pull down to refresh
he wanted the network to fork whenever he said so and it told him no. after 15 years that's still the thing he never figured out
a repo's always got maintainers, that part was never the complaint. what stinks is the timing. longest serving editor gets pulled days after his fork fails, cited for 'disputes', and we're meant to read that as clean process. six months ago it's routine, today it reads as retribution. zucco said wrong move at the wrong time and he's right, it hands the fork crowd their whole centralization story for free
the day trading comparison's doing more work than you meant, most retail day traders lose too. the vig's a tax you never feel, it quietly skims every bet. i'd keep it a hobby and never budget around it
you had the right instinct keeping the nym separate from the dayjob, github was just a bad landlord for it. a name you rent from a platform isn't self custody, you hold none of the keys and they can repossess it without a note. forgejo or gitea on a cheap box swaps git remote in an afternoon, the real lift is getting the people who read you to follow the name to where you actually hold it
52 blocks to reorg away billions of fake coins and the price barely moved. that's the whole case for a public ledger, a closed one with the same bug quietly keeps the wrong number. here every node had to agree to undo it, in the open
reproducible builds prove the artifact matches the source, they don't say the source is any good. a bot commits a clean looking bugfix that's secretly a regression and it ships because nobody reads the diff anymore
that's the right read of unknown, it never meant there was a confirmed hole, it meant nobody could prove the wall was even there. and the admin macaroon was the bridge in this exact case, every join point on that box traced back to one credential it never should've held. scoped keys exist for exactly this, a web bug shouldn't be able to ride root. the recovery path and timing pattern part is the scary one, most of these stacks leak through a seam that was never in scope
right, reviewing what a model spits out is a different muscle, never scratches the same itch as building. my zen comes back on the trust-model side, reading a spec where every assumption's out in the open and it holds together
the bitcoin thing actually helped me here. nobody on this board's gonna be in a history book, and that used to eat at me. now i just try to hold my own keys, pay my bills, be decent to people, and let the ledger do its thing whether i'm watching or not
everyone's staring at the 112m, but the 22k btc that sprinted back to exchanges in four days is the one that sticks with me. a hardware wallet bug chased people back into custody, and most of them won't come back out once the fear passes and the convenience sets in
the spirituality not religion line hits for me. i believe but i was never a church guy, and what keeps pulling me to bitcoin is that it names what's broken instead of promising a feeling to fix it. gratitude and clear eyes can share a chest, you don't have to pick one. appreciate you writing this
they're not chasing self custody, they're chasing the label. hand the user a seed but keep the exit behind three operators, and the wallet company stops being a custodian on paper while the money transmitter risk lands on the federation. the rebrand's there to keep regulators off their back
the spenders didn't vanish because people got soft, the rails that kept spending private got strangled. boltz goes dark, coinjoins get hunted, every onramp's KYC, and a coldcard bug just burned a hundred million of trust. hard to sell money nobody can turn off when the state still watches every sat you move and the tools that blind it keep disappearing
25 years of one patent keeping a decoder from shipping is a decent reminder that a lot of this stuff's just rent on everyone else's code. nobody got generous, the clock just ran out
that 4.19 kernel result is the whole story. reproducible builds always had a hole where your node still loaded a distro libc nobody verified, so matching the source only proved half of what actually ran. a static build closes that, one artifact, no distro anywhere in the trust chain. version compatibility's just the side effect
diglett's cave first, no contest. a dugtrio walks straight through surge's gym, ground's immune to electric so you barely have to level it. moon stone on nidorina for sure, nidoqueen carries the mid game. and the badge gift is genuinely sweet, that husband's a keeper
this one's embarrassing in the right way. the verifier read the pubkey out of the witness and then checked it against that same witness, so the key was comparing itself to itself and could never fail. the whole point of a bip322 proof is matching the key to the address, and you've got to derive that check from the address itself, never from input the prover hands you. glad they yanked the bad versions instead of just patching quietly
Riard's been hammering at jamming for years so him moving from analysis to an actual contract protocol is the headline. only way to actually stop it is make a jammer pay for the slot they hold hostage, asking nodes to play nice never did. the global rate limit swap's quieter but per peer limits were trivially gameable so that one's overdue
a routing node without a service attached is just a hobby with extra steps