Veronica's share.google/1Xy6KRvJDqeVzlaZr One of my fav and tres leches at magnolia post
ivc
Android on iOS! The best use case I see is basically getting access to games (which I don't play) however, getting a separate space for installing apps from Aurora / F Droid is quite nice.
In other news, surp.ivc.lol is now available in the hermes plugin directory to one-click install on desktop.
surp.ivc.lol — cheapest LLM API on the internet | x402 pay-per-request
Pay-per-request LLM inference at live cheapest prices. We watch the Surplus Intelligence marketplace and route every request to whichever model is cheapest right now. USDC on Base, no account needed.
surp-hermes-x402-llm · Plugin Catalog | Hermes Agent
Marketplace-aware LLM routing for Surp — live prices for all models, provider TPS and trusted flags, mid-chat seller switching, secret masking, and approval-gat
wowow
My test with FATLINE.ink Machine: AWS Instance type: t2.xlarge Disk: 2050GB NodeJS didn’t come built in the instance, so had to install node js. curl -o- raw.githubusercontent.com/nvm-sh/nvm/v0.40.8/install.sh | bash # 2. Reload your terminal session config to start using nvm source ~/.bashrc # 3. Install the latest version of Node.js nvm install 26 # 4. Verify that Node.js and npm (package manager) are successfully installed node -v npm -v  Supercool animation intro… but took a loooong time to render on the vps.. like 5 fps. Docker wasn’t installed on the VPS, so it just told me that the node can’t be installed, I ctrl+c’d, then installed docker and re-ran the script … If it could auto-install docker, that would be great. After installing docker in the ubuntu instance, the recommendation was super, I didn’t have to go back into docs to refer to the commands, I simply did “fatline node farcaster” and it was on it. Currently downloading shards.
Tried out @webthe3rd.eth 's webthethird.github.io/snapstart/#quickstart and it was a pretty chill experience! #vouch
Your iPhone is sitting there doing nothing. Backburner puts it to work. 🔥 Plug it into your Mac and it becomes the second half of a local LLM — Mac runs layers 1-40, iPhone GPU takes the rest. 27B models, 184k context, 400+ tok/s prefill. Fully offline. Fully yours. No cloud. No subscription. Just the hardware you already own, finally pulling its weight. github.com/StayLameBro/backburner
You're all caught up
