Hot take on a weak White House Accord on “Super Intelligence”
From Gary Marcus.
Top now
6 storiesWhat is climbing fastest across every source
OpenAI6.2xAnthropic7.2xGPT-6.17.0xMuse7.2xClaude7.1xLanguage6.7xDevDay7.1xGPT-6.1 Sol7.0xMeta7.2xCodex7.6x
openai/codex 0.159.1
OpenAI updated Codex version 0.159.1 to make GPT-6.1 Sol the default model in its bundled and Amazon Bedrock catalogs. ·
Quoting Anthropic Frontier Red Team
Anthropic's red team found that GLM-5.3 and Claude Mythos Preview achieved full control flow hijacks in 4% and 6% of 100 binary exploitation tasks, respectively, while earlier models succeeded in none. ·
From Math Olympiads to Navier-Stokes: How Fast Is AI Progressing? with Greg Burnham - #778
Greg Burnham of Epoch AI discusses how AI systems have advanced from basic arithmetic to assisting with long-standing research problems like Navier-Stokes. ·
anthropics/claude-code v2.1.285
New release v2.1.285 of anthropics/claude-code on GitHub.
AI needs $6T in annual revenue to justify data centre boom
185 points and 257 comments on Hacker News, linking thenationalnews.com.
Jeeves. Reasoning improves Jev-like decision models
220 points and 85 comments on Hacker News, linking github.com.
GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price
746 points and 683 comments on Hacker News, linking openai.com.
anthropics/claude-code v2.1.285
New release v2.1.285 of anthropics/claude-code on GitHub.
openai/codex 0.159.1
OpenAI updated Codex version 0.159.1 to make GPT-6.1 Sol the default model in its bundled and Amazon Bedrock catalogs. ·
ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-Coder-GGUF
A image text to text model from ISTA-DASLab with 135 likes and 19.5k downloads on Hugging Face.
Post-Training Leaves Behavioral Shadows on Unrelated Decisions
Researchers introduced Active Taskless Distillation, a method transferring capabilities using single-word prompts. In coding tests, this approach yielded a 5.34 percentage point gain on HumanEval+. ·
PostHog/jeeves
220 points and 85 comments on Hacker News, linking github.com.
GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price
746 points and 683 comments on Hacker News, linking openai.com.
Language models for text classification: From bag-of-words to Jev
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. ·
JoasASantos/Offensive-Security-AI-Models
JoasASantos maintains a repository of AI models that are uncensored or fine-tuned for cybersecurity tasks. ·
anthropics/claude-code v2.1.285
New release v2.1.285 of anthropics/claude-code on GitHub.
openai/codex 0.159.1
OpenAI updated Codex version 0.159.1 to make GPT-6.1 Sol the default model in its bundled and Amazon Bedrock catalogs. ·
Trending topics today: heat
Table
| OpenAI | 4.93 |
| Anthropic | 3.75 |
| GPT-6.1 | 1.44 |
| Muse | 1.36 |
| Claude | 1.35 |
| Language | 1.32 |
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 | 4505 |
| Qwen/Qwen-Image-2.1 | 2656 |
| abenzerps/Qwen-Image-2.1-Uncensored-GGUF | 2432 |
| Edge0/Audio8-ASR-Infinite | 1507 |
| XingChen-AGI/TeleOCR | 867 |
| Contrastive-LM/CLM-v0.1-8B | 527 |
PirateBench: score out of 100
Table
| 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/
google-gemini/gemini-cli Release v0.62.0
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 v0.62.0 of google-gemini/gemini-cli on GitHub.
Show HN: A working 3D model of an Enigma machine
A developer shared a working 3D model of an Enigma machine that allows users to explore its electrical circuits and rotor mechanisms. ·
New on OpenRouter: OpenAI: GPT-6.1 Sol (batch)
OpenAI added GPT-6.1 Sol to OpenRouter. This model is an upgrade to GPT-6 Sol and sits below the flagship GPT-6 Astra in the series. ·
New on OpenRouter: OpenAI: GPT-6.1 Sol Pro
OpenRouter added GPT-6.1 Sol Pro, a variant of OpenAI's GPT-6.1 Sol model. It uses a pro reasoning mode for complex tasks but costs significantly more. ·
GPT-6.1 Sol in GitHub Copilot
OpenAI's GPT-6.1 Sol model is now generally available in GitHub Copilot for agentic coding and terminal workflows. ·
Jeeves. Reasoning improves Jev-like decision models
220 points and 85 comments on Hacker News, linking github.com.
Language models for text classification: From bag-of-words to Jev
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. ·
Show HN: A working 3D model of an Enigma machine
A developer shared a working 3D model of an Enigma machine that allows users to explore its electrical circuits and rotor mechanisms. ·
Claude partial outage
169 points and 143 comments on Hacker News, linking status.claude.com.
Show HN: Jevstiller – Distill Jev into a local model, with a disagreement bound
Jevstiller distills Jev into a local model and provides a disagreement bound. The project includes benchmarks that can be rerun locally without an API key. ·
Language models for text classification: From bag-of-words to Jev
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. ·
13 points and 0 comments on Hacker News, linking magazine.sebastianraschka.com.
Claude Sonnet 5.5
191 points and 138 comments on Hacker News, linking anthropic.com.
JoasASantos/Offensive-Security-AI-Models
JoasASantos maintains a repository of AI models that are uncensored or fine-tuned for cybersecurity tasks. ·
Claude partial outage
169 points and 143 comments on Hacker News, linking status.claude.com.
Show HN: A working 3D model of an Enigma machine
A developer shared a working 3D model of an Enigma machine that allows users to explore its electrical circuits and rotor mechanisms. ·
X voices
5 storiesX posts other sources are linking to, then named voices on Bluesky and Mastodon
@thsottiaux
OpenAI: Tomorrow we are re-opening the Pro $200 subscription
74 points and 90 comments on Hacker News, linking an X post.
@philipkiely
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.
@0x0SojalSec
Best Open-Source AI Models for Cybersecurity
1 points and 0 comments on Hacker News, linking an X post.
@thsottiaux
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.
@lukaspet
Claude suddenly stopped cheating – Drone Bench
1 points and 1 comments on Hacker News, linking an X post.
@Geiger_Capital
Trump, Musk, Jensen. and Somewhere, Anthropic
2 points and 0 comments on Hacker News, linking an X post.
@usemuna
The Next 3x in Inference Won't Come from Faster Kernels
2 points and 1 comments on Hacker News, linking an X post.
@thsottiaux
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.
@0x0SojalSec
New Cyber-OSINT model released
53 points and 16 comments on Hacker News, linking an X post.
@thsottiaux
OpenAI: Tomorrow we are re-opening the Pro $200 subscription
74 points and 90 comments on Hacker News, linking an X post.
@thsottiaux
OpenAI: Tomorrow we are re-opening the Pro $200 subscription
74 points and 90 comments on Hacker News, linking an X post.
@0x0SojalSec
New Cyber-OSINT model released
53 points and 16 comments on Hacker News, linking an X post.
@Geiger_Capital
Trump, Musk, Jensen. and Somewhere, Anthropic
2 points and 0 comments on Hacker News, linking an X post.
@usemuna
The Next 3x in Inference Won't Come from Faster Kernels
2 points and 1 comments on Hacker News, linking an X post.
@thsottiaux
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.
GitHub
5 storiesNew repos, releases and changelog
New AI repos this week: stars
Table
| KKKKhazix/AIHOT | 3062 |
| yetone/magpie | 2627 |
| dzhng/jevgrep | 1770 |
| mikehasa/golive-skill | 1110 |
| feitangyuan/onetake | 979 |
| ollaya-dev/ollaya | 974 |
Show HN: TurboGPT: train 22KiB transformer in 13s
41 points and 5 comments on Hacker News, linking github.com.
google-gemini/gemini-cli Release v0.62.0
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. ·
openai/openai-python v3.22.0
New release v3.22.0 of openai/openai-python on GitHub.
openai/openai-python v3.21.0
New release v3.21.0 of openai/openai-python on GitHub.
GPT-6.1 Sol in GitHub Copilot
OpenAI's GPT-6.1 Sol model is now generally available in GitHub Copilot for agentic coding and terminal workflows. ·
Jeeves. Reasoning improves Jev-like decision models
220 points and 85 comments on Hacker News, linking github.com.
Show HN: TurboGPT: train 22KiB transformer in 13s
41 points and 5 comments on Hacker News, linking github.com.
openai/codex 0.159.0
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. ·
KKKKhazix/AIHOT
A website framework automatically identifies trending topics and generates daily reports. Users can customize its sources and selection criteria to create industry-specific news sites. ·
openai/mcp-extensions
OpenAI released a repository for building plugins that function as native features within ChatGPT. ·
PostHog/jeeves
220 points and 85 comments on Hacker News, linking github.com.
JoasASantos/Offensive-Security-AI-Models
JoasASantos maintains a repository of AI models that are uncensored or fine-tuned for cybersecurity tasks. ·
openai/codex 0.159.0
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. ·
Show HN: TurboGPT: train 22KiB transformer in 13s
41 points and 5 comments on Hacker News, linking github.com.
KKKKhazix/AIHOT
A website framework automatically identifies trending topics and generates daily reports. Users can customize its sources and selection criteria to create industry-specific news sites. ·
Data centers & energy
5 storiesBuilds, power deals and the grid
Data-center and energy stories per day
Table
| 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.
Free space optics: To the moon and back
Researchers are exploring optical communication for space applications. The technology's success depends on its ability to function reliably despite adverse weather conditions. ·
Free space optics: To the moon and back
Researchers are exploring optical communication for space applications. The technology's success depends on its ability to function reliably despite adverse weather conditions. ·
Every Household in This Rural Town Receives $10k If a Data Center Gets Built
10 points and 5 comments on Hacker News, linking wsj.com.
Hardware prices
2 storiesGPUs, memory, and what tokens cost
Output price, $ per million tokens (newest models)
Table
| 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 474: Platonic mindspace; TPUs in space; Zhipu starts an outer RSI loop
Import AI issue 474 covers Platonic mindspace, space-based TPUs, and Zhipu's outer RSI loop. ·
Import AI 474: Platonic mindspace; TPUs in space; Zhipu starts an outer RSI loop
Import AI issue 474 covers Platonic mindspace, space-based TPUs, and Zhipu's outer RSI loop. ·
Virtio-nvgpu: Near-native Nvidia GPU access inside a KVM guest
An experimental project provides a virtio device that allows KVM virtual machines to access NVIDIA GPUs with near-native performance. ·
