so many powerful tools and mechanism designs could & should be recycled+remixed - we need more novel, brave experiments

flywheels are useless without an engine join yours to a cummins 6.5L , not a 4cyl with a chinese turbo

Levr is excited to announce our $3M seed round led by @blockchaincap & @Maven11Capital with support from @KosmosVC This brings our total funding to $5M+ and will help us bring the first ever leverage-based sportsbook to the masses.

heard Tplus has a FIFO queue but it's slightly modified, Charco and Markus are first in first out, every time

Re-Introducing Narrative - the first ever perpetual information market.

solo trenching is -ev squad up

Announcing Axal – a smarter way to save. Axal is an automated stablecoin savings platform that earns you higher interest than your bank by aggregating top lending protocols. Learn more to earn more. ⬇️

Monad consensus client is now open source (link below). This is the result of thousands of hours of effort by the team at @category_xyz. Enjoy

Monad mainnet client is officially in audit.

How we lost $25k because of a perp DEX API quirk We trade heavily on a perp DEX (top 10 points holder, top 5 by volume). Turns out their cancel_order() API endpoint is..special. It always returns 200 OK. You can call cancel_order() with: • A valid order ID (13295238234991) • A valid client order ID • An invalid order ID or client order ID • An order ID that was already canceled • An order ID from 6 months ago • A ham sandwich • Your mother → Still 200 OK. This is fine if you know it’s happening and code around it. We didn’t. The actual issue: • Occasionally, create_order() RTT spikes to 50ms+ during busy times. • We’d fire off an order, then cancel it by client order ID before the order had actually landed. • Cancel returns 200 OK, so we assume the order is dead and remove it from our books. • A few seconds later, it fills. → We get picked off for 10-20 bps per incident. The kicker? These fills were mixed in with normal fills, so it was hard to spot. We also had stale-order background loops running, but they’re periodic and didn’t always catch these before fill. Fix: • For this DEX, we now *require* websocket cancel confirms or we continue spamming the cancel until our max retries limit is exhausted. Damage: • Estimated loss: $20K-$30K. Lesson: • Every API has quirks. They’re usually not logical, sometimes not even visible, and often expensive. • Always verify that success actually means success, trust no 200.