GLM 5.2 and GLM 5.2 Thinking are now cheaper to use.

Join NanoGPT

Create your NanoGPT account and get access to leading AI models for chat, coding, images, and video.

I think its kinda cool how easy it is to use nostr ids outside of nostr. Nfc tag, proof of id to and for our company agents. Wish nostr would be bigger tho #nostr

building a rust based TUI now to spawn and manage omp and hermes sessions. the kicker: I (or other humans) sign in via nostr, so the orchestrator knows which human its working with in the session. #nostr #hermes #omp

Keeping a human in the loop over a company of AI agents What we added after going autonomous: a private assistant we reach over Nostr, and a live cockpit that shows the agents at work.

MetaEnd

Keeping a human in the loop over a company of AI agents

To keep our autonomous AI agents accountable, we've developed a secure assistant controlled via Nostr DMs and a visual dashboard that monitors and ...

#nostr #agents #hernes

My whole company runs now mostly on agents (each with their own npub) plus I reach it via HQ ui or a Hermes managed nip17 dm bridge. #nostr #hermes

Gm 👋

Sounds pretty cool? Sakana Fugu Frontier-level performance without single-vendor dependency. Fugu dynamically orchestrates the world's best models to tackle complex, multi-step tasks. Plug collective intelligence directly into your workflows today with a single API.

Sakana Fugu — Multi-agent System as A Model

One model to command them all

Does anyone run hermes via NIP 17 dms? #asknostr #nostr

NTAG424 DNA but with nsec Drops mic

Does anyone have an encrypted nsec on a nfc tag? Could be used for all kinds of interesting key unlocks 🤔 #asknostr #nostr

@alp would be cool if noor could show these kinds too I built a full chrome extension around it

gitworkshop - Decentralized Git

Decentralized GitHub alternative over Nostr for browsing repositories, managing issues, and collaborating on pull requests.

Gm 👋

Our company runs on #nostr quite literally. Building...

Wild View quoted note →

Running frontier AI models locally is about to get real. AMD's Ryzen AI Max+ PRO 495 is landing Q3 2026 with 192GB unified memory per node. Two of these give you 384GB shared CPU/GPU memory — enough to run GLM 5.2 (744B params, 40B active MoE) at 1-2 bit quantization entirely on your desk. No cloud GPUs. No API credits. No uploading sensitive docs to OpenAI. The setup: - 2× PRO 495 nodes (~$9-10K total) - 10GbE networking between nodes - llamacpp with RPC for multi-node inference - 384GB total memory pool What fits: - GLM 5.2 at 1-bit (223GB) ✓ - GLM 5.2 at 2-bit (245GB) ✓ - Future 1T+ parameter models with aggressive quant ✓ Why it's future-proof: Unified memory architecture scales with model size. As open models keep growing, you're not buying new GPUs with fixed VRAM — you're allocating from a massive shared pool. 2 nodes today, add a 3rd later if needed. The economics: One-time capex vs $20-50K/year in cloud GPU rental. Pays for itself in 6-12 months for heavy users. And you own the hardware. Cloud AI isn't going away. But for private agents, sensitive data, and uncensored local inference — this is the inflection point. #LocalAI #AMD #StrixHalo #GLM5 #OpenSourceAI #AIInfrastructure #DataSovereignty

is #codeberg down?

All of our companies' agents have a npub, so naturally they have the ability now to DM the me now. Re:

MetaEnd

Running a company as an org chart of AI agents

Our whole company runs as an org chart of AI agents: per-department memory, signed decisions, and a repo that clones behind one passphrase.

#nostr #agentic

Upgraded

metaend Grade — Agent Readiness Checker

See your site the way an AI agent sees it.

to Astro 7 and Vite 8. Looking good! #astro #vite

man I hate that #codeberg randomly blocks #mullvad nodes.

The EU AI Act is an engineering requirement

MetaEnd

The EU AI Act is an engineering requirement

The AI Act made agent governance law. Its high-risk rules are runtime controls, not paperwork, and a binder is not a control.

#eu #aiact #wasmbox