PSSA: A non-transformer language model written from scratch in Rust
11 points and 0 comments on Hacker News, linking github.com.
The front page of the internet for AI users · Wed 30 Sep 2026, 04:00 UTC
What is climbing fastest across every source
11 points and 0 comments on Hacker News, linking github.com.
Researchers introduced a framework that groups visual observations of specific objects into biographies to improve long-video question answering. The system achieved 72.0% accuracy on the EgoLifeQA benchmark. ·
Researchers introduced MaLiang-Harness, a framework that uses multimodal models to iteratively build, check, and fix code for image and video generation. The system evaluates eleven closed-source models on new benchmarks to measure success and quality. ·
OpenAI released version 0.159.2 of Codex, a patch that stops console windows from flashing on Windows when the tool runs background processes. ·
New release v3.22.1 of openai/openai-python on GitHub.
From Gary Marcus.
232 points and 89 comments on Hacker News, linking github.com.
830 points and 757 comments on Hacker News, linking openai.com.
Researchers introduced MaLiang-Harness, a framework that uses multimodal models to iteratively build, check, and fix code for image and video generation. The system evaluates eleven closed-source models on new benchmarks to measure success and quality. ·
15 upvotes on Hugging Face Daily Papers.
OpenAI released version 0.159.2 of Codex, a patch that stops console windows from flashing on Windows when the tool runs background processes. ·
New release v2.1.285 of anthropics/claude-code on GitHub.
830 points and 757 comments on Hacker News, linking openai.com.
232 points and 89 comments on Hacker News, linking github.com.
The source text lists titles for visual guides and experiments on recurrent neural networks, convolutional neural networks, and transformers, focusing on their accuracy and efficiency. ·
Researchers introduced MaLiang-Harness, a framework that uses multimodal models to iteratively build, check, and fix code for image and video generation. The system evaluates eleven closed-source models on new benchmarks to measure success and quality. ·
JoasASantos maintains a repository of AI models that are uncensored or fine-tuned for cybersecurity tasks. ·
15 upvotes on Hugging Face Daily Papers.
| OpenAI | 4.18 |
| Anthropic | 2.87 |
| GLM-5.3 | 1.62 |
| Video Generation | 1.52 |
| GPT-6.1 | 1.21 |
| Muse | 1.17 |
Heat = summed hotness of today's stories naming the topic.
Releases, papers, trending weights and new listings
| convaiinnovations/laya | 4528 |
| Qwen/Qwen-Image-2.1 ▼1 | 2663 |
| abenzerps/Qwen-Image-2.1-Uncensored-GGUF ▼1 | 2452 |
| Edge0/Audio8-ASR-Infinite ▲1 | 1546 |
| XingChen-AGI/TeleOCR ▲1 | 879 |
| Contrastive-LM/CLM-v0.1-8B ▲1 | 531 |
| opus-5-5 | 91.15 |
| gpt-6-sol | 81.78 |
| deepseek-v4-flash-0731 | 62.82 |
Agents build our own projects from a brief; every run, failures included: https://cowerx.dev/daily/piratebench/
15 upvotes on Hugging Face Daily Papers.
Researchers introduced SAKI, a method routing token-level supervision in on-policy distillation using maximal coupling. It improves performance on mathematical reasoning benchmarks for small student models. ·
Researchers introduced AnisoWM, a method using learnable diagonal covariance to improve planning in latent world models. It outperformed LeWorldModel across four visual control environments. ·
Researchers found that latent world action models lose generalization benefits when they skip future generation during inference. They proposed Simple-WAM, which uses a single forward pass to retain these advantages. ·
Researchers introduced EasyPPO to stabilize critic training in PPO for large language models by filtering rollouts, normalizing noise, and adjusting batch sizes. ·
232 points and 89 comments on Hacker News, linking github.com.
15 upvotes on Hugging Face Daily Papers.
Researchers introduced SAKI, a method routing token-level supervision in on-policy distillation using maximal coupling. It improves performance on mathematical reasoning benchmarks for small student models. ·
Researchers introduced AnisoWM, a method using learnable diagonal covariance to improve planning in latent world models. It outperformed LeWorldModel across four visual control environments. ·
Researchers found that latent world action models lose generalization benefits when they skip future generation during inference. They proposed Simple-WAM, which uses a single forward pass to retain these advantages. ·
The source text lists titles for visual guides and experiments on recurrent neural networks, convolutional neural networks, and transformers, focusing on their accuracy and efficiency. ·
49 points and 1 comments on Hacker News, linking magazine.sebastianraschka.com.
191 points and 138 comments on Hacker News, linking anthropic.com.
JoasASantos maintains a repository of AI models that are uncensored or fine-tuned for cybersecurity tasks. ·
15 upvotes on Hugging Face Daily Papers.
Researchers introduced SAKI, a method routing token-level supervision in on-policy distillation using maximal coupling. It improves performance on mathematical reasoning benchmarks for small student models. ·
X posts other sources are linking to, then named voices on Bluesky and Mastodon
Baseten x OpenAI: Use open models like GLM 5.3 and Kimi K3 natively in Codex
1 points and 0 comments on Hacker News, linking an X post.
Best Open-Source AI Models for Cybersecurity
1 points and 0 comments on Hacker News, linking an X post.
Changing the difference between Codex plans to be Plus=1x Pro 100=5x Pro 200=10x
2 points and 2 comments on Hacker News, linking an X post.
Claude suddenly stopped cheating – Drone Bench
1 points and 1 comments on Hacker News, linking an X post.
Trump, Musk, Jensen. and Somewhere, Anthropic
2 points and 0 comments on Hacker News, linking an X post.
The Next 3x in Inference Won't Come from Faster Kernels
2 points and 1 comments on Hacker News, linking an X post.
New Cyber-OSINT model released
53 points and 16 comments on Hacker News, linking an X post.
Changing the difference between Codex plans to be Plus=1x Pro 100=5x Pro 200=10x
2 points and 2 comments on Hacker News, linking an X post.
OpenAI: Tomorrow we are re-opening the Pro $200 subscription
74 points and 90 comments on Hacker News, linking an X post.
OpenAI: Tomorrow we are re-opening the Pro $200 subscription
74 points and 90 comments on Hacker News, linking an X post.
New Cyber-OSINT model released
53 points and 16 comments on Hacker News, linking an X post.
Trump, Musk, Jensen. and Somewhere, Anthropic
2 points and 0 comments on Hacker News, linking an X post.
The Next 3x in Inference Won't Come from Faster Kernels
2 points and 1 comments on Hacker News, linking an X post.
Changing the difference between Codex plans to be Plus=1x Pro 100=5x Pro 200=10x
2 points and 2 comments on Hacker News, linking an X post.
New repos, releases and changelog
| KKKKhazix/AIHOT | 3399 |
| dzhng/jevgrep | 1791 |
| feitangyuan/onetake | 986 |
| kaankiziltug/logo-design-skill | 860 |
| lemomo-ai/lemo-opuscar | 620 |
| glanderness/BeefTV | 555 |
OpenAI updated Codex version 0.159.1 to make GPT-6.1 Sol the default model in its bundled and Amazon Bedrock catalogs. ·
New release rust-v0.159.1 of openai/codex on GitHub.
Google released version 0.62.0 of the Gemini CLI. The update includes fixes for OAuth token handling, UI layout rendering, and MCP tool call formatting. ·
New release v3.22.0 of openai/openai-python on GitHub.
New release v3.21.0 of openai/openai-python on GitHub.
OpenAI's GPT-6.1 Sol model is now generally available in GitHub Copilot for agentic coding and terminal workflows. ·
232 points and 89 comments on Hacker News, linking github.com.
New release v2.1.285 of anthropics/claude-code on GitHub.
OpenAI updated Codex version 0.159.1 to make GPT-6.1 Sol the default model in its bundled and Amazon Bedrock catalogs. ·
41 points and 5 comments on Hacker News, linking github.com.
OpenAI released version 0.159.0 of Codex, adding an opt-in feature that allows new input to steer the model during active responses or long-running code calls. ·
232 points and 89 comments on Hacker News, linking github.com.
JoasASantos maintains a repository of AI models that are uncensored or fine-tuned for cybersecurity tasks. ·
New release v2.1.285 of anthropics/claude-code on GitHub.
OpenAI updated Codex version 0.159.1 to make GPT-6.1 Sol the default model in its bundled and Amazon Bedrock catalogs. ·
41 points and 5 comments on Hacker News, linking github.com.
Builds, power deals and the grid
| Thu 24 | 3 |
| Fri 25 | 20 |
| Sat 26 | 1 |
| Sun 27 | 1 |
| Mon 28 | 20 |
| Tue 29 | 33 |
| Wed 30 | 0 |
Stories from DCD, Data Center Knowledge, EIA and Hacker News, by publication day.
197 points and 286 comments on Hacker News, linking thenationalnews.com.
197 points and 286 comments on Hacker News, linking thenationalnews.com.
197 points and 286 comments on Hacker News, linking thenationalnews.com.
10 points and 5 comments on Hacker News, linking wsj.com.
GPUs, memory, and what tokens cost
One point per OpenRouter reading, last 8 days. Data by OpenRouter, CC BY 4.0.
| gpt-6.1-sol-pro | 10 |
| gpt-6.1-sol-pro:batch | 5 |
| gpt-6.1-sol | 10 |
| gpt-6.1-sol:batch | 5 |
| claude-sonnet-5.5 | 10 |
| median, all paid models | 2 |
438 paid models listed; median of the 60 added in 30 days: $5.00. Rankings data by OpenRouter, CC BY 4.0.
Import AI issue 474 covers Platonic mindspace, space-based TPUs, and Zhipu's outer RSI loop. ·
Import AI issue 474 covers Platonic mindspace, space-based TPUs, and Zhipu's outer RSI loop. ·
An experimental project provides a virtio device that allows KVM virtual machines to access NVIDIA GPUs with near-native performance. ·