Solana, What’s Right And What’s Wrong
Summary
TLDRIn this engaging conversation, Eugene Chin, co-founder at Ellipsis Labs and known as zerox on Twitter, shares his insights on the crypto space, particularly focusing on high-performance DeFi products like Phoenix on Solana. With a background in high-frequency trading and product engineering, Eugene discusses the evolution of his thoughts on DeFi, the challenges faced by liquidity providers, and the potential of Solana's performance to revolutionize the trading experience on-chain. He also addresses the importance of balancing technological innovation with core values in the blockchain community.
Takeaways
- 🌐 Eugene Chin is a co-founder at Ellipsis Labs, focusing on high-performance DeFi products like Phoenix on Solana.
- 🤖 Eugene's background includes high-frequency trading and product engineering, with a foundation in mathematics.
- 💡 Eugene's transition to crypto was driven by a desire to improve upon the inefficiencies he observed in traditional DeFi liquidity provision models.
- 📈 His perspective on AMMs has evolved over time, recognizing their value as a liquidity bootstrapping mechanism, especially for new tokens.
- 🔄 Eugene sees a future where AMMs serve as a liquidity backbone, complementing more sophisticated order book systems.
- 🚀 Phoenix is a proof of concept for demonstrating that active liquidity can outperform passive liquidity in an environment like Solana.
- 🛠️ The Solana ecosystem has room for improvement, especially regarding its fee model and network layer issues.
- 🔧 Solana's priority has been on performance and user experience, sometimes at the expense of incentive compatibility.
- 📊 The success of Solana's token (SOL) is tied to the platform's ability to support a robust DeFi ecosystem, rather than its market cap ranking.
- 🌟 In the long term, Eugene envisions a future where Solana can facilitate the trading of non-crypto assets, integrating with traditional financial systems.
Q & A
What is Eugene Chin's background and how did he get involved in the crypto space?
-Eugene Chin is a co-founder at Ellipsis Labs, where they build high-performance DeFi products like Phoenix on Solana. He has a background in high-frequency trading and product engineering, with a focus on math during his studies. He entered the crypto space during the pandemic, initially spending time looking at DeFi and its developments.
What were the main issues Eugene saw in the current DeFi market that led him to start his company?
-Eugene was not satisfied with the state of DeFi, particularly with the XY=K model and its replacement with Uniswap V3's concentrated liquidity. He felt that these models did not make sense from a trading background and seemed inefficient compared to traditional market-making practices.
How does Eugene view the role of AMMs in the DeFi space?
-Eugene sees AMMs as a liquidity bootstrapping mechanism suitable for new coins, but believes they are not designed to be the ultimate solution for liquidity provision. He thinks AMMs serve a purpose but are not as efficient as traditional market-making models, especially for more liquid, large market cap assets.
What is the main goal of Phoenix on Solana according to Eugene?
-The main goal of Phoenix on Solana is to prove that active liquidity can be competitive with passive liquidity in an environment like Solana. It aims to provide better liquidity than centralized exchanges by building a product that market makers can use to provide really good liquidity to end users.
What are some of the challenges that Eugene sees with running a high-performance trading platform on a blockchain like Solana?
-Some challenges include the need to pay gas fees for every on-chain order placement or cancellation, which can add up quickly for active market makers. Additionally, the environment is significantly different from centralized venues, requiring a new primitive for on-chain liquidity provision.
How does Eugene view the future of Solana and its ability to support high-performance applications?
-Eugene believes that Solana has been ambitious and successful in its mission, offering high throughput and credible decentralization. However, he acknowledges that there are long-term issues that need to be addressed, particularly on the engineering and research sides, to ensure the platform's continued growth and success.
What is Eugene's take on the Solana fee model and its impact on user experience?
-Eugene criticizes the Solana fee model, stating it does not make much sense as it does not effectively throttle demand or provide a good user experience. He advocates for a model that offers better UX, such as one with a maximum escalation limit and a clear transaction fee structure.
What are Eugene's thoughts on the recent changes to Solana's MEV (Miner Extractable Value) situation with the removal of the mempool by Jto?
-Eugene views the removal of the mempool by Jto as a positive step towards prioritizing user experience over incentive compatibility. However, he believes this is not a permanent fix and that a better long-term solution is needed to support the on-chain trading economy.
How does Eugene see the role of blockchain in the future of finance and beyond?
-Eugene believes that blockchain, particularly through DeFi, is well-suited for disrupting the traditional financial system by offering permissionless innovation and access. He sees a future where blockchain technology can onboard a significant portion of non-digital assets and provide better financial outcomes for users.
What are Eugene's thoughts on the economic security of L1 tokens and their role in the ecosystem?
-Eugene thinks that the economic security of L1 tokens is overrated. He believes that as long as there is an honest minority that can fork out malicious actors, the economic security argument does not hold much weight. He also suggests that the focus should be more on building a platform that optimizes for applications rather than token price.
How does Eugene view the competition between Solana and other high-performance chains?
-Eugene believes that Solana has an incumbency advantage due to its early adoption and network effects. He thinks that while other high-performance chains are making valuable contributions to the space, Solana's head start and continued innovation give it an edge. However, he also acknowledges that no blockchain is in a finished state and that ongoing innovation is necessary for the growth of the entire crypto industry.
Outlines
🤝 Introduction and Background
The video begins with an introduction to Eugene Chin, also known as ZeroX Trader on Twitter, who is a co-founder at Ellipsis Labs. Eugene shares his background in high-frequency trading and product engineering, and how he transitioned into the crypto space during the pandemic. He discusses his dissatisfaction with the DeFi space and the motivation behind starting his company in late 2022. The conversation delves into the inefficiencies of the traditional DeFi liquidity provision models and the need for a more efficient order book system.
💡 Evolution of DeFi and AMMs
Eugene shares his perspective on the evolution of automated market makers (AMMs) in DeFi, from the XY=K model to the introduction of concentrated liquidity AMMs like Uniswap V3. He discusses the challenges faced by AMMs, particularly in terms of providing competitive liquidity compared to traditional high-frequency trading firms. Eugene also highlights the role of liquidity mining rewards in sustaining AMMs and his changing views on their utility, especially as a liquidity bootstrapping mechanism for new tokens.
🚀 The Vision for Phoenix and Solana
Eugene outlines the vision behind Phoenix, a product developed by Ellipsis Labs, which aims to prove the viability of active liquidity provision on Solana. He discusses the challenges of providing liquidity on-chain, the importance of profitability for market makers, and the need to improve the underlying primitive for active liquidity. Eugene also shares his thoughts on the ambitious goals of Solana's founder, Anatol, and the potential for a shared global state in DeFi.
🌐 Solana's Current State and Challenges
The discussion shifts to the current state of Solana, with Eugene acknowledging the platform's high throughput and credible decentralization levels. He highlights the various challenges Solana faces, including engineering issues related to the networking layer and the need for long-term research solutions. Eugene also touches on Solana's fee model, incentive compatibility, and the recent removal of the mempool by Solana's leading node operator, Jeto.
🔄 MeV, Blockchain Speed, and Decentralization
Eugene delves into the concept of miner extractable value (MeV) and its implications on Ethereum and Solana. He discusses the differences in how each platform has addressed MeV, the potential for faster blockchains to reduce MeV, and the trade-offs between speed and efficiency. Eugene also shares his thoughts on the ideal level of decentralization for a blockchain, emphasizing the importance of aligning with the platform's goals and use cases.
🌟 The Future of Solana and DeFi
Eugene envisions a future where Solana and DeFi can compete with traditional finance, offering better trading experiences and guarantees of transparency and auditability. He predicts a future where a significant portion of non-digital assets are onboarded onto crypto rails. Eugene also reflects on the broader potential of blockchain technology beyond trading, acknowledging the ongoing innovation and the need for a pragmatic approach to evaluating the value of blockchain applications.
🤔 Final Thoughts and Outlook
In the concluding part of the conversation, Eugene shares his personal views on the future of Solana and the broader DeFi ecosystem. He discusses the importance of continued innovation, the potential for new high-performance chains, and the value of differentiating Solana from its competitors. Eugene emphasizes the need for blockchains to evolve and improve, while also maintaining a focus on applications and user experience.
Mindmap
Keywords
💡Cryptocurrency
💡DeFi (Decentralized Finance)
💡High-Performance Trading
💡Ellipsis Labs
💡Solana
💡AMM (Automated Market Maker)
💡Liquidity Provision
💡Market Maker
💡Product Engineering
💡XY=K Model
💡Uniswap V3
Highlights
Eugene Chin, co-founder at Ellipsis Labs, discusses the development of high-performance DeFi products like Phoenix on Solana.
Eugene's background in high-frequency trading and product engineering led him to crypto during the pandemic, where he observed inefficiencies in traditional DeFi liquidity models.
The transition from XY=K to Uniswap V3's concentrated liquidity model was a significant shift that raised questions for Eugene about the profitability of simple liquidity provision.
Eugene notes that traditional market making is highly competitive and sophisticated, questioning the viability of simple on-chain liquidity models.
The concept of AMMs as a liquidity bootstrapping mechanism is highlighted, recognizing their value in facilitating new coin markets despite their limitations.
Eugene's vision for Phoenix is to prove that active liquidity can outperform passive liquidity in an environment like Solana.
The importance of addressing the fee model and providing a better user experience is emphasized for the success of on-chain trading platforms.
Eugene discusses the need for Solana to improve its networking layer to address engineering challenges and enhance the platform's scalability.
The removal of the mempool by Solana is seen as a positive step towards prioritizing user experience over incentive compatibility.
Eugene shares his views on the future of DeFi, predicting a shift towards on-chain trading that offers better user experience, liquidity, and transparency.
The potential for on-chain trading platforms like Phoenix to eventually handle non-digital assets and disrupt traditional financial systems is explored.
Eugene addresses the criticisms of Solana from the Ethereum community, highlighting the differences in values and priorities between the two blockchains.
The discussion touches on the importance of continuous innovation in blockchain technology, emphasizing that no platform is in a finished state.
Eugene's perspective on the role of blockchain in finance is that it should be fast, efficient, and permissionless, disrupting traditional systems to deliver better outcomes.
The interview concludes with Eugene's thoughts on the future of Solana and the broader blockchain ecosystem, emphasizing the need for constant improvement and adaptation.
Transcripts
[Music]
hey guys I'm here today with Eugene chin
um known on Twitter as zerox [ __ ] Trader
uh Eugene is somebody who's been in
crypto for a little while now has very
good thoughts I think on the space and
also has become somewhat of a personal
friend so um Eugene you want to give a
little background and uh tell everybody
what you do thanks for having me on G uh
yeah I'm a co-founder at Ellipsis
Labs we build product in di specifically
high performance uh di products like
Phoenix on
salana I have a pretty standard
background spent some time in high
frequency trading more time in product
engineering studied math in school came
to crypto during the pandemic spent a
lot of time looking at dii and uh yeah I
was not too happy with what I saw and so
my co-founder and I started the company
in late 2022 and here we are
nice uh so was Phoenix or Ellipsis Labs
which is your company uh was this the
first work you did in dii or were you
doing stuff before um before you founded
this like were you actually like writing
code before that or no depends what you
mean by work uh you know in crypto we
call ourselves like independent
researchers which usually just means
unemployed so that was me for quite a
while in crypto sort of exploring and
trying to figure out what's going on got
it yeah I think that's what I would put
myself in that category as well as well
wellth I don't know how much research I
do but um okay interesting so you you
did you sp you spent a little time in
high frequency trading um I would assume
this is what sort of brought you to
looking at how Marcus function in Defi
and eventually to the conclusion that
defi needed a very uh efficient order
book is this like would would you agree
with this and then also what was sort of
like your your view on how trading was
occurring um yeah like the structure and
then like how you thought that you could
like
potentially present a solution yeah so I
came to crypto in 2021
and spent a lot of time looking at what
was going on in Defi and there was this
whole XY equals k thing uh which I guess
very soon after I got into crypto
was uh replaced with unisat V3 the
concentrated
liquidity amm and this is just like so
foreign to someone with a little bit
more of a trafi background and you know
there were all these papers saying
things like if you provide liquidity on
the XY equals K amm you're going to make
money and it sort of just did not make
very much sense to me and then as all
the coins were going
up uh I kept thinking okay like I must
have something wrong here um all these
positions are making tons of
money uh despite being like really
really
unsophisticated liquidity providers
because coming from the uh the more
traditional side electronic Market
making is extremely extremely
competitive industry where you have
these high frequency trading firms
electronic market makers who have been
in the business for decades at this
point spent tens of billions of dollars
on R&D to compete in providing really
really good liquidity which means tight
spreads and uh lots of
depth and this one click XY equals k
liquidity provisioning just seemed like
there's no way it should be able to
compete with
that um and I think that's pretty much
born out to be true it became like much
more clear when the prices started going
down uh and you still see
this in crypto today where on chain
especially with amm the spreads tend to
be wider and the liquidity providers
tend to not make money tend to lose
money in fact uh often times they have
to be compensated with uh liquidity
mining Rewards
um and I think I have a lot more respect
for the amm today than I did maybe like
a year or two ago uh I think it really
makes sense as a liquidity bootstrapping
mechanism where the onboarding cost for
a market maker to start providing
liquidity on a new coin can be
relatively
high and this does give like a one-click
way for someone who is not particularly
concerned with making money on providing
liquidity and more concerned with um
liquidity existing at all for example a
team that is deploying a token um so I
think there is a real use case there um
but it does seem relatively Fringe uh
and if defi is to compete with grafi
which I think should be the goal I don't
think the goal should just be defi for
people who don't have access to binance
or defi for trading coins that you know
coinbase won't list
um so for defi to be able to compete
with trafi you actually need to also
compete on product you need to compete
on efficiency you need to build really
good products for market makers so they
can provide really good liquidity to the
end user yeah and you've been like
pretty vocal about amms I would say um
at least from what I've seen on your
Twitter you've long like been sort of
critical of the model or at least like
you know you you voiced um what you've
just said now uh so you say that you've
like sort of had a change of opinion but
I want to ask you like specifically do
you think that amms
are I don't want to say useless but like
not nearly as useful when it starts to
come to like assets that are not on this
long tail like do you see a future where
amm are useful for these like very very
liquid large market cap assets or do you
think that like at that point like
you're saying like we're going to just
be trading exclusively on order books in
that capacity I think they're likely to
continue existing in some capacity more
as uh like the liquidity of Last Resort
the backbone
liquidity
um where you know there are times where
market makers don't want to provide and
so generally the way this works on
centralized venues whether it's in trafi
or in crypto is someone whether it's the
exchange or a token team is going to
strike a deal with the market maker
to guarantee that there's going to be
some liquidity at some depth uh like 99%
of the time or almost 100% of the time
even when it might be not profitable uh
directly on the market making and maybe
you can replace that with some sort of
amm like
construct uh but the amm is really not
designed to be the Super General thing
in my mind it's designed to have some
form of completely onchain liquidity in
an environment where
one there's no guarantee there's going
to be sophisticated market makers who
want to show up and onboard onto your
system and two when the whole blockchain
like the entire Global state does call
it 15 or 20 transactions per second it's
just not conducive to any sort of active
Market making activity so you do need a
new primitive for that environment so
for example I think like order books
will never make sense on ethereum Main
net unless ethereum main net someday
scales to you know 100,000
TPS yeah and that's long like that's
basically been the rationale at least
like for since the Inception of amm is
that like you realistically can't have
these like very high frequency order
books on top of a blockchain that you
know uh moves every 12 seconds but um so
it to to that end um what did you like
what were you looking for when you uh
started like building out Phoenix like
what was in your mind what you needed to
do and then also we can kind of like I
suppose move on to like salana as a
whole and like how you felt that salana
specifically would like enable what you
were looking to do yeah in my mind
Phoenix is like very far from a
completed product it's just a naive
vanilla order book implementation on top
of
salana with spot trading only
and yeah it's really a proof of concept
where first we need to prove out the
product thesis that active liquidity can
be competitive with passive liquidity on
an environment like salana and it's not
obvious that this is going to be the
case because providing liquidity on
chain is very different from providing
liquidity on a centralized
venue I think the most important piece
is you have to pay gas the market maker
has to pay gas anytime they want to
place an onchain order anytime they want
to cancel an onchain order and yeah that
number might be very small but if you're
placing and cancelling orders very
rapidly uh you know those cost just add
up and so you need enough volume to to
make up for that and then the second
piece is these active market makers
really care about being profitable
they're like sophisticated entities they
know how to measure the profitability of
a strategy they're not looking at
you know some big green number on the
screen that says 800% apy and thinking
like okay that's like probably good to
go um so the market liquidity providers
are more sophisticated and more
sensitive to profitability than a
passive liquidity provider and the
environment is significantly worse than
it is providing on a centralized venue
so it's I think like not obvious a
priori that active liquidity is even
viable on chain and so we had to prove
that at first I think at least in the
current market conditions or the market
conditions over the last year which have
obviously changed quite a
bit
uh yeah the the market makers make money
on Phoenix they make uh quite a bit of
money and are able to provide tighter
spreads than the the amm now now that
like we've proven that active liquidity
can outperform passive liquidity you
know we still need to make the core
underlying primitive better because the
goal isn't to provide better liquidity
than Unis swap uh the goal is to provide
better liquidity than than coinbase or
binance or in the long run like the New
York Stock Exchange yeah and this is
something that anatol the founder of
Sona is like very fixated on I mean this
has been like a lot like one of his uh
selling points on SOL he says like I
think we can create this Global state
that competes with Finance competes with
coinbase first of all is this
realistic um second of all just out of
curiosity a little bit on your last
question you said that market makers are
profitable on Phoenix do you think this
is a function of the flow being
primarily non-toxic I don't know I
actually don't know if it is or I mean
is this like a long run profitable
Market making strategy for them or is
this like you said under current market
conditions so those are kind of two
questions but you can um answer how you
want I'll answer the second one first I
think it is always very difficult to
make a long run general statement uh
which is why is going to hedge with
something like in the current market
conditions there's a lot of non-toxic
flow on salana today although the market
makers were still profitable when the
volumes were 1% of what they are today
and there definitely are uh shall we say
toxic uh actors uh I don't mean toxic in
a derogatory way I mean it just in the
nature of the flow
and you know they're making money as
well and as a market maker your job is
basically you want to provide to the
non-toxic flow while losing as little as
possible to the toxic flow and so sort
of the ratio of how much non-toxic flow
to toxic flow there is uh matters quite
a bit and yeah in the market conditions
from six months ago when the volumes are
1% of what they are today and also in
the current market conditions the the
market makers are profitable and of
course they're adjusting their
strategies given the volumes given the
nature of the flow um you know very
actively monitoring as opposed to again
a passive liquidity provider and then on
your first
question one thing I've always respected
about toly and his vision for salana is
it's extremely ambitious I think Global
shared State makes a ton of sense
for uh defi specifically I think we look
at the rollup landscape today the L2
landscape and there's so much
fragmentation
and it really limits the usability and
usefulness of a lot of these UHC ations
are really these like pieces of infra
where one of the biggest benefits of
doing things on chain is the shared
state where you have this notion of
composability assets that can be used
across apps in a very very seamless way
and I think at some point you need to
probably have some opinion on what
pieces of State need to be shared and
what can have more like asynchronous
composability so for example I think it
makes a lot of sense for uh in the long
run of defi to live on the same shared
State whether that's salana or somewhere
else uh but maybe that doesn't need to
be the same place as nfts or or dpin or
whatever else is going on uh in
crypto yeah I think that that's I think
a lot of people are realizing this at at
this point um okay so I want to ask you
more specifically about like salana what
you what what are your scene right now
with
like uh well transactions are very hard
to land on chain as as you probably know
um and there's like you have done a lot
of research on this I would actually I
don't know I'm probably just talking you
up right now but I think it's also true
that you're like one of the foremost
experts on the salana fee structures you
have some strong opinions what's salana
doing right what's salana doing wrong
like what are the issues and like how
rectifiable are these you can just
ramble however you want yeah there's a
lot of problems with salana today um I
think the most important thing though is
like salana like kind of works it mostly
works and it works at quite high
throughput with very credible levels of
decentralization uh which is extremely
differentiated in the market today uh
and one mental model for salana is
basically since uh main net beta
launched in March I think March of
2020 Sal's basically been putting out
fires the entire time where things are
just completely breaking on the
engineering side or
uh on the business side or
whatever and over the last year there's
been a ton of time to you know pay down
some of this technical debt think
through some of the more like long-term
research problems which salana just
hasn't really had the luxury of being
able to deal with this all has been in
like this survival firefighting
mode and yeah like the again the mission
is extremely ambitious like you're
building this massive distributed system
that's permissionless permissionless to
participate in whether it's as a
validator um you know contributing to
consensus or as like a user application
developer and that's just an extremely
hard problem so I think salana was
founded
in 2017 or
2018 and the output since then has just
been extremely impressive and I think
it's super unrealistic to expect that
Sal would look anything like a completed
product
today um so yeah mentioned that volumes
have picked up by about 100x in the last
six to to eight months and it's brought
a lot of problems to the chain uh
there's definitely some some of these
like long-term issues that need to be
addressed some of which are on the
engineering side like the what we call
the networking layer on salana which is
really just part of the uh client
implementation has some issues where
effectively every validator has these
like ports that are exposed to the rest
of the world and there's you know
trading opportunities very directly
profitable opportunities to sort of spam
this layer with massive amounts of
traffic and it's a very difficult
engineering problem to figure out okay
how are we going to throttle this how
are we going to assign bandwidth to
different types of users or rpcs or uh
or
stake uh so there's a lot of engineering
problems there are plenty of research
problems as well uh I think salana
historically is not focused very much on
incentive compatibility very different
from the ethereum side and I think it's
worked out quite well for salana just
really focusing on engineering really
focusing on performance and sort of
kicking the C can down the road on some
of these uh thornier research problems
which we'll need to get addressed at
some point but actually have not been a
significant issue to
date on the topic of fees yeah I think
like the fee model and just doesn't make
a ton of sense right now where there's a
couple things you're optimizing for with
fees first you have some limit on block
space
and so like you can't increase Supply
infinitely and that means you need some
mechanism for throttling the demand
which is almost always going to be on
something like price and one really good
way to do this is with uh like a GL fee
model where you also care about the ux
and we saw this with ethereum you know
1559 where yeah there was the burn and
all that but actually much more
important in my opinion was the ux
Improvement where you basically know I
can send a transaction with some you
know like small tip and then I'm just
going to pay the base fee whatever it is
and I know it can't increase by like a
100x in the next block there's a Max
escalation and you're just going to get
your transaction in with quite high
probability in most market conditions
not all market conditions but most
market
conditions and this type of mechanism is
really really good for the ux where the
user doesn't need to spend time or the
wallet doesn't need to make a decision
on behalf of the user of oh how much fee
do I think I need to bid to get in in
some sort of like first price auction
mechanism salana also has continuous
block building so any sort of first
price auction is still going to have uh
inefficiency where I think you could
make an argument for the efficiency
being just as good with the first price
auction on ethereum as um as a 1559 like
model or or probably Superior I think
the mechanism design designers would say
yeah um okay also I wanted to ask you so
jeto recently like did away with their
me pool like I think you had talked
about how bad sandwiches were getting on
salana um yeah what do you think of this
like is this one of the incentive
compatibility problems that you're
talking about or the others I actually
like wanted to ask more what you think
the incentive incompatibilities are and
then yeah like so what is the situation
like with with me now where is it headed
is somebody going to step in and like
just take a take goo's place and then
also sorry I keep like just thrusting
questions at you but the other thing I
was going to ask is just like if you
could um like synthesize the difference
between meev that exists on like um amm
that we see versus order books um and
like if order books eventually are just
going to like sort of look like
traditional stat ARB or if there's like
you talked about the difficulties of
Market Mak or or the differences Market
making on something like binance versus
Phoenix so um yeah like I hope you
remember all those things and because
I'm not sure that I do but uh yeah you
can just address any of those sure let's
start with jto uh I think removing the
men pool was a really really good
decision
uh and it is another example of salana
prioritizing I mean salana here in like
this very like vague nebulous sense of
uh you know it's comprised of a bunch of
different actors and of which like jeto
is one and it's another example of
salana prioritizing the user experience
over incentive compatibility because
yeah in the long run let's say salana
makes no changes moving forward on the
engineering uh and consensus side of
things yeah the long
run uh you lose continuous block
building you end up with something that
looks like PBS you might even end up
with something looks like four slot PBS
because leaders uh on salana get four
slots in a
row um and you also definitely end up in
this PBS construction with some sort of
front running
capability and that's just like really
bad for the user where you have these uh
you know today the main use case on
salana I would say like on any
blockchain
is playing the shitcoin game
and a big part of shitcoin game is you
want to place a market order Market
order really only makes sense in a world
where you have this first in first out
type
ordering uh because when you come into
the market and you say I'll take
literally any price or I'll take up to
50% slippage or 99%
slippage if that order is not executed
right way if someone else gets to look
at it first then you're going to get
front run you're going to get the worst
possible execution which is just a
terrible outcome for the end user and
yeah there's sort of ways around this
with private order flow uh we're seeing
this game start to play out or actually
it's played out quite a bit on the
ethereum side where uh you know some of
the largest block Builders three main
block Builders and a couple of them have
exclusive order flow deals with some of
these uh large senders of Market order
shall we say these telegram trading apps
where users are relatively execution
insensitive
and it's quite centralizing at the Block
building layer of course where if you
have more market share
today uh the you're able to provide
better guarantees to the incremental
order flow so the cost to to serve the
order flow is is lower for you
and yeah if you play everything out to
the equilibrium you end up with
something very very similar on salana I
think jto turning off the men pool is
not some permanent fix I don't think
it's ever been marketed like that
certainly not by the jto team
and what it does do is buy a little bit
of time where uh yeah sort of going to
have to figure out how to do Market
order on chain
if if you're going to be able to support
some sort of shitcoin economy and you
can do this with like a trusted
sequencer which is what the layer twos
do today um but if you want the sequence
of set to be fully permissionless then
you're going to need a better long-term
solution so now I have two questions and
that's I'm G to remember which ones they
are okay so the first one is you have
observed and honestly I would say very
well informed on ethereum me and you've
watched what's sort of UN unfolded in
that uh in that domain what are your
thoughts on just like the general
direction of me on ethereum and like do
you think that so like obviously there's
been some I mean there's been a lot of
opinions on this but there's been some
opinions that ethereum has in some way
like attempted to maximize the me well
actually this is sort of a stated goal
right the meev accelerationist whatever
you want to call it they've said look
we're going to like maximize it and they
we're going to work out the Kinks but
then there's people who have said okay
well it's not so clear we're working out
the Kinks and it seems to just like be
accelerating um so I want I want to know
like what your general view is on
ethereum MV and then the other thing is
how much of this because this is
something that toally talks about a lot
and you know there's also different
opinions on this how much of me like
just functionally is solved by by
blockchain's gun really really fast like
is it is there a way to
um uh is there a way to quantify that
like such that you can I I know that you
know like it's the difference between
onchain prices and offchain prices like
this is like a lot of what contributes
to M on ethereum for example but is
there a way for you to say like we can
push me down to a negligible amount with
xeed or you know like the blockchain
going this fast in in your mind so those
were two two specific questions but okay
I'll try to answer them in order this
time so the first was on ethereum
me I
think ethereum has not had the luxury
that salana has had in addressing the me
problem where I think it was another
case of oh [ __ ] is like on fire and we
us to address this right away uh I think
this was starting in like late 2019
early 2020 as defi really started to
pick up and you know there were all
these mempool games getting played you
all these failed transactions landing on
chain which is much more painful on
ethereum than it is on salana because
the block space is so
constrained
and because there were like so few
mining pools that were uh contributing
like a large market share of Mining and
there's like sort of no way to kick them
out you can't like slash them there's
like actually no option to
to uh and I don't think a the would have
considered this anyway
but there wasn't really a way to uh you
know nicely ask a minor to not do the
profit maximizing thing and in that
market we
saw ethereum
miners spinning up deals with Searchers
sort of under the table we
saw some mining pools attempt to extract
Meb on their own
and so this democratization angle I
think made a lot of sense in that proof
of work World
um and you actually had to you probably
had to enable sandwiching front running
because all the order flow was public at
the time maybe you could have just gone
the private order flow
route I think the real issue is you know
moving towards proof of stake and
PBS really having an emphasis on this in
set of compatibility uh really having
this focus on the solo Staker being able
to make as much yield as the
institutional Staker which is never
actually going to be the case but the
goal is to at least like decrease the
Gap and hope that I don't know solo
stier air drops will make up for the
difference or more than make up for the
difference
uh and to contrast with salana the me
infrastructure just didn't really
exist and only started to get built out
during this Market when there was very
very little activity on salana and
so the actors extracting Meb I would say
in general like less sophisticated on
the salana side than on the ethereum
side which is just a function of the the
market side
and I think that will not be true in the
medium to long run J has a very strong
position in this market and I think they
are able
to influence For Better or For Worse the
like where the Meb markets go on Sal
today they have more than 70% of the
stake I believe which is not them owning
the stake it's just 70% of the stake
running the jetto salana
client uh and yeah I think like removing
the men pool is one thing that I mean
well to be fair they added it in in the
first place there's no me pool by by
default on salana transactions are sent
directly to the
leaders there's no PTP gossip uh of of
pending
transactions uh yeah I think removing
that was just like clearly a good
decision uh and then your second
question was good but I forgot what it
was my second question was just I mean
you started to move that direction but
it was just if you make it really really
fast you make a block really really fast
can you
decrease me to I don't know a negligible
amount I don't know if that's really uh
a stated goal or I mean how how much
does just making it really fast because
this is something like anat says right
like this is like one of his big points
is like you make it really fast there's
less than e it's just like period end of
end of story and I wanted to hear your
take on that like coming from like the
more more looking at this as like
somebody building an application top uh
salana and yeah you can run with it from
there sure yeah uh I guess my first
point here is the application developer
really needs to care about me because
you can think of an application sort of
as this Clos system so on Phoenix for
example you have uh like market makers
and you have uh takers you have two
sides of the the flow and then what me
sort of is is this like extraction from
this Clos system and so if you're trying
to build something that is sustainable
you want to minimize that in the same
way that uh you know if you're doing any
sort of Market design you want to
minimize the the toxic flow relative to
the non-toxic flow with the
understanding that you'll never be able
to get to zero toxic flow and that
that's
okay
um running the blockchain
faster definitely reduces some types of
me increasing amount of block
space
Also increases
the protocol design space
and so there's more things application
developer can do to mitigate
me at the end of the day though what you
really care about is the proposer or
whoever has the right
to insert and order the
transactions uh what really matters is
how much agency they have so on ethereum
they have this not easy but relatively
easy problem of okay you have call it
150 transactions and you have 12 seconds
to like jam it into a block and me the
[ __ ] out of all the retail trades and on
salana it's like you have 1.6
seconds uh across your four slots and
far more transactions so it's like a
little bit harder to to meev it uh no
matter what though the most
sophisticated me actors are going to be
smarter and have more uh compute
capability than like the the weakest
validator and so if you're an L1 and you
care about
someone being able to run a node on a
Raspberry Pi or be a validator with you
know spending less than ,000 dollar a
month on on compute and and Cloud
hosting the most sophisticated Searchers
are going to have much more
capabilities and basically the only
thing you can do here is reduce how much
agency the proposer has which will have
other effects like the transaction
ordering will be less efficient
efficient in the economic sense which I
think might be okay
because like the max efficiency means
Max extraction which is bad for the
users and bad for the
applications there's this interesting
proposal on salana from
toi uh I think it's called concurrent
block proposers or something like that
where instead of having a single
proposer per slot you have multiple
proposers
and let's pretend there's two
proposers
they each have a right to call it half
the block space and then they pick some
set of transactions that they're going
to put in that block space and then the
consensus mechanism finds a way to merge
them into a single block and so you're
going to lose a lot of efficiency here
because two independent parties packing
blocks just going to be less efficient
than one one you're going to have
probably more failed transactions
because if I build a transaction uh if I
send a transaction I don't know at
execution time or sorry
at uh at the time it's inserted into the
block you don't know if it's going to
revert or not because execution depends
on
ordering and so yeah you just like lose
efficiency in all these places but you
really you also reduce the amount of
agency that proposers have which reduces
the amount of extraction there can be so
if you can run a system like that really
really fast where you have like a ton of
block space and you know we're just
we're just going to eat the failed
transactions and it's not a huge deal I
could see something like
that mitigating some of the problem in
the medium to long run is a huge
engineering lift to get there and
there's also some tradeoffs where today
on salana you get pre-confirmation for
free because the leader is streaming out
the block as they're building it and
that is not just was your transaction
included or not it's the state
transition that comes with the the
transaction and once you separate
transaction inclusion from transaction
execution then you have to wait for
consensus and these multiple partial
blocks get merged to know what the state
transition is so there's there's going
to be trade-offs all over the place here
but that's the nature of this
distributed permissionless system uh I
wanted to ask your person opinion you
don't have to speak on party lines here
I want to ask your personal opinion on
the level
of I don't know I would I guess
decentralization I can't think of a
better word right now but what do you
think the ideal chain looks like in
terms of do people need to be able to
run a Raspberry Pi or is it this idea of
um a hundreds of data centers around the
world um like what to you is
I don't know what should we strive for
in this capacity well it depends on what
the the end goal is so for example if
your goal is for the chain to be
resilient to an event like World War III
then you're also going to want to be
robust to things like massive data
center
censorship and if your goal is to build
a financial system that runs on chain
and provides really good outcomes to
users
then you know your requirements might
look a little bit different going off
that a little bit what is your view on
the sole token this isn't like Financial
advice obviously I'm just like thinking
as as the token used on salana and like
what is the long-term vision for this I
think CU I've talked toally and what he
says is like what we're trying to do is
build a platform and if salana if the
soul token is like the 100th largest by
market cap I think token on salana then
we've done a good job because we've
optimized for applications right um does
this work in your mind in terms
of yeah like developing a sustainable
fee and validator structure so the L1
token price matters if you care about
Economic Security and I think it's not
actually super clear how important
Economic Security is but you could
imagine a world where let's say the soul
token was extremely cheap there's like
no demand whatsoever call it like $1 or
one cent and then that means some can
come in and buy up like a massive chunk
of it um are they able to force an
invalid State transition answer is still
no uh which I think is totally argument
you see like one honest node to go out
to the world and say like hey this guy
with 99% of the stake is is doing
something that's [ __ ] up so we need to
Fork the chain and I think that's
all quite
reasonable
um the
l1s I think have the best argument for
why the application actually needs a
token uh it makes a lot of sense in the
proof of work context makes a lot of
sense in the proof of State context um
whereas I think a lot of other shall we
say like applications with the token or
sort of fishing for a reason to have the
token doesn't actually make sense for
the product uh more makes sense as like
a you know
a uh like a capital generation mechanism
shall we say and
yeah I just don't really see any reason
why you know if if the soul token is the
H hundredth
largest by market cap on salana why the
chain like doesn't function I think the
economic security arguments don't make a
ton of sense uh I think it's quite
unrealistic though uh maybe the way
tokens trade will change in the future
today they very obviously do not trade
on any sort of fundamental basis so if
you ascribe to that then like the fee
model and the revenue model don't
actually matter for the valuation of the
token but I think it's CBD if that will
continue in
perpetuity yeah so you kind of ascribed
to the economic security is a meme
category I wouldn't say I wouldn't put
it that strongly but I think it is
overrated yeah I think and also like as
you can tell toally goes way out of his
way to troll about this right like with
the Chihuahua chain stuff like he goes I
mean he's like like it's like
performance art you know it's crafty but
it also does like it certainly does make
it it definitely makes you think let's
put that way uh okay I think the main
point the main point that he has is the
honest minority can always Fork out and
yeah it's going to be extremely costly
but if you have a dishonest majority
that's the only thing you can do and
yeah definitely if you're if like that's
part of the
uh the plan against some
yeah very large actor that accumulates a
ton of stake then yeah I think Economic
Security matters a lot less
than than most people in crypto think
yeah 100% agree and also like the thing
is like the part of the part of the
premise of proof of stake is that you
can Fork out you know malicious actors
and so like ascribing some not like um
amount that can be denominated as
economic security is like somewhat of a
frivolous Endeavor because like the end
of the day like it's the it's the notion
that their state could be forked out
that should in theory deter any
dishonest actors like at least at least
in the model so like I do kind of wonder
about that myself but anyways um okay I
want to like zoom out now or I don't
know zoom out but just like take this a
slightly different direction um we can
start with like okay so in five years F
okay you can start with 5 years and then
after that I'm going to push you to 15
years what are we using Phoenix for what
are we trading um like give me the give
me like the hyper bullish outcome not
like for investment reason just give me
the hyper bullish outcome
on trading on chain uh with Phoenix
specifically and then we'll move to like
def5 as a whole but like what what where
where is Phoenix headed and like what is
your hope for what you're building in
five years we should be able to get to a
point where trading on
chain
is as good or maybe even a significantly
better experience in trading on a
centralized crypto exchange and
concretely what that means is the
product experience is better the
liquidity is as good or better uh plus
you have these guarantees
around auditability
transparency permissionless access
permissionless Innovation on top of it
which don't exist with centralized
exchanges or at least don't
today 15
years on boarding of
a significant chunk of non-digital
assets onto this um onto these crypto
rails I don't know what those I don't
know what those um onboarding mechanisms
are going to look like but by that point
I think if we haven't been able to build
a system that is you know defy as a
whole better than trafi I would view
that as a pretty significant failure
yeah so like in 15 years are we Trading
apple and Amazon
shares on Phoenix
specifically no but just on
chain I think so okay that's quite
optimistic uh all right so then like
sort
of adjacent to this question uh what is
like just your overall view on Defi and
maybe blockchain as a whole like what is
it very useful for like so you seem to
think that it can be very useful for
trading obviously building a trading
application what else is like what is
the future for this industry and like
what is this stuff really good for and
I'm GNA press you so be prepared yeah so
I don't think about the philosophy of
crypto all too much I'm really really
focused on the defi side I'm not super
imaginative when it comes to uh what the
future of nfts or crypto gaming might
look like where those applications are
uh I
think there is like sober realiz ation
that needs to be had which is that not
everything belongs on chain you have to
think about what are the costs and what
are the benefits of putting something on
chain so the costs are extremely high
where just any application you put on
chain uh is going to be from like a
developer perspective much less
ergonomic much more expensive much more
expensive from the user point of view
because you're eating this overhead of
you know thousands of nodes coming to
consensus uh and sharing the state
rather than a single copy
in
AWS um in Virginia or something like
that and then the benefits are
around permissionless access or
sensorship resistance they're
around composability and permissionless
Innovation and this is why like Finance
I think makes a ton of sense for crypto
where you have this massive Trad F
system which I think is like mostly been
built by well-meaning
actors um or at the very least like
profit seeking actors um but it's been
built iteratively over like a hundred
years right and that means there's like
a lot of crft in the system there's a
lot of entrenched actors is very
difficult to
innovate and I don't think Finance isn't
is in a completed State I don't think
it's near a completed
State and so crypto specifically with
its permissionless Innovation and
access is perfect for disrupting the
traditional Financial system and not to
just like build some shitcoin because
you know on the side but actually to
build a financial delivers better
outcomes to the end users which are like
real people and real companies real
businesses as opposed to again
uh like pure shitcoin nihilism when you
look at ethereum right now and like the
we sort of touched on this but now like
zooming out just talking about like I
guess defi as a whole like are you what
do you
see the future of rollups um cross chain
interoperability all of this stuff like
you you mentioned the fragmentation like
this is something that's discussed like
at infin itm right and like uh it seems
like this stuff can be well I don't know
honestly I don't write code so I don't
know if this stuff can be like solved so
easily um some of it is more intuitive
some of it's like I'm not quite sure
about but uh it is annoying like to it
it is annoying when you have like your
metamask and you have to switch between
eight chains like it's like there's no
denying that right um and I think that
salana has stepped up like Solana saw
like the benefits of like just sticking
this all in one layer and like running
with it and like we'll fix problems like
you know move fast break things and like
things break like um so is ethereum
still innovating in your mind and what
do you think about rollups as a whole
like is this the direction that should
be taken I think it's a very reason
reable approach to take I guess to
answer the first question ethereum is
definitely still innovating ethereum is
on the Forefront uh specifically on the
research side of so much that relates to
blockchain I think salana owes a ton to
ethereum uh for Paving the way and
sewing some of the initial Vision even
if uh you know salana is opinionated in
ways that might disagree with some of
the ethereum core
philosophies or core values let's
say on rollups specifically yeah of
course the interop is a huge problem
today I think there's a bunch of
different narratives floating around
here the rollup companies or the L2
developers will tell
you well the big ones will tell you they
want everything to happen on their own
on like their own L2 so uh you know
believe in this vision of one true L2
and you just have to Duke it out to
figure out who's going to win and then
some of the infro companies
are like pedaling this narrative around
oh you can just like solve the interupt
problem and I think today we just
haven't seen a good solution there and
it depends on what you mean by interupt
so there's a bunch of different forms
like uh for example synchronous
composability versus asynchronous
composability can I like touch multiple
applications across multiple layer twos
or multiple chains in like the single
Atomic action which uh you know
synchronous composability and
this almost certainly requires um some
sort of shared some some notion of
shared validity which is like sharing
the same sequence or set
and having one Chain's validity depend
on anothers and this sort of just looks
like one Mega rollup you just called it
n different rollups so you get n tokens
plus a sequence or token so maybe there
is some product Market fit for that but
from a architectural standpoint I think
doesn't make too much sense uh I call
that the salon at home where you have
this uh multiple you know single
threaded machines that sometimes
interact with each other which actually
looks very similar to a massive uh
single state machine
where you have a locking mechanism on
different pieces of state which is how
salana is
built
um but again not everything needs to be
composable with each other in the same
way that you know we use tons of
different internet applications today
but my slack doesn't need to know
everything that is going on on my Gmail
it's actually okay if one of them runs
on like AWS and one of them runs on on
gcp it's just like not a problem
whatsoever and again like yeah your
crypto gaming might just happen
somewhere else from your Defi and I
think that's that's totally fine yeah
what criticisms of salana that ether
Community wages do you think are
legitimate and which ones do you think
are
illegitimate I think most of them really
just come down to differences in values
and so that will affect if you think
it's like legitimate or illegitimate for
example ethereum really has this value
of anyone should be able to run a node
from home anyone should be able to
participate in consensus from like a
home internet connection and uh
relatively small
hardware and the reason for that is for
censorship resistance and some notion of
like World War III robustness and this
is just not really something that salana
prioritizes that highly and that's just
a values
difference um I think maybe where I
would disagree with Sana is
more on the research side and how
important some of these you know getting
some of these core mechanisms right
rather than slapping a bandaid on here
or there which is uh you know like an
engineering team where everything is on
fire that's what you're going to
do one example is around the fees where
yeah just like the fee model on Sal does
not make sense has not made sense for a
very long time sort of comes from a
point of view where supply of block
space is just always going to be greater
than demand for Block space which
actually is like quite a pessimistic
view um and I don't think is expressed
explicitly but kind of comes out in the
an engineering decisions or
prioritization and you know we see some
of the effects of that today where yeah
you just like don't know how much you
need to bid to land your transaction and
so there's massive inefficiency in the
bidding it affects the ux etc etc yeah
that makes sense um all right well I was
gonna I want to
ask yeah maybe one like one last
question just you're looking at some of
these other chains that are cropping up
like maybe less so on the L2 front
but some of these high performance
chains like for example U manad is
coming out soon it's like a paralyzed
parallelized evm um I want so I'm just
using this as an example I want to hear
and I'm not sure exactly what I'm asking
here but I want to hear like is the
vision in your mind for these chains to
just continue to innovate and that
sounds like um that sounds like it may
be obvious or may not be obvious
depending on like what chain you sort of
ascribe to but what in your mind like
will make salana different from the next
chain that comes out that's also
extremely fast and after that like
technological innovation versus maybe
some of like value sets inscribed um
because I think that like I guess what
I'm getting at here is that salana I
think was Saul okay so this is my
opinion right like this isn't I don't
think what anybody else say but this is
my opinion right like Shana said okay
Finance should go really fast actually
Finance should go really fast and what I
think happened actually is that a lot of
what was occurring on ethereum was not
like people maximizing for
decentralization or like maximizing for
like some like of these values now the
chain May in theory be designed in such
a capacity but that's not what people
were really using it for and so I think
like when like a lot of what salana has
seen recently it's just people
maximizing for more degeneracy over here
and so I want to know like is this
something where we just continue to hop
from chain to chain to chain and like
just continue to follow technological
advancements or is there a way that like
salonic continues to stay a breast of
technology and can ensure like some of
these ideals of decentralization I don't
know maybe this wasn't such an
articulate question but like hopefully
you start philosophizing for me well
it's a bit of an odd question because I
think prior you should expect if salana
is the fastest or most performing chain
above some level of decentralization
however you want to mention however you
want to measure that um it has some
incumbency advantage and again it's not
a finished product it should continue to
get better I don't think anyone views
Salon as in like a finished state today
and so you know it has like in my mind a
huge Head Start Over a chain like monad
or you know compared to some of these
high performance chains that came out a
little a little bit later like Aptos and
andu uh there's just this difference in
adoption
and that comes from being out earlier
and gaining some mind share earlier
but I think I have a more practical view
towards blockchains which like they're
pretty valueless by themselves and they
really depend on the applications that
are built on top and you know as an
application developer
when we look at any sort of infra we're
just you know we're picking the tool
that's best for the job and so when we
take a look at salana today it's good
enough for Phoenix and we look at the
other chains like not really good enough
uh from a Pure Performance perspective
and there's a bunch of other things that
come into play as well if you know we
need to do some tiebreaking like uh what
does the asset base look like are there
users are there active Traders and maybe
if the interop problem gets you know
solved in in some capacity by I don't
know uh a really great multi-chain
wallet that is actually able to abstract
the chain away then some of those chain
level Network effects are going to
matter less uh that's not the world
we're in today um and yeah I think like
none of these blockchains are close to a
finished State and all these attempts to
innovate like uh you know moned is a
great
example um are really good for crypto as
a whole because infra is not good enough
for like real applications that are
dealing with tons and tons of value um
they're not efficient enough they're not
performance enough and it's really
really hard to get there while
maintaining some level of
decentralization yeah okay that was a
good answer to a pretty non-descript
question um okay well thank you Eugene
where oh also um okay so they can find
you at on Twitter at Ox [ __ ] Trader and
then also um you write like some
research stuff as well if you want to
like show that and then anything else
you want to show as well yeah we
published some research on umra
research.
XYZ or umra Research on Twitter uh
various blockchain
related research pieces more on the
infra side more on the um like infro
technical design side relating to
primarily ethereum and salana
but hope not to be limited by that in
the future awesome well thank you so
much for coming on we'll have you on
again sometime I thought it was great
thanks for having me
G
Посмотреть больше похожих видео
Kamino 2.0: Brand new Borrow / Lend market on Solana
The MakerDAO-Aave Debacle: Risks and Rewards of Collateralizing DAI | Analyst Round Table
Where are we in the Cycle? w/ Darryl Wang
Layer N: Hyper Performant and Hyper Composable Execution. An Interview with Co-Founder David Cao
Trading Psychology - Dr David Paul
Solana BLINKS | The Blockchain Links that opens PANDORA's Box
5.0 / 5 (0 votes)