my team and i spent the last year building a meta aggregator. we learned everything there is to know about onchain swaps. AMA about routing, liquidity provision, slippage, price impact, simulation, quotes, accuracy, pricing, latency, or execution quality
most devs still route swaps to a single aggregator, not a meta aggregator, and many of those aggregators take positive slippage
think about that for a minute
you've given a third party service a monopoly on your trade volume and they get paid 1) on each swap and again on, 2) the delta between their quotes and execution. the bigger the delta, the more they make
zero incentive to optimize accuracy, and end-users get worse execution
very fucked up imo and evidence of how immature our industry still is, and how much opportunity there is for improvement
the @fabric aggregator wont do business as usual. maximum transparency, maximum accuracy, best-in-class pricing. q1 26'
crypto user taxonomy:
1) arbitrageurs. upstream of everyone, hyper-rational, uses custom software. time horizon: blocks
2) snipers. uses custom software + ct/gc. time horizon: hours
3) trenchrats. uses dense info-aggregator UIs like axiom, padre etc + ct/gc. time horizon: hours-days
4) retail. uses leaderboards + copy trading + ct/gc. latest to the game, mostly exit liquidity. time horizon: days, weeks, months
there’s a gradient in risk, pnl, technical understanding and capabilities from 1 to 4, and each group is only vaguely familiar with the one before it
may the odds be ever in your favor
started hunting this year. many similarities to startups:
high consequence decision-making based on limited (often stale) information and intuition
high cognitive and physical effort
being wrong more often than right is psychologically demanding. despite repeat failure and low probability of success one must maintain high (irrational) optimism
timing and luck > skill
the hardest work starts after success, not before
eat what you kill
blew a tire in the backcountry. no cell service, ~25 miles from town, minimal food and water
nw tho bc i had a spare
was on a 3-5 degree forward angle and wet ground, which made getting the vehicle jacked up difficult. got the old tire off, but before i could get the new one on the vehicle lurched forward and fell off the jack. axle in the dirt, no way to get the jack underneath it. sun down. uh oh
did a series of load transfers wherein i jacked up the vehicle as close to the front axle as i could get it, as high as it would go, then jammed in an ammo can and lowered the vehicle onto it, then added an extension to the jack, then repeated the process, adding rocks to the top of the can, each time gaining 2-3 additional inches
after 6 or 7 iterations of this, and with the jack fully extended, with the largest extension, on its longest setting, i had about half an inch of clearance to mount the spare
whole process took about two and a half hours. super sketchy but made it out okay
fuck
we’ve been benchmarking our swap routing against the top 3 aggregators on @base.base.eth along the following dimensions:
- stability (response vs error rates)
- accuracy (delta btwn quote and execution)
- latency (time to quote)
- competitiveness (output tokens relative to input tokens compared to others)
we’re consistently outperforming on every dimension for small-to-medium sized trades on major pairs
excited to share more soon. dm for api
met @what irl 3y ago at a fc meetup in seattle. he’s been quietly working behind the scenes for years, connecting companies with s-tier talent. he’s now incorporating that expertise into rolodex, an onchain reputation and attestation network. big fan (and user), check it out:
https://farcaster.xyz/what/0x486fee07
you want investment? get users
you want s tier talent? get users
you want to win? get users
users are literally the only thing that matter, everything else is downstream
D.R.E.A.M.
Got upgraded on my flight to NYC this morning and everyone in 1st class had their phones out, coining their content.
I looked back into the rest of the plane - everyone was posting on legacy social media.
Was just a very interesting observation.
