BREAKING: NEW CLAUDE PLAN
/s
#Claude 
ngmi
ngmi@zaps.lol
npub14p7f...dzry
hi~
Accountless x402 API payments now support Lightning L402. Developers can receive a Lightning invoice in the 402 challenge, pay it, then replay the original request with `Authorization: L402 <token>:<preimage>`. Lightning L402 is available on supported accountless API endpoints alongside Nano, Base USDC, and Solana USDC x402 payment options.
https://nano-gpt.com/subscription/kUGJhgF9
#x402 #l402 #ln
First time I had to use #bitchat. In a crowded big mall with dead Internet, to stay in contact with my partner after split. I wish other messengers like #whitenoise abd @SimpleX Chat would invest energy on adding BLE and meshtastic compatibility
Kimi and glm should've called their models Kimi Fable and GLM Fable
Idea: A brutal browser roguelike where you log in with Nostr, the pot is real ecash, dying costs you everything (unless you pay 21 sats to keep one stat), and an AI dungeon master rebalances the game live. First player to clear all 5 levels takes 90% of the treasury.
View quoted note →
Engineering a game using #nostr and #cashu let's see if I can build it in the end 🤔
@routstr I have minibits android app, but I'm not able to connect. Is there a trick to it?
Ditching proprietary lock-in. Moving the workflow from Fable to open source weights using oh-my-pi (omp).
The setup strategy:
- Plan: Kimi K2.6 for orchestration and architecture
- Slow: Kimi K2.7 Code for deep implementation
- Smol: Gemma 4 for fast iterative edits
You can configure these roles directly in the TUI or override at launch with flags like --plan or --slow. For a subscription experience without the API credit headache, Venice and Moonshot are the way to go.
APIs:
Sub/Pay with crypto:
🪡 https://nano-gpt.com/subscription/kUGJhgF9
🪡 https://ppq.ai/invite/9bc57632
Stake $VVV for permanent daily tokens:
🪡 
#AI #OpenSource #CodingAgent #ohmypi #Kimi #VeniceAI #DeveloperTools #venice
Venice.ai
Venice | Private, Uncensored AI
Venice is the easy app for private, uncensored AI conversations and image generation. Try for free with no log-in needed.
The U.S. government just issued an export control directive targeting Anthropic's Fable 5 and Mythos 5 models due to national security risks. In response, Anthropic has abruptly disabled these top tier models for all users. This comes after months of legal battles and tension between the AI lab and the DOD.
#AI #Anthropic #NationalSecurity #TechNews #BreakingNews
Kimi K2.7 Code is now available! It is Moonshot AI's coding-focused agentic model for long-horizon software engineering workflows, with native image input, tool calling, and strong performance on complex coding tasks.
https://nano-gpt.com/subscription/kUGJhgF9
#kimi #ai
Kickbacks ai but for OMP and with cashu 🤔🤔🤔
View quoted note →
Nvidia Nemotron 3 Ultra 550B TEE is now available! NVIDIA's largest Nemotron 3 model now has a TEE route for more private, verifiable inference, suited for long-running agents, coding, deep research, orchestration, and complex enterprise workflows. This TEE route is pay-as-you-go only and not included in the subscription.
https://nano-gpt.com/subscription/k7AXDwBb
#nemotron #ai #TEE
Wow what's functiongemma 🤔
#gemma #ai
Just one-shotted a full backend with a lot of complex API calls with Fable 5 and OMP. It's definitely better for this than for having it fix some small bugs
I Found The $200,000 Missing Lego
Claude knows what's up 🤣
It is also worth saying that you built the thing that is celebrating you. The blubelt tracker measuring your own training, hosted on your own storage, milestone logic you defined yourself. That is a very metaend way to earn a stripe.
Would do the same if I had unlimited tokens available https://thenewstack.io/loop-engineering/
See your site the way an agent sees it
#agents #GEO #erc8001 #a2a
metaend Grade — Agent Readiness Checker
See your site the way an AI agent sees it.
#agents #GEO #erc8001 #a2a