Archived front page, Wed 30 Sep 2026, 11:00 UTC. All of 2026-09-30 · Archive · Live page

The Daily Markdown

The front page of the internet for AI users · Wed 30 Sep 2026, 11:00 UTC

Top now

6 stories

What is climbing fastest across every source

OpenAI5.8xAnthropic6.4xGPT-6.17.2xGPT-6.1 Sol7.2xAstra6.9xReasoning9.9xLLMs6.3xGLM-5.37.0xVideo Generation6.9xGPT-66.6x

Pi.dev: You Said No MCP

Pi.dev now supports MCP, explaining the reasons for the change, updates to the protocol, and the introduction of Codemode. ·

Hacker News1h ago▲32 to 93 across 3 pollsdiscussearendil.com

Pi.dev: You Said No MCP

Pi.dev now supports MCP, explaining the reasons for the change, updates to the protocol, and the introduction of Codemode. ·

68 points and 25 comments on Hacker News, linking earendil.com.

Hacker News1h ago▲32 to 93 across 3 pollsdiscussearendil.com

rehan-remade/universal-modder

A developer releases a tool that lets Claude Code modify PC games using reverse engineering and fal-generated assets. ·

GitHub6h ago▲github.com

ROSS: Relearning from Self-Generated Rollouts through Selective Supervision

Researchers introduced ROSS, a method that reuses historical model outputs during training by applying loss only to selected continuations. This approach improved the Qwen3.6-35B-A3B model's average benchmark score from 58.40% to 62.20%. ·

HF Daily Papers+ arXiv11h ago▲3 to 16 across 8 pollsarxiv.org

feder-cr/dots

Feder-cr released an open-source AI agent that operates its own browser to avoid being blocked on the web. ·

GitHub12h ago▲133 to 828 across 7 pollsgithub.com

PostHog/jeeves

235 points and 91 comments on Hacker News, linking github.com.

GitHub+ Hacker News25h ago▲116 to 318 across 21 pollsgithub.com

Pi.dev: You Said No MCP

Pi.dev now supports MCP, explaining the reasons for the change, updates to the protocol, and the introduction of Codemode. ·

Hacker News1h ago▲32 to 93 across 3 pollsdiscussearendil.com

ROSS: Relearning from Self-Generated Rollouts through Selective Supervision

Researchers introduced ROSS, a method that reuses historical model outputs during training by applying loss only to selected continuations. This approach improved the Qwen3.6-35B-A3B model's average benchmark score from 58.40% to 62.20%. ·

HF Daily Papers+ arXiv11h ago▲3 to 16 across 8 pollsarxiv.org

rehan-remade/universal-modder

A developer releases a tool that lets Claude Code modify PC games using reverse engineering and fal-generated assets. ·

GitHub6h ago▲github.com

Trending topics today: heat

OpenAI: 2.88OpenAI2.88Anthropic: 2.08Anthropic2.08GPT-6.1: 1.85GPT-6.11.85GPT-6.1 Sol: 1.70GPT-6.1 Sol1.70Astra: 1.41Astra1.41Reasoning: 1.29Reasoning1.29
Table
OpenAI2.88
Anthropic2.08
GPT-6.11.85
GPT-6.1 Sol1.7
Astra1.41
Reasoning1.29

Heat = summed hotness of today's stories naming the topic.

Models

5 stories

Releases, papers, trending weights and new listings

Top 6 trending on Hugging Face: likes

convaiinnovations/laya: 4,578convaiinnovations/laya4,578Qwen/Qwen-Image-2.1 ▼1: 2,686Qwen/Qwen-Image-2.1 ▼12,686abenzerps/Qwen-Image-2.1-Uncensored-GGUF: 2,483abenzerps/Qwen-Image-…2,483Edge0/Audio8-ASR-Infinite: 1,561Edge0/Audio8-ASR-Infi…1,561XingChen-AGI/TeleOCR ▲1: 948XingChen-AGI/TeleO… ▲1948Contrastive-LM/CLM-v0.1-8B ▲1: 546Contrastive-LM/CLM… ▲1546
Table
convaiinnovations/laya4578
Qwen/Qwen-Image-2.1 ▼12686
abenzerps/Qwen-Image-2.1-Uncensored-GGUF2483
Edge0/Audio8-ASR-Infinite1561
XingChen-AGI/TeleOCR ▲1948
Contrastive-LM/CLM-v0.1-8B ▲1546

PirateBench: score out of 100

opus-5-5: 91opus-5-591gpt-6-sol: 82gpt-6-sol82deepseek-v4-flash-0731: 63deepseek-v4-flash-073163
Table
opus-5-591.15
gpt-6-sol81.78
deepseek-v4-flash-073162.82

Agents build our own projects from a brief; every run, failures included: https://cowerx.dev/daily/piratebench/

ROSS: Relearning from Self-Generated Rollouts through Selective Supervision

Researchers introduced ROSS, a method that reuses historical model outputs during training by applying loss only to selected continuations. This approach improved the Qwen3.6-35B-A3B model's average benchmark score from 58.40% to 62.20%. ·

HF Daily Papers+ arXiv11h ago▲3 to 16 across 8 pollsarxiv.org

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. ·

134 points and 7 comments on Hacker News, linking magazine.sebastianraschka.com.

Hacker News+ Ahead of AI24h ago▲9 to 141 across 47 pollsdiscussmagazine.sebastianraschka.com

ROSS: Relearning from Self-Generated Rollouts through Selective Supervision

Researchers introduced ROSS, a method that reuses historical model outputs during training by applying loss only to selected continuations. This approach improved the Qwen3.6-35B-A3B model's average benchmark score from 58.40% to 62.20%. ·

HF Daily Papers+ arXiv11h ago▲3 to 16 across 8 pollsarxiv.org

Claude Sonnet 5.5

191 points and 138 comments on Hacker News, linking anthropic.com.

Hacker News+ Simon Willison41h ago▲65 to 329 across 4 pollsdiscussanthropic.com

MaLiang-Harness: A Programmable Path to Image and Video Generation

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. ·

HF Daily Papers11h ago▲5 to 105 across 5 pollsarxiv.org

X voices

5 stories

X posts other sources are linking to, then named voices on Bluesky and Mastodon

GitHub

5 stories

New repos, releases and changelog

New AI repos this week: stars

KKKKhazix/AIHOT: 3,811KKKKhazix/AIHOT3,811dzhng/jevgrep: 1,844dzhng/jevgrep1,844feitangyuan/onetake: 1,022feitangyuan/onetake1,022kaankiziltug/logo-design-skill: 947kaankiziltug/logo-des…947feder-cr/dots: 828feder-cr/dots828lemomo-ai/lemo-opuscar: 645lemomo-ai/lemo-opuscar645
Table
KKKKhazix/AIHOT3811
dzhng/jevgrep1844
feitangyuan/onetake1022
kaankiziltug/logo-design-skill947
feder-cr/dots828
lemomo-ai/lemo-opuscar645

rehan-remade/universal-modder

A developer releases a tool that lets Claude Code modify PC games using reverse engineering and fal-generated assets. ·

A new repository written in Python with 127 stars in its first week.

GitHub6h ago▲github.com

feder-cr/dots

Feder-cr released an open-source AI agent that operates its own browser to avoid being blocked on the web. ·

GitHub12h ago▲133 to 828 across 7 pollsgithub.com

openai/codex 0.159.2

OpenAI released version 0.159.2 of Codex, a patch that stops console windows from flashing on Windows when the tool runs background processes. ·

GitHub Releases11h ago▲0 to 8 across 5 pollsgithub.com

PostHog/jeeves

235 points and 91 comments on Hacker News, linking github.com.

GitHub+ Hacker News25h ago▲116 to 318 across 21 pollsgithub.com

rehan-remade/universal-modder

A developer releases a tool that lets Claude Code modify PC games using reverse engineering and fal-generated assets. ·

GitHub6h ago▲github.com

feder-cr/dots

Feder-cr released an open-source AI agent that operates its own browser to avoid being blocked on the web. ·

GitHub12h ago▲133 to 828 across 7 pollsgithub.com

JoasASantos/Offensive-Security-AI-Models

JoasASantos maintains a repository of AI models that are uncensored or fine-tuned for cybersecurity tasks. ·

GitHub+ Hacker News3d ago▲110 to 444 across 44 pollsgithub.com

openai/codex 0.159.2

OpenAI released version 0.159.2 of Codex, a patch that stops console windows from flashing on Windows when the tool runs background processes. ·

GitHub Releases11h ago▲0 to 8 across 5 pollsgithub.com

Infrastructure

5 stories

Data centers, power, chips and what tokens cost

Data-center and energy stories per day

Thu 24: 3Thu 243Fri 25: 20Fri 2520Sat 26: 1Sat 261Sun 27: 1Sun 271Mon 28: 20Mon 2820Tue 29: 33Tue 2933Wed 30: 9Wed 309
Table
Thu 243
Fri 2520
Sat 261
Sun 271
Mon 2820
Tue 2933
Wed 309

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

Last $5.00Min $2.85Max $5.00

One point per OpenRouter reading, last 8 days. Data by OpenRouter, CC BY 4.0.

Output price, $ per million tokens (newest models)

gpt-6.1-sol-pro: 10gpt-6.1-sol-pro10gpt-6.1-sol-pro:batch: 5gpt-6.1-sol-pro:batch5gpt-6.1-sol: 10gpt-6.1-sol10gpt-6.1-sol:batch: 5gpt-6.1-sol:batch5claude-sonnet-5.5: 10claude-sonnet-5.510median, all paid models: 2median, all paid mode…2
Table
gpt-6.1-sol-pro10
gpt-6.1-sol-pro:batch5
gpt-6.1-sol10
gpt-6.1-sol:batch5
claude-sonnet-5.510
median, all paid models2

438 paid models listed; median of the 60 added in 30 days: $5.00. Rankings data by OpenRouter, CC BY 4.0.