In 2013 Yonatan Sompolinsky and Aviv Zohar asked a simple question: what happens when blocks come faster than the network can agree on them? The answers became GHOST, then GHOSTDAG, then Kaspa. The next one is DAGKnight. The whole story 🧵 2013: GHOST. Instead of the longest chain, follow the heaviest tree, so fast blocks stop wasting security. Ethereum's uncle blocks grew out of this idea. 2018: PHANTOM, and its practical form GHOSTDAG. Blocks stop racing each other. Parallel blocks all land in a DAG and get ordered instead of orphaned. Nov 2021: Kaspa launches on GHOSTDAG. No premine, 1 block per second. 2025: Crescendo takes it to 10. GHOSTDAG still has one number baked in: k, set from a guess about the worst network delay. Guess low and security suffers. Guess high and confirmations slow down. DAGKnight drops the guess. It reads the real delay from the DAG itself and adapts. "Parameterless", in the paper's own title. @coderofstuff_ is building DAGKnight in the open, so we put the current code under load: • 70,000 simulated blocks • 7.2 million votes, each computed twice: by the optimized code and by a direct implementation of the paper's algorithm • 0 disagreements • and @kascovio already follows a private DAGKnight network reorgs and all Anyone can rerun it: simpa's baseline-debugging feature, on the public dagknight branch. github.com/kaspanet/rusty-ka… kaspa:native #Kaspa #KAS #GHOSTDAG #DAGKnight