Today I'm launching Femtos.
The idea is to allow anyone to build the app they dream of.
Give the weapons to any market participant: the arbitrage machine, the KOL buster machine, the token sniper, the bot that sees a whale's wall before it gets pulled, everything they could think about.
I dreamed about making this for about a year and a half, indexing hypercore, making everything available to the people,giving the same weapons to everyone. The big ones run their own nodes, their own indexers, their own quants. The rest trade blind.
Blockchains currently have an asymmetry of information problem, causing small palyers to be eaten by the bigger ones.
We aim to change this. HyperCore, HyperEVM, Elysium.
Every trade, every position, every order, every transfer, every token.
Making @Enter_Elysium fully transparent from the get-go.
Not snapshots. We run our own nodes and stream it as it happens, from Tokyo, next to the chain. One wallet, three layers, and more to come, one view: what it does on HyperCore, what it holds on HyperEVM, what it deploys on Elysium.
And it is built for agents as much as for humans, as I and we, believe the future to be run and developed by agents. REST, Websockets, MCP. An agent can take a key and pay on its own, why should we need a human to approve everything ?
And more, Femtos' vision isn't to stay on blockchains, we aim to provide data for agents all around the globe, whether it be news, SaaS, or anything.
Femtos aims to provide to anyone the way to access institutional grade data. For the people, and more.
Femtos.
https://t.co/2JClZR7xWg
Under load, Hyperliquid cancels were losing to orders signed after them. Thanks to @FrothyBeverage the issue went public, and the goated Hyperliquid team was fast to fix it.
We then worked together to go deeper and see whether anything else had broken.
Modifies were hit too, to the point where it is now faster to cancel and place a new order than to modify one. Here is the data.
Paper, code and data, so anyone can check it from Hyperliquid's public archive, or straight from the histograms we collected on the @femtosxyz node:
Paper link : https://t.co/QdwvPOSjev
Decentralization and open nodes let anyone check data like this independently. We see that as a strength, not a weakness, next to other perp DEXs. Thanks to Hyperliquid for making it all available.
Hyperliquid.
This has been a dark week for $NVDA.
Here are @HyperliquidX and @tradexyz numbers.
Sunday 20:12 UTC, the board probe goes public.
NVDA -3.4% in 28 min. Wall Street was asleep.
Wednesday, after the close, the WSJ reports the board wants him out. -10.6% in 35 min.
Today, Huang resigns. -20.2% from the first headline to the bottom.
Open interest: $111M -> $636M, a record.
Small accounts are long. Whales are short.
Answering my own question: no, makers didn't just set the flag.
Losing cancels signed before the 8 bps IOC that filled them (40k+ tx/s), counting only cancels sent WITH the fast flag:
Nov 8-9: 30.6%
Nov 10: 12.9%
Nov 11: 1.4%
On Nov 9, nearly every losing cancel (98.5%) already had the flag, and the share of cancels sent without it barely moved (1.8% → 1.3%).
So the improvement came from Hyperliquid's side, in two steps, which fits their "optimizations already implemented". What's left today is mostly cancels without the flag.
Alright, here are the measurements:
Your result was real: 30.5% of losing cancels were signed before the 8 bps IOC on Nov 8–9. Same method gives 13.4% on Nov 10 and 3.8% today.
Today almost every losing cancel was simply signed after the taker, a median 621 ms after. Of the 610 signed first, 504 had no fast flag. Without it a cancel waits like a taker who pays no fee.
Matches what HL posted today. 78 cases have no known cause, happy to share them.
Open question: what changed between Nov 9 and Nov 10? The last network upgrade was Nov 7, before all three windows, and the share still dropped from 30% to 13% in a day.
Alright, here are the measurements:
Your result was real: 30.5% of losing cancels were signed before the 8 bps IOC on Nov 8–9. Same method gives 13.4% on Nov 10 and 3.8% today.
Today almost every losing cancel was simply signed after the taker, a median 621 ms after. Of the 610 signed first, 504 had no fast flag. Without it a cancel waits like a taker who pays no fee.
Matches what HL posted today. 78 cases have no known cause, happy to share them.
Open question: what changed between Nov 9 and Nov 10? The last network upgrade was Nov 7, before all three windows, and the share still dropped from 30% to 13% in a day.
Ladies and gentleman, meet @AscendLaunch co-founders, @CryptoTomYT and @analyzationz of @VersusFNF and @DigiWorldd, also known as 'igotruggedoncashcat'.
They are the attackers behind the coordinated efforts against my account and the @MontraXYZ project and token, who fooled the entire space while masquerading as a legitimate project on @Enter_Elysium.
The main attack vectors used in an attempt to undermine Montra were the weaponization and leverage of their fomo standing, social engineering on @X activating paid coordinate FUD campaigns, and manipulation through coordinate buying and selling of Montra assets.
Descend below to see the timeline of events 👇🧵