Archived front page, Fri 02 Oct 2026, 05:00 UTC. All of 2026-10-02 · Archive · Live page

The Daily Markdown

The front page of the internet for AI users · Fri 02 Oct 2026, 05:00 UTC

Top now

6 stories

What is climbing fastest across every source

Cloudflare7.1xAgents6.0xTe Hiku Media7.0xBeyond6.1xOpenAI4.0xNvidia7.4xContext Language8.2xDots6.1xExcited7.2xIntel6.8x

DeepSeek Harness

DeepSeek Harness lets users process documents, analyze spreadsheets, write code, and schedule tasks while extending functionality through composable plugins. ·

42 points and 14 comments on Hacker News, linking deepseek.com.

Hacker News2h ago▲6 to 56 across 5 pollsdiscussdeepseek.com

Cloudflare/clef-flash

A image text to text model from Cloudflare with 133 likes and 18 downloads on Hugging Face.

Hugging Face4h ago▲108 to 133 across 8 pollshuggingface.co

DeepSeek Harness

DeepSeek Harness lets users process documents, analyze spreadsheets, write code, and schedule tasks while extending functionality through composable plugins. ·

42 points and 14 comments on Hacker News, linking deepseek.com.

Hacker News2h ago▲6 to 56 across 5 pollsdiscussdeepseek.com

Cloudflare/clef-flash

A image text to text model from Cloudflare with 133 likes and 18 downloads on Hugging Face.

Hugging Face4h ago▲108 to 133 across 8 pollshuggingface.co

Hierarchical Continuous Diffusion Language Models

Researchers propose Hierarchical Continuous Diffusion Language Models, which couple discrete token generation with a continuous latent trajectory to improve structured reasoning and mathematical planning tasks. ·

HF Daily Papers5h ago▲9 to 43 across 3 pollsarxiv.org

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. ·

GitHub Releases6h ago▲0 to 1 across 3 pollsgithub.com

Hierarchical Continuous Diffusion Language Models

Researchers propose Hierarchical Continuous Diffusion Language Models, which couple discrete token generation with a continuous latent trajectory to improve structured reasoning and mathematical planning tasks. ·

HF Daily Papers5h ago▲9 to 43 across 3 pollsarxiv.org

DeepSeek Harness

DeepSeek Harness lets users process documents, analyze spreadsheets, write code, and schedule tasks while extending functionality through composable plugins. ·

Hacker News2h ago▲6 to 56 across 5 pollsdiscussdeepseek.com

Cloudflare/clef

A image text to text model from Cloudflare with 387 likes and 18 downloads on Hugging Face.

Hugging Face10h ago▲121 to 387 across 21 pollshuggingface.co

Cloudflare/clef-flash

A image text to text model from Cloudflare with 133 likes and 18 downloads on Hugging Face.

Hugging Face4h ago▲108 to 133 across 8 pollshuggingface.co

Trending topics today: heat

Cloudflare: 1.46Cloudflare1.46Agents: 1.41Agents1.41Te Hiku Media: 1.09Te Hiku Media1.09Beyond: 0.93Beyond0.93OpenAI: 1.12OpenAI1.12Nvidia: 0.83Nvidia0.83
Table
Cloudflare1.46
Agents1.41
Te Hiku Media1.09
Beyond0.93
OpenAI1.12
Nvidia0.83

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 ▲1: 4,871convaiinnovations/… ▲14,871Qwen/Qwen-Image-2.1 ▲1: 2,795Qwen/Qwen-Image-2.1 ▲12,795abenzerps/Qwen-Image-2.1-Uncensored-GGUF ▲1: 2,720abenzerps/Qwen-Ima… ▲12,720XingChen-AGI/TeleOCR ▲1: 1,235XingChen-AGI/TeleO… ▲11,235Contrastive-LM/CLM-v0.1-8B ▲1: 628Contrastive-LM/CLM… ▲1628Cloudflare/clef new: 387Cloudflare/clef new387
Table
convaiinnovations/laya ▲14871
Qwen/Qwen-Image-2.1 ▲12795
abenzerps/Qwen-Image-2.1-Uncensored-GGUF ▲12720
XingChen-AGI/TeleOCR ▲11235
Contrastive-LM/CLM-v0.1-8B ▲1628
Cloudflare/clef new387

PirateBench: score out of 100

gpt-6.1-sol: 99gpt-6.1-sol99opus-5-5: 96opus-5-596gpt-6-sol: 92gpt-6-sol92gpt-6-luna: 65gpt-6-luna65haiku-4-5: 4.28haiku-4-54.28
Table
gpt-6.1-sol98.9
opus-5-596.46
gpt-6-sol91.6
gpt-6-luna65.25
haiku-4-54.28

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

Hierarchical Continuous Diffusion Language Models

Researchers propose Hierarchical Continuous Diffusion Language Models, which couple discrete token generation with a continuous latent trajectory to improve structured reasoning and mathematical planning tasks. ·

42 upvotes on Hugging Face Daily Papers.

HF Daily Papers5h ago▲9 to 43 across 3 pollsarxiv.org

Agent Priors-guided Policy Learning

Researchers introduce APPL, a method where a construction agent segments robot demonstrations into skills and trains policies using structural priors to improve generalization. ·

HF Daily Papers5h ago▲1 to 18 across 2 pollsarxiv.org

ROWBench: Do Video Models Render What the Program Specifies?

Researchers introduce PROWBench, a benchmark containing 170 programmatically constructed episodes and 600 proxy videos that tests whether video models accurately render specified world events. ·

HF Daily Papers5h ago▲5 to 15 across 2 pollsarxiv.org

Sharpening Tax in Post-Training

Researchers introduce Sharpening Tax, a metric quantifying reduced test-time scalability in post-trained language models, based on analysis of fourteen model pairs across three agentic benchmarks. ·

HF Daily Papers5h ago▲5 to 12 across 2 pollsarxiv.org

Hierarchical Continuous Diffusion Language Models

Researchers propose Hierarchical Continuous Diffusion Language Models, which couple discrete token generation with a continuous latent trajectory to improve structured reasoning and mathematical planning tasks. ·

42 upvotes on Hugging Face Daily Papers.

HF Daily Papers5h ago▲9 to 43 across 3 pollsarxiv.org

Agent Priors-guided Policy Learning

Researchers introduce APPL, a method where a construction agent segments robot demonstrations into skills and trains policies using structural priors to improve generalization. ·

HF Daily Papers5h ago▲1 to 18 across 2 pollsarxiv.org

ROWBench: Do Video Models Render What the Program Specifies?

Researchers introduce PROWBench, a benchmark containing 170 programmatically constructed episodes and 600 proxy videos that tests whether video models accurately render specified world events. ·

HF Daily Papers5h ago▲5 to 15 across 2 pollsarxiv.org

Sharpening Tax in Post-Training

Researchers introduce Sharpening Tax, a metric quantifying reduced test-time scalability in post-trained language models, based on analysis of fourteen model pairs across three agentic benchmarks. ·

HF Daily Papers5h ago▲5 to 12 across 2 pollsarxiv.org

X voices

5 stories

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

@OfficialLoganK

5 points and 1 comments on Hacker News, linking an X post.

X, via Hacker News33h ago▲3 to 6 across 33 pollsView on X

GitHub

5 stories

New repos, releases and changelog

New AI repos this week: stars

KKKKhazix/AIHOT: 4,722KKKKhazix/AIHOT4,722Louis-CFM/coucou: 2,531Louis-CFM/coucou2,531feder-cr/dots: 2,401feder-cr/dots2,401dzhng/jevgrep: 2,006dzhng/jevgrep2,006rehan-remade/universal-modder: 1,624rehan-remade/universa…1,624kaankiziltug/logo-design-skill: 1,371kaankiziltug/logo-des…1,371
Table
KKKKhazix/AIHOT4722
Louis-CFM/coucou2531
feder-cr/dots2401
dzhng/jevgrep2006
rehan-remade/universal-modder1624
kaankiziltug/logo-design-skill1371

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. ·

New release v3.23.0 of openai/openai-python on GitHub.

GitHub Releases6h ago▲0 to 1 across 3 pollsgithub.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. ·

GitHub Releases9h ago▲0 to 1 across 4 pollsgithub.com

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. ·

GitHub Releases6h ago▲0 to 1 across 3 pollsgithub.com

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. ·

GitHub Releases15h ago▲6 to 16 across 7 pollsgithub.com

PostHog/jeeves

PostHog released a project named Jeeves that uses reasoning to improve Jev-like decision models. ·

GitHub+ Hacker News3d ago▲116 to 388 across 63 pollsgithub.com

Infrastructure

5 stories

Data centers, power, chips and what tokens cost

Data-center and energy stories per day

Sat 26: 1Sat 261Sun 27: 1Sun 271Mon 28: 20Mon 2820Tue 29: 33Tue 2933Wed 30: 35Wed 3035Thu 01: 25Thu 0125Fri 02: 0Fri 020
Table
Sat 261
Sun 271
Mon 2820
Tue 2933
Wed 3035
Thu 0125
Fri 020

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 $4.00Min $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)

pareto-26.10-preview: 3.20pareto-26.10-preview3.20gpt-6.1-sol-pro: 10gpt-6.1-sol-pro10gpt-6.1-sol: 10gpt-6.1-sol10claude-sonnet-5.5: 10claude-sonnet-5.510claude-sonnet-5.5:batch: 5claude-sonnet-5.5:bat…5median, all paid models: 2median, all paid mode…2
Table
pareto-26.10-preview3.2
gpt-6.1-sol-pro10
gpt-6.1-sol10
claude-sonnet-5.510
claude-sonnet-5.5:batch5
median, all paid models2

437 paid models listed; median of the 56 added in 30 days: $4.00. Rankings data by OpenRouter, CC BY 4.0.

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. ·

Hacker News39h ago▲6 to 247 across 97 pollsdiscussalphapixeldev.com