policy Distillation
Researchers introduce Waypoint-OPD, a distillation method that gates policy shifts using intermediate checkpoints to improve transfer between Qwen3 model sizes in reasoning tasks. ·
anthropics/claude-code v2.1.301
New release v2.1.301 of anthropics/claude-code on GitHub.
anujasid/mcp-for-blender-pro
Anujasid releases an AI assistant and MCP server for Blender that supports visual chat, scene tools, and multi-agent workflows. It integrates with OpenAI, Claude, Gemini, Ollama, and OpenCode. ·
Trending topics today: heat
Table
| Anthropic | 2.11 |
| Chinese | 1.44 |
| nvidia | 1.13 |
| A42B | 0.91 |
| Open-816B | 0.91 |
| meta-models | 0.91 |
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
| abenzerps/Qwen-Image-2.1-Uncensored-Quantized ▼1 | 4424 |
| google/embeddinggemma-2 | 2496 |
| jialinyyzz/humanizer | 1813 |
| Qwen/Qwen-Image-2.1-Turbo ▼1 | 931 |
| openbmb/MiniCPM-V-4.7-35B-A3B ▲1 | 835 |
| fishaudio/drama-3 ▲2 | 539 |
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/
Show HN: Gemini 4 Argon plays StarCraft
127 points and 65 comments on Hacker News, linking starskirmish.com.
meta-models/Muse-Spark-Open-816B-A42B-NVFP4
A image text to text model from meta-models with 145 likes and 814 downloads on Hugging Face.
MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments
MAI-Gym is an open platform for evaluating and training GUI agents across mobile and computer environments. It supports 20,000 parallel instances and trains a 9B agent. ·
Waypoint-OPD: Transferring RL-Induced Policy Shifts via Gated On-Policy Distillation
Researchers introduce Waypoint-OPD, a distillation method that gates policy shifts using intermediate checkpoints to improve transfer between Qwen3 model sizes in reasoning tasks. ·
OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understanding
Researchers introduce OmniCompass, a benchmark for evaluating multimodal models on video grounding and understanding. It includes 820 vision-language videos and 240 audio-visual videos. ·
MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments
MAI-Gym is an open platform for evaluating and training GUI agents across mobile and computer environments. It supports 20,000 parallel instances and trains a 9B agent. ·
92 upvotes on Hugging Face Daily Papers.
Show HN: Gemini 4 Argon plays StarCraft
127 points and 65 comments on Hacker News, linking starskirmish.com.
Waypoint-OPD: Transferring RL-Induced Policy Shifts via Gated On-Policy Distillation
Researchers introduce Waypoint-OPD, a distillation method that gates policy shifts using intermediate checkpoints to improve transfer between Qwen3 model sizes in reasoning tasks. ·
OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understanding
Researchers introduce OmniCompass, a benchmark for evaluating multimodal models on video grounding and understanding. It includes 820 vision-language videos and 240 audio-visual videos. ·
FDE-Gym: The Last Mile of AI Delivery
FDE-Gym is a benchmark of 95 delivery tasks from 20 customer scenarios across 18 sectors, where agents interact with stakeholders to improve customer success metrics. ·
X voices
5 storiesX posts other sources are linking to, then named voices on Bluesky and Mastodon
@SemiAnalysis_
Nvidia Barred Neoclouds Running AMD Chips from GTC Berlin
1 points and 0 comments on Hacker News, linking an X post.
@OpenAI
GPT-Live-1.1 – New voices. Better accents. Less "checking."
2 points and 0 comments on Hacker News, linking an X post.
@SEGA
Total War Remaster Removed from Steam After AI Allegations, Brings Back Original
4 points and 0 comments on Hacker News, linking an X post.
@HelloSurgeAI
Can Coding Agents Act Like Staff Engineers?
2 points and 1 comments on Hacker News, linking an X post.
@testingcatalog
Claude Fable 5.5 traces have been spotted in Azure configs
1 points and 0 comments on Hacker News, linking an X post.
@SemiAnalysis_
Nvidia Barred Neoclouds Running AMD Chips from GTC Berlin
1 points and 0 comments on Hacker News, linking an X post.
@OpenAI
GPT-Live-1.1 – New voices. Better accents. Less "checking."
2 points and 0 comments on Hacker News, linking an X post.
@SEGA
Total War Remaster Removed from Steam After AI Allegations, Brings Back Original
4 points and 0 comments on Hacker News, linking an X post.
@opencode
DeepSeek v4.1 Flash is now free in OpenCode
4 points and 0 comments on Hacker News, linking an X post.
@HelloSurgeAI
Can Coding Agents Act Like Staff Engineers?
2 points and 1 comments on Hacker News, linking an X post.
GitHub
5 storiesNew repos, releases and changelog
New AI repos this week: stars
Table
| onehumanbeing/esp32-iphone-agent | 2832 |
| deeplethe/solid-glass | 1347 |
| rociiu/talorys | 1242 |
| marclou/mailcheap | 1128 |
| zhuyansen/new-term-radar | 1063 |
| feder-cr/git-ai | 1044 |
OpenMOSS/MOSS-Image
OpenMOSS releases MOSS-Image, a 5B-parameter base DiT model that generates high-quality text and layouts while supporting efficient fine-tuning. ·
A new repository written in Python with 188 stars in its first week.
anujasid/mcp-for-blender-pro
Anujasid releases an AI assistant and MCP server for Blender that supports visual chat, scene tools, and multi-agent workflows. It integrates with OpenAI, Claude, Gemini, Ollama, and OpenCode. ·
PointWormGazebo/panel-all-in-one-any-game
PointWormGazebo offers a universal cheat panel that provides AI-enhanced aimbot and per-game precision tuning for various video games. ·
openai/codex 0.163.0
OpenAI releases Codex version 0.163.0, which adds parallel read-only tool execution and enables Daybreak by default for eligible accounts. ·
anthropics/claude-code v2.1.301
New release v2.1.301 of anthropics/claude-code on GitHub.
anthropics/claude-code v2.1.301
New release v2.1.301 of anthropics/claude-code on GitHub.
anujasid/mcp-for-blender-pro
Anujasid releases an AI assistant and MCP server for Blender that supports visual chat, scene tools, and multi-agent workflows. It integrates with OpenAI, Claude, Gemini, Ollama, and OpenCode. ·
OpenMOSS/MOSS-Image
OpenMOSS releases MOSS-Image, a 5B-parameter base DiT model that generates high-quality text and layouts while supporting efficient fine-tuning. ·
openai/codex 0.163.0
OpenAI releases Codex version 0.163.0, which adds parallel read-only tool execution and enables Daybreak by default for eligible accounts. ·
ggml-org/llama.cpp v0.7.0
llama.cpp v0.7.0 adds LiliCorr speculative decoding, a GPU cache for host-memory experts, and support for models like K2 Horizon and Prism Bonsai 2. ·
Infrastructure
5 storiesData centers, power, chips and what tokens cost
Data-center and energy stories per day
Table
| Sat 10 | 3 |
| Sun 11 | 7 |
| Mon 12 | 41 |
| Tue 13 | 29 |
| Wed 14 | 34 |
| Thu 15 | 41 |
| Fri 16 | 33 |
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
| beam-1.0-501b-a23b | 1 |
| moondream3-preview | 2.5 |
| moondream3.1-9b-a2b | 2.5 |
| muse-spark-open | 4.25 |
| nemotron-3.5-super-vl | 0.5 |
| median, all paid models | 2 |
439 paid models listed; median of the 48 added in 30 days: $4.62. Rankings data by OpenRouter, CC BY 4.0.
A story of WiFi secrets, UART lockdowns and a damaged flash memory chip
Researchers reverse engineer the Eufy Homebase2, revealing its WiFi secrets, UART lockdowns, and a damaged flash memory chip. ·
14 points and 1 comments on Hacker News, linking adepts.of0x.cc.
DCN Warsaw 2026: Poland deserves a greater share of global data center capacity
Poland is positioned to gain a larger share of global data center capacity due to its evolving energy situation and the ongoing industry expansion. ·
Google's DeepMind hiring for nuclear expert
Google DeepMind is hiring a nuclear expert to evaluate and mitigate the safety risks of frontier AI models in the nuclear domain. ·
Market selloff, rising rates test AI data centre deal boom
49 points and 40 comments on Hacker News, linking reuters.com.
A story of WiFi secrets, UART lockdowns and a damaged flash memory chip
Researchers reverse engineer the Eufy Homebase2, revealing its WiFi secrets, UART lockdowns, and a damaged flash memory chip. ·
Banks are now trying to offload the AI data center debt they just underwrote
382 points and 277 comments on Hacker News, linking startupfortune.com.
Nvidia's first GPU: GeForce 256
60 points and 39 comments on Hacker News, linking dfarq.homeip.net.
DCN Warsaw 2026: Poland deserves a greater share of global data center capacity
Poland is positioned to gain a larger share of global data center capacity due to its evolving energy situation and the ongoing industry expansion. ·