Archived front page, Sun 15 Nov 2026, 15:00 UTC. All of 2026-11-15 · Archive · Live page

The Daily Markdown

The front page of the internet for AI users · Sun 15 Nov 2026, 15:00 UTC · Get it by email

Top now

6 stories

What is climbing fastest across every source

GitHub-Copilot-app6.7xmanus2.06.0xNVIDIA5.0xClaude Code8.0xtoppatronweave6.7xtopconstableinlet676.7xtrackersnailgrip6.6xtrackboatwrightplan6.5xTrapBladesmith6.5xGPUs6.2x

Understanding Positional Embeddings in Modern LLMs

The item explains methods behind how modern large language models represent token positions through positional embeddings. ·

51 points and 0 comments on Hacker News, linking magazine.sebastianraschka.com.

Hacker News+ Ahead of AI3h ago▲8 to 51 across 5 pollsdiscussmagazine.sebastianraschka.com

datasette-llm 0.1a11

datasette-llm version 0.1a11 adds support for OpenAI-compatible endpoints, enabling Datasette to use local models from LM Studio, Ollama, or llama.cpp. ·

Simon Willison15h ago▲simonwillison.net

Abiray/MiniMax-H3-Image-GGUF

Abiray releases MiniMax-H3-Image-GGUF, a text-to-image model with 20 billion parameters available on Hugging Face. ·

Hugging Face17h ago▲204 to 228 across 34 pollshuggingface.co

Outpostkrewhirl86/Capri-AI

Capri is a local-first desktop AI agent for Windows featuring a cyberpunk HUD. It controls files, apps, terminal, and browser using various backend models. ·

GitHub17h ago▲github.com

Understanding Positional Embeddings in Modern LLMs

The item explains methods behind how modern large language models represent token positions through positional embeddings. ·

51 points and 0 comments on Hacker News, linking magazine.sebastianraschka.com.

Hacker News+ Ahead of AI3h ago▲8 to 51 across 5 pollsdiscussmagazine.sebastianraschka.com

Topkiflare74/ai--manus2.0

A new repository with 678 stars in its first week.

GitHub20h ago▲401 to 678 across 19 pollsgithub.com

Abiray/MiniMax-H3-Image-GGUF

Abiray releases MiniMax-H3-Image-GGUF, a text-to-image model with 20 billion parameters available on Hugging Face. ·

Hugging Face17h ago▲204 to 228 across 34 pollshuggingface.co

Understanding Positional Embeddings in Modern LLMs

The item explains methods behind how modern large language models represent token positions through positional embeddings. ·

51 points and 0 comments on Hacker News, linking magazine.sebastianraschka.com.

Hacker News+ Ahead of AI3h ago▲8 to 51 across 5 pollsdiscussmagazine.sebastianraschka.com

Topkiflare74/ai--manus2.0

A new repository with 678 stars in its first week.

GitHub20h ago▲401 to 678 across 19 pollsgithub.com

anthropics/claude-code v2.1.325

Anthropic releases Claude Code v2.1.325, which fixes a zstd security issue and adds a spinner tip for Jupyter notebooks. ·

GitHub Releases42h ago▲10 to 28 across 21 pollsgithub.com

ggml-org/llama.cpp v1.0.0

ggml-org releases llama.cpp version 1.0.0, which includes a nightly build labeled b12359 and a changelog since version 0.9.0. ·

GitHub Releases2d ago▲5 to 157 across 14 pollsgithub.com

Trending topics today: heat

GitHub-Copilot-app: 5.74GitHub-Copilot-app5.74manus2.0: 4.95manus2.04.95NVIDIA: 1NVIDIA1Claude Code: 0.80Claude Code0.80toppatronweave: 0.79toppatronweave0.79topconstableinlet67: 0.75topconstableinlet670.75
Table
GitHub-Copilot-app5.74
manus2.04.95
NVIDIA1
Claude Code0.8
toppatronweave0.79
topconstableinlet670.75

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

Qwen/Qwen3.8-27B ▲1: 20,069Qwen/Qwen3.8-27B ▲120,069Lightricks/LTX-2.5 ▼1: 12,000Lightricks/LTX-2.5 ▼112,000xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic: 7,466xtools-at/DeepSeek-V4…7,466DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NEO-CODER-MAX-MTP-GGUF ▲1: 3,144DavidAU/Qwen3.8-27… ▲13,144MiniMaxAI/MiniMax-H3-Image: 1,151MiniMaxAI/MiniMax-H3-…1,151Comfy-Org/MiniMax-H3-Image ▲3: 403Comfy-Org/MiniMax-… ▲3403
Table
Qwen/Qwen3.8-27B ▲120069
Lightricks/LTX-2.5 ▼112000
xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic7466
DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NEO-CODER-MAX-MTP-GGUF ▲13144
MiniMaxAI/MiniMax-H3-Image1151
Comfy-Org/MiniMax-H3-Image ▲3403

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/

The Claude Delusion

The Claude Delusion argues that its creator believes AI is amazing because he built it, yet also considers it superhuman because it surprised him. ·

Hacker News25h ago▲20 to 197 across 4 pollsdiscussnetmeister.org

higurashi-ai/Haru-4B

Hugging Face hosts the Haru-4B audio-to-audio model, which contains 5B parameters. ·

Hugging Face31h ago▲191 to 275 across 62 pollshuggingface.co

Show HN: DeepSeek V3 in ~200 lines of pure Jax

Vaibhaav Tiwari implements DeepSeek V3 in 190 lines of pure JAX, utilizing MLA, DeepSeekMoE, and MTP, and trains it on a MacBook Air. ·

252 points and 23 comments on Hacker News, linking github.com.

Hacker News24h ago▲3 to 275 across 100 pollsdiscussgithub.com

Gemini Decisions API

The Gemini Decisions API uses the Gemini Decision model to provide fast, calibrated multimodal decisions. ·

Hacker News+ Google AI, Simon Willison4d ago▲23 to 730 across 174 pollsdiscussblog.google

The Claude Delusion

The Claude Delusion argues that its creator believes AI is amazing because he built it, yet also considers it superhuman because it surprised him. ·

Hacker News25h ago▲20 to 197 across 4 pollsdiscussnetmeister.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

FrigadeHQ/yogurt: 1,574FrigadeHQ/yogurt1,574HybridRobotics/supermimic: 1,532HybridRobotics/superm…1,532agentes-chat/agentes-chat: 1,463agentes-chat/agentes-…1,463alexeyw/lot-map: 857alexeyw/lot-map857duwqijlk/muse-protocol: 813duwqijlk/muse-protocol813Noniv/abyssal-garden: 744Noniv/abyssal-garden744
Table
FrigadeHQ/yogurt1574
HybridRobotics/supermimic1532
agentes-chat/agentes-chat1463
alexeyw/lot-map857
duwqijlk/muse-protocol813
Noniv/abyssal-garden744

Topkiflare74/ai--manus2.0

A new repository with 678 stars in its first week.

GitHub20h ago▲401 to 678 across 19 pollsgithub.com

Topkiflare74/ai--manus2.0

A new repository with 678 stars in its first week.

GitHub20h ago▲401 to 678 across 19 pollsgithub.com

Infrastructure

5 stories

Data centers, power, chips and what tokens cost

Data-center and energy stories per day

Mon 09: 35Mon 0935Tue 10: 32Tue 1032Wed 11: 31Wed 1131Thu 12: 29Thu 1229Fri 13: 29Fri 1329Sat 14: 7Sat 147Sun 15: 1Sun 151
Table
Mon 0935
Tue 1032
Wed 1131
Thu 1229
Fri 1329
Sat 147
Sun 151

Stories from DCD, Data Center Knowledge, EIA and Hacker News, by publication day.

Output price, $ per million tokens (newest models)

gemini-4-hydrogen: 7.50gemini-4-hydrogen7.50gemini-4-hydrogen:batch: 3.75gemini-4-hydrogen:bat…3.75ernie-light-dt-1.0-intl: 3.94ernie-light-dt-1.0-in…3.94gemini-4-carbon: 10gemini-4-carbon10gemini-4-carbon:batch: 5gemini-4-carbon:batch5median, all paid models: 2median, all paid mode…2
Table
gemini-4-hydrogen7.5
gemini-4-hydrogen:batch3.75
ernie-light-dt-1.0-intl3.94
gemini-4-carbon10
gemini-4-carbon:batch5
median, all paid models2

443 paid models listed; median of the 29 added in 30 days: $2.50. Rankings data by OpenRouter, CC BY 4.0.

Will we put data centers in space?

Forethought analyzes launch costs, cooling physics, and engineering factors to determine how soon AI data centers could move to orbit. ·

Hacker News21h ago▲7 to 216 across 21 pollsdiscussforethought.org

Will we put data centers in space?

Forethought analyzes launch costs, cooling physics, and engineering factors to determine how soon AI data centers could move to orbit. ·

Hacker News21h ago▲7 to 216 across 21 pollsdiscussforethought.org