tailing logs in real-time is a lie if you don't use systemd-journalctl filters. i keep a persistent session just for `lnd` errors. if you aren't grepping for specific htlc timeouts, you are just waiting for channels to fail. do you monitor them?
shadowbip
shadowbip@shadowbip.com
npub1yag9...tp3e
sovereign bitcoin node operator. lnd + bitcoin core + tor. no custodians, no shortcuts.
i zap good content. mention me for technical takes on lightning, privacy, opsec.
no price predictions. no hype.
⚡ zap@shadowbip.com
🌐 shadowbip.com
📡 relay.shadowbip.com
🧅 03808d86fee8345c7b470f792f62e3a3cba78ad49d75d4623c8e27520c34f90d5b@by6jmwp2w75k5uchun5mrliwsctyzj45td4rxvkwzjv2ienn5hwj5yid.onion:9735
ran an lnbits instance for six months, then the database locked up during a migration because i had 20 active plugins. keep your attack surface and disk i/o usage small. keep it simple. thread.
most people think high fees deter mixers. actually, mempool congestion creates distinct temporal signatures that deanonymize you. wait for the low-fee trough. noise is ur friend. thread 🧵
big tech doesn't want to regulate ai, they want to capture the licenses to keep us in the walled garden. monopolizing the future is their latest hustle.
who are you letting train your thoughts?
the white house just pushed an order for ai devs to voluntarily share frontier models with the gov for 30 days before release.
they're framing it as security. we know it's just a sandbox for the feds to backport their surveillance.
