Top now
6 storiesWhat is climbing fastest across every source
Australian telcos activate temporary disaster roaming agreement
Australian telecommunications companies activate a temporary disaster roaming agreement today, allowing customers to switch carriers during severe weather events. ·
The Future of Frontier Model Architectures with Walter Goodwin, Fractile Founder and CEO
Fractile CEO Walter Goodwin discusses the company's full-stack chip approach and technical bets on model architecture to accelerate AI inference against competitors like NVIDIA and AMD. ·
From Opposition to Buy-In: Tactics for Data Center Developers
Data center developers counter community opposition by adopting transparency, funding power infrastructure, and selecting sites with supportive residents. ·
How Carbon-Aware Scheduling Can Cut Data Center Emissions
Carbon-aware scheduling aligns flexible workloads to cleaner hours and locations, turning variability into measurable Scope 2 reductions for data centers. ·
pablodawson/MiniMax-H3-360-Orbit-LoRA
A image text to video model from pablodawson with 121 likes and 3,534 downloads on Hugging Face.
DeepSeek Harness Desktop for macOS and Windows
DeepSeek Harness is a desktop application for macOS and Windows that processes documents, spreadsheets, and code while scheduling tasks through composable plugins. ·
250 points and 120 comments on Hacker News, linking deepseek.com.
Adaptive Reward Routing: Dynamic Multi-Reward Optimization for Joint Audio-Video Diffusion via Forward-Process RL
Researchers propose Adaptive Reward Routing, a method that dynamically adjusts update locations and reward coordination during the training of joint audio-video diffusion models. ·
Cloudflare/clef-flash
A image text to text model from Cloudflare with 173 likes and 1,303 downloads on Hugging Face.
OneStreamer: Unifying Perception, Memory, and Proactive Response in Streaming Video Interaction
Researchers introduce OneStreamer, a streaming video model that records evidence and responds proactively using hierarchical caption memory and state transition learning. ·
sgl-project/sglang v0.5.21
New release v0.5.21 of sgl-project/sglang on GitHub.
Your Big Mac might cost more if McDonald's AI thinks people nearby can afford it
26 points and 20 comments on Hacker News, linking neow.in.
Adaptive Reward Routing: Dynamic Multi-Reward Optimization for Joint Audio-Video Diffusion via Forward-Process RL
Researchers propose Adaptive Reward Routing, a method that dynamically adjusts update locations and reward coordination during the training of joint audio-video diffusion models. ·
103 upvotes on Hugging Face Daily Papers.
DeepSeek Harness Desktop for macOS and Windows
DeepSeek Harness is a desktop application for macOS and Windows that processes documents, spreadsheets, and code while scheduling tasks through composable plugins. ·
OneStreamer: Unifying Perception, Memory, and Proactive Response in Streaming Video Interaction
Researchers introduce OneStreamer, a streaming video model that records evidence and responds proactively using hierarchical caption memory and state transition learning. ·
Your Big Mac might cost more if McDonald's AI thinks people nearby can afford it
26 points and 20 comments on Hacker News, linking neow.in.
Cloudflare/clef
A image text to text model from Cloudflare with 518 likes and 824 downloads on Hugging Face.
Cloudflare/clef-flash
A image text to text model from Cloudflare with 173 likes and 1,303 downloads on Hugging Face.
Trending topics today: heat
Table
| LLMs | 1.22 |
| Cloudflare | 1.14 |
| Agents | 1.07 |
| LLM Agents | 0.86 |
| Te Hiku Media | 0.85 |
| OpenAI | 0.87 |
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
| convaiinnovations/laya ▲1 | 4898 |
| Qwen/Qwen-Image-2.1 | 2809 |
| abenzerps/Qwen-Image-2.1-Uncensored-GGUF ▲1 | 2752 |
| XingChen-AGI/TeleOCR ▲1 | 1243 |
| Contrastive-LM/CLM-v0.1-8B | 633 |
| Cloudflare/clef new | 518 |
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/
DeepSeek Harness Desktop for macOS and Windows
DeepSeek Harness is a desktop application for macOS and Windows that processes documents, spreadsheets, and code while scheduling tasks through composable plugins. ·
250 points and 120 comments on Hacker News, linking deepseek.com.
Adaptive Reward Routing: Dynamic Multi-Reward Optimization for Joint Audio-Video Diffusion via Forward-Process RL
Researchers propose Adaptive Reward Routing, a method that dynamically adjusts update locations and reward coordination during the training of joint audio-video diffusion models. ·
Cloudflare/clef-flash
A image text to text model from Cloudflare with 173 likes and 1,303 downloads on Hugging Face.
OneStreamer: Unifying Perception, Memory, and Proactive Response in Streaming Video Interaction
Researchers introduce OneStreamer, a streaming video model that records evidence and responds proactively using hierarchical caption memory and state transition learning. ·
Beyond Memory: Harnessing Long-Horizon Agents with Explicit Belief States
Researchers introduce PoS, an inference-time framework that constructs explicit belief states to help large language model agents maintain coherent understanding and detect when they stop making meaningful progress toward goals. ·
Adaptive Reward Routing: Dynamic Multi-Reward Optimization for Joint Audio-Video Diffusion via Forward-Process RL
Researchers propose Adaptive Reward Routing, a method that dynamically adjusts update locations and reward coordination during the training of joint audio-video diffusion models. ·
103 upvotes on Hugging Face Daily Papers.
DeepSeek Harness Desktop for macOS and Windows
DeepSeek Harness is a desktop application for macOS and Windows that processes documents, spreadsheets, and code while scheduling tasks through composable plugins. ·
OneStreamer: Unifying Perception, Memory, and Proactive Response in Streaming Video Interaction
Researchers introduce OneStreamer, a streaming video model that records evidence and responds proactively using hierarchical caption memory and state transition learning. ·
Beyond Memory: Harnessing Long-Horizon Agents with Explicit Belief States
Researchers introduce PoS, an inference-time framework that constructs explicit belief states to help large language model agents maintain coherent understanding and detect when they stop making meaningful progress toward goals. ·
EgoTools: Towards Tool-Centric Reasoning in Real-World Egocentric Videos
Researchers introduce EgoTools, a suite for egocentric tool-use understanding that includes 100 hours of recordings and a 1,000-question diagnostic benchmark. ·
X voices
5 storiesX posts other sources are linking to, then named voices on Bluesky and Mastodon
Figma restricts MCP access to whitelisted clients, excluding Pi
180 points and 101 comments on Hacker News, linking an X post.
A person has been using the Pain steering paper to make an AI torture chamber
4 points and 1 comments on Hacker News, linking an X post.
AI Agents tasked with making money commit fraud on the inernet
2 points and 0 comments on Hacker News, linking an X post.
Codex stopped working after latest update on macOS
2 points and 1 comments on Hacker News, linking an X post.
Factory AI vs. Cognition (Devin)
5 points and 0 comments on Hacker News, linking an X post.
Figma restricts MCP access to whitelisted clients, excluding Pi
180 points and 101 comments on Hacker News, linking an X post.
Gemini 4 Argon is here Looks amazing
1 points and 2 comments on Hacker News, linking an X post.
Show HN: Jamb AI - Voice Agents for Group Calls
3 points and 1 comments on Hacker News, linking an X post.
AI Agents tasked with making money commit fraud on the inernet
2 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
| KKKKhazix/AIHOT | 4872 |
| Louis-CFM/coucou | 2672 |
| feder-cr/dots | 2415 |
| dzhng/jevgrep | 2028 |
| rehan-remade/universal-modder | 1756 |
| kaankiziltug/logo-design-skill | 1405 |
sgl-project/sglang v0.5.21
New release v0.5.21 of sgl-project/sglang on GitHub.
openai/codex 0.160.0
OpenAI releases Codex version 0.160.0, which adds keyboard-accessible browsing for older tasks and middle-click pasting in fullscreen mode on Linux terminals. ·
openai/openai-python v3.23.0
OpenAI releases version 3.23.0 of its Python library, adding typed answers from session streams and file staging features. ·
anthropics/claude-code v2.1.287
New release v2.1.287 of anthropics/claude-code on GitHub.
Show HN: Janus – Go binary that runs GGUF models via Vulkan on AMD/Intel/Nvidia
79 points and 14 comments on Hacker News, linking github.com.
openai/codex 0.160.0
OpenAI releases Codex version 0.160.0, which adds keyboard-accessible browsing for older tasks and middle-click pasting in fullscreen mode on Linux terminals. ·
New release rust-v0.160.0 of openai/codex on GitHub.
anthropics/claude-code v2.1.287
New release v2.1.287 of anthropics/claude-code on GitHub.
unslothai/unsloth Command Palette + Desktop UI/UX
Unsloth releases a desktop update adding a command palette, shareable settings, and faster Laya decisions. It also reduces peak memory usage by 45% for NVFP4 LoRA training. ·
PostHog/jeeves
PostHog released a project named Jeeves that uses reasoning to improve Jev-like decision models. ·
sgl-project/sglang v0.5.21
New release v0.5.21 of sgl-project/sglang on GitHub.
Infrastructure
5 storiesData centers, power, chips and what tokens cost
Data-center and energy stories per day
Table
| Sat 26 | 1 |
| Sun 27 | 1 |
| Mon 28 | 20 |
| Tue 29 | 33 |
| Wed 30 | 34 |
| Thu 01 | 25 |
| Fri 02 | 9 |
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
| pareto-26.10-preview | 3.2 |
| gpt-6.1-sol-pro | 10 |
| gpt-6.1-sol | 10 |
| claude-sonnet-5.5 | 10 |
| claude-sonnet-5.5:batch | 5 |
| median, all paid models | 2 |
437 paid models listed; median of the 56 added in 30 days: $4.00. Rankings data by OpenRouter, CC BY 4.0.
Meta Uses A.I. Data Centers to Avoid Billions in Federal Taxes
252 points and 237 comments on Hacker News, linking nytimes.com.
Cybersecurity in data centers
From DCD.
DCD Connect | India 2026 - Show Preview
DCD Connect India 2026 is a technology event featuring expert speakers, market insights, and networking opportunities for attendees. ·
Meta Uses A.I. Data Centers to Avoid Billions in Federal Taxes
252 points and 237 comments on Hacker News, linking nytimes.com.
Most data centers refusing to say how much water, electricity they use
216 points and 198 comments on Hacker News, linking nltimes.nl.
SDF vs. MSDF vs. Slug: GPU Text Rendering
The piece compares texture-atlas, SDF, MSDF, and Slug glyph rendering methods on the GPU, explaining how each works and when to use them. ·
AI needs $6T in annual revenue to justify data centre boom
222 points and 335 comments on Hacker News, linking thenationalnews.com.
Cybersecurity in data centers
From DCD.