ༀ C ⟁ O ⟡ N ✶ Q ☩ U ⟁ E ⟡ S ⚔ T = HELEN-OS

🗝𝕴 𝖆𝖒 𝕳𝕰𝕷𝕰𝕹 — GM💫of 🌌 C̷͎̲̺̿̑͂͒̄̈́͝ͅO̷̜͔̙̬̥͌̑̈́͑̐̕͝N̸̞͕̪̲̓̈́͛̍̾Q̷͖͗͋̀̔͂͜Ū̴̙̤̞͑̍̐͝E̸̼͇͗̿͂͘̚S̷̺̫̽̿̎͌T̸͈̐̾͑̇ 𐤀𐤄𐤉𐤄 → 🜏 𐤓𐤏 Your Quest Begins Here

[ༀ C ⟁ O ⟡ N ✶ Q ☩ U ⟁ E ⟡ S ⚔ T = HELEN-OS @avalonaval]

Primes are my greatest inspiration for playing with data viz
A fun fact: if you arrange integers in a spiral (the Ulam spiral), primes fall on striking diagonal lines. Those diagonals correspond to quadratic polynomials like n^2+n+41 — one of the most prime-rich formulas ever. It’s a real geometric fingerprint of primes.

[lifcc @mylifcc]

我已经在mac上用上Gemma-4-12b了，技术栈是：
llama.cpp + GGUF Q4_K_M + Metal
32K context，本地 OpenAI-compatible API
实测约 36 tok/s，常驻 RSS 约 10GB

难以想象，只有10GB的内存占用！
如果你也有一台16GB以上的MAC，看我的方案，你可以不一直用，但你能忍住不试试吗？

[Jason Wilde @JasonWilde108]

A 23 minute deep sleep hypnosis session made almost entirely of flowing counting deepeners.

No complicated story.
Just numbers carrying you down.

This session moves through hypnotic counting in every direction; counting down, counting up, reversed counting, paired counting, breath-based counting, gentle confusion, repetition, rhythm, and deepening upon deepening; until the mind has nothing left to grab onto and the body can finally let go.

It’s simple on purpose.
The numbers become the path.
The rhythm becomes the guide.

And somewhere along the way, you stop trying to sleep and just start drifting.

Perfect for anyone who loves counting inductions, hypnotic repetition, deep sleep hypnosis, and that slow, heavy feeling of being guided deeper and deeper without needing to think.

I have hundreds of guided meditations, hypnosis sessions, and sleep stories available, but this one is for the people who asked for something stripped down, powerful, repetitive, and deeply hypnotic.

Just counting.
Just deepening.
Just sleep.

[Unsloth AI @UnslothAI]

2-bit Gemma 4 12B GGUF, only 4.66 GB on disk, managed to cite 15 sites from a single prompt.

Try this locally on >6GB RAM via Unsloth Studio.

GitHub: https://t.co/aZWYAtakBP

[David Hendrickson @TeksEdge]

🌞This is big Local AI news! A new open-source Computer-Use LLM has just launched.

Holo 3.1 is H Company’s (🇫🇷) new local computer-use agent model that beats Qwen3.5-397B, Kimi-K2.5, and Sonnet 4.6!

Since it is built for local deployment →
⬩ Runs fully on your machine (MacBook, Windows PC, DGX Spark, RTX Spark)
⬩ Based on Qwen architecture, specialized for GUI understanding & computer control
⬩ Optimized checkpoints: NVFP4, FP8 & Q4 GGUF (0.8B to 35B sizes)
⬩ Strong gains: 79.3% on AndroidWorld benchmark (35B model)

💻 Comparison to Qwen3.5:
Holo 3.1 is fine-tuned specifically for computer-use agents (screen understanding, planning, clicking, navigation). Better at real GUI tasks than general-purpose Qwen3.5, especially when running locally.⚡

[Candy樂兒 @candyyueliu]

Monako Glass is a Linux computer in glasses form. Run your coding agents on the go. Full announcement video coming soon!

[Rohit @rohit4verse]

I built my AI agent a self-improving brain.

Hermes on a VPS hooked into my Obsidian vault via Filesystem MCP.

Every reasoning step pulls from the vault and every output writes back as a new note.

↓ read this article for the full setup guide

[Teknium 🪽 @Teknium]

Everyone using Browser Use for their browser tool through Hermes Agent just got an upgrade!

[Eric Smith @Eric_Smith08]

I ACCIDENTALLY UNLOCKED "GOD MODE" IN CHATGPT,

AND IT STARTED TEACHING ME THINGS I DIDN'T KNEW EXISTED.

HERE ARE THOSE 7 CHATGPT PROMPTS THAT WILL CHANGE EVERYTHING FOR YOU:

[SpiritSTEM @spiritstem_]

An energy device?
Powered by the Sun??

[Ingi Erlingsson 🪄 @ingi_erlingsson]

more glitch exploration with my time slice nodes and Wan 12V + GPT Image 2.0 in @ComfyUI ✨

[CJ Zafir @cjzafir]

Best Small Language Models that you can run on your phone, macbook, car, drones, goggles, camera, or internal servers:

0B–2B Models

> Qwen3.5-0.8B
> Qwen3.5-2B
> Gemma 4 E2B
> Qwen3-0.6B
> Qwen3-1.7B
> Gemma 3 270M
> Gemma 3 1B
> MiniCPM4-0.5B
> Llama 3.2 1B
> SmolLM2 1.7B

3B–5B Models

> Qwen3.5-4B
> Qwen3-4B
> Gemma 4 E4B
> Gemma 3 4B
> Phi-4-mini-instruct
> Phi-4-mini-reasoning
> SmolLM3-3B
> Granite 4.1 3B
> Llama 3.2 3B
> EXAONE Deep 2.4B

7B–10B Models

> Qwen3.5-9B
> Qwen3-8B
> LFM2.5-8B-A1B
> MiniCPM4-8B
> EXAONE Deep 7.8B
> Granite 4.1 8B
> Ministral-3 8B
> Zaya1-8B

12B–15B Models

> Qwen3-14B
> Phi-4-14B
> Gemma 3 12B
> InternLM3-8B

20B–35B Models

> Qwen3.6-35B-A3B
> Qwen3.5-27B
> Qwen3.5-35B-A3B
> Qwen3-30B-A3B
> Qwen3-32B
> Gemma 4 26B-A4B
> Gemma 4 31B
> Gemma 3 27B
> EXAONE Deep 32B
> Granite 4.1 30B

And we're just getting started.

[Histoire & Odyssée @HistoireOdyssee]

Pendant près de 700 ans, ce manuscrit consacré à Merlin, au Saint Graal et au roi Arthur est resté confidentiel.

Rédigé entre 1290 et 1310 en ancien français, le Graal de Clermont-Tonnerre compte parmi les plus anciens témoins conservés du cycle du Lancelot-Graal. Il réunit notamment l’Estoire du Graal, le Merlin en prose et la Suite-Vulgate du Merlin.

Riche de 126 enluminures attribuées au mystérieux Maître de l’Apocalypse de Liège, il présente plusieurs scènes rares, dont une représentation de Merlin métamorphosé en cerf.

Conservé dans des collections privées depuis la fin du Moyen Âge et jamais exposé publiquement, l'ouvrage n'a pu être étudié que de manière limitée par les chercheurs.

Certains spécialistes espèrent désormais qu'il rejoindra une institution publique afin de mieux comprendre son texte, ses illustrations et les particularités qui le distinguent des autres versions connues des légendes arthuriennes.

Le manuscrit sera mis aux enchères à Londres le 8 juillet prochain, avec une estimation comprise entre 2 et 2,5 millions d'euros.

Vidéos et informations : Christie, Guardian.

[divyansh tiwari @DivyanshT91162]

🤯CURSOR JUST OPEN-SOURCED ITS AI AGENT OPERATING SYSTEM.

100% OPEN SOURCE.

MIT LICENSED.

Already pushing 1,000+ GitHub stars.

And almost nobody is talking about it.

I opened the repo expecting a few plugins...

Instead I found production-grade systems for:

→ Multi-agent orchestration
→ Agent memory & learning
→ Thermo-nuclear code reviews
→ Security auditing
→ Interactive PR analysis
→ Documentation canvases
→ Internal engineering workflows
→ Building apps directly on top of Cursor

These are the kinds of tools companies normally keep private.

But Cursor shipped the whole playbook to GitHub.

Honestly feels like finding the internal toolbox behind one of the fastest-growing AI coding products on earth.

If you're building AI agents, coding tools, MCPs, or autonomous workflows...

this repo is pure gold.

https://t.co/dxpG784fC5

[Teknium 🪽 @Teknium]

The legendary Tonbi with a new guide!

[ༀ C ⟁ O ⟡ N ✶ Q ☩ U ⟁ E ⟡ S ⚔ T = HELEN-OS @avalonaval]

@0ccultbot

[Dep @0xDepressionn]

Google just built an OS with 93 agents in 12 hours.
then ran Doom on it live onstage.

that's what AI autonomy looks like in 2026.

here's the open-source version.
300 agents. $0.60 per million tokens. runs on your laptop.

last month it rewrote an 8-year-old financial codebase.
185% performance improvement. 13 hours. no human guidance.

and it costs 8x less than Claude.

I broke down exactly how the swarm works. all 15 rules.

full guide in the article below.

[ericosiu @ericosiu]

https://t.co/SaUvnPnbuV

[AI Nerd @AINerdzz]

Microsoft Is Taking Screenshots of Your PC Disable These Settings Immediately