Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
Yearly Wrap Up Spaces W/ Sei Labs Co-Founder Jay
Recorded: Dec. 29, 2023 Duration: 0:59:01Space Recording
Full Transcription
GM, GM, Mic Test123.
Good morning, good morning,
how is everybody doing?
Excellent, thank you sir, excellent.
And yourself?
I got a little bit of a raspy voice,
I apologize.
It's the winter cold
but doing my best to power through with some tea.
Excellent, excellent.
I find this time of year after Christmas Day
and Boxing Day
too many glasses of beer wine whatever yeah it's uh it's pretty slow but um sweet who we got
piloting the say account today i think joe you over there yes sir good morning good evening
good afternoon sans happy to be here uh nice to see you grover nice to see you jay nice to see you
logan gm gm guys gm sir well uh without further ado shall we get started let's do it let's go sir
awesome awesome so i think we should start out with some uh brief introductions uh i'll go ahead
and go first uh my name is is joe or the average joe um i'm on the growth and marketing team for
about a year and a half now i'd say and um you know residence uh spaces host and uh resident say
poster so um you know i'm super excited about this one it's been one hell of a year for say
um and i'm just excited to kick it off and reflect um so grover would you like to go next sweet yes
sir so yeah hi guys my name is grover i'm head of marketing here at say uh i've been with the team
since it feels like 10 years but i think it's it's coming up to one um so yeah um it's been it's been
quite an adventure so uh i think maybe it's worth saying i've been in crypto since 2017 uh been through
the ringer as a as a normal user working through coinbase all the way through to uh speculating on
my own coins and all the way to to working here at say so it's been it's been great fun
all right i'll go next um my name is jay i'm a co-founder of say labs um which is the uh development
entity for uh supporting the save blockchain so excited to be speaking with everyone and i think
the last person is logan do you want to go ahead logan let's do it uh my name is logan uh co-founder
and managing partner of frictionless capital uh we're a thesis driven fund but primarily focused
on high throughput blockchains ultimately we kind of say the blockchains that enable uh really the
development from the users and uh user and engineering standpoint going from dial-up internet to kind of
broadband and fiber optics beyond and the unique and novel applications that ultimately can be built
built on those blockchains like say awesome thanks for joining us today logan and and thank you jay um
i know you're busy building and a ton of exciting stuff coming up especially with with savey too so
um why don't we go ahead and just kind of get started we have some uh questions sourced from the
community um a lot of a lot of people are excited for this one and to hear from you know the gigabrain
that that's actually building say so um you know it's been a wild wild four months uh since the
since the beta mainnet launch of say um and there's so much more to come in the future but i would love
to ask jay um you know what what are some of your reflections uh from the past four months and and uh you
know since we launched uh mainnet uh only four months ago yeah i mean it's actually pretty crazy how it's
only been four months um i mean kind of resonating with what george or sorry with grover was saying
before um it does definitely feel like it's been a lifetime and i mean 2023 the way this entire year
kicked off like we were definitely not i think in the same kind of spot as we are right now like at
the start of this year um ftx has just had just collapsed uh november of last year so it was
definitely i think depths of the bear market um we'd raised our seed round a few months before then
and uh i definitely wouldn't say that say it was the hottest deal on the market at that time
um i think that there was a ton of stuff that we had to go through um kind of get to where we are
right now so i mean since the mainnet launch has happened i think that it's been just kind of
fantastic to see um the amount of support that the community has been giving us especially in the
past month month and a half um from a technical standpoint everything has been going phenomenally
um say launched uh it was very smooth um sustained 390 millisecond time to finality so
say has continued to be from launch um the fastest chain um just out there right now uh and yeah i mean
in terms of a lot of the ecosystem traction that we've been seeing in the past a couple of months it's
definitely been phenomenal to see like nfts inscriptions a lot of the new um trading activity
that's happening on chain as well so i would say overall it's been um pretty fantastic uh it's
interesting because today also marks the one month um anniversary since we announced day v2 uh which
was the proposal to build the first paralyzed edm um and i i mean it kind of seems crazy that it was
only a month ago um there's just been so much that has happened since then um so much excitement
around savey too so many technical discussions around the nuances around kind of how we're
um implementing it so yeah overall things have been just fantastic and um yeah thanks for
uh the community for supporting us this entire way yeah echoing a lot of that what jay just said
there about um how we've seen a huge change in so the community people waking up to say on mainnet
seeing i'm seeing like a lot of groups and projects that planted seeds over the past few months and
they're all sort of started starting to sprout um so really really proud of what the community
are building on top of say um especially with the dragon ball memes which is great as a dragon ball
fan so uh yeah that's my two cents
yeah i i think everybody loves um the the new uh kind of mascot for say and everybody loves the the
say and memes um you know uh memes control the universe is as uh elon musk would say so it's been
incredible to see like um how the community just took the same meme and just ran with it um so um i
think i think this is a good kind of uh point to pivot and talk a little bit more about uh savey too
um there's so much more to come for say and it really feels like say is is just starting um so i would
love to kind of hear more about um you know what say v2 entails um kind of um what uh you know what
is in the pipeline uh for say when it comes to this so you know i recently a month ago say labs put out
a proposal for say v2 um jay i would love to uh hear you kind of take us through some of the specifics
about say v2 and share how say labs came up with the proposal um and and and logan if you have any
questions as well i would love to hear you chime in and um kind of ask away feel free to to go ahead
and speak i know you have a lot of experience in this industry and i would love to hear your insights
as well yeah i mean dude logan is one of the biggest kicker bands i've met um especially around
paralyzed chains so i mean dude it's like fantastic that you joined us i'm happy to have a much more
nuanced technical discussion as well um i'll kind of keep things higher level and then let logan ask more
detailed questions so a little bit of like where we are as i was mentioning before um august of 2023
so a little over four months ago they went live on mainnet um from a technical side i think things
were going normally um but the biggest pain point for developers was the lack of evm compatibility
like if you look at crypto native developers right now almost all of them are evm developers there's
some folks that are building on solana right now i would say that's almost certainly the second
biggest camp um especially in the past uh around three months solana started getting a lot more
traction again um but outside of maybe like solana and then some smaller camps out there um almost
everyone right now that is crypto native is an evm developer and asking an evm developer to leave
the evm and start writing rust-based smart contracts is a very tough app because once you've already
built up expertise in the evm transitioning over to a new language a new execution environment
uh it represents massive risk for you because even a single like bug in your rust-based smart contract
um that you might introduce because you don't completely understand how things work well that
can result in your entire protocol getting drained so it's just extremely high risk for people to move
over um and the biggest request from developers um from a lot of technical folks in the say ecosystem
has just been evm compatibility um this isn't something that completely caught us by surprise
um since last year we were actually building out evm support um i tweeted about this i think a week ago
our initial um foray into evm support came in august of 2022 so we actually have like our first code
requests um from august of 2022 um and what we basically decided to do from the say lab side was
okay there's this very clear developer kind of sentiment that we need to have evm support
um so let's start supporting the evm and that's exactly what the proposal for savee2 is um savee2
will be the first paralyzed evm and what this means is it allows us to get the best of both solana
and ethereum um we're able to get a hyper optimized execution layer um while still supporting the ethereum
virtual machine and all the mindshare tooling um and support that already exists for it
so in terms of the actual proposal um there's three separate parts to it um i'll just give a very
brief kind of minute two minute overview and then um let folks dive into questions here
so the first part is evm support more specifically we're going to be using um get which is the goal
implementation of the ethereum virtual machine um to be able to give full evm by code um compatibility
to say and this will be implemented as a first class citizen of the chain so similar to how we
support cosmon wasm smart contracts right now we'll be supporting evm smart contracts the exact same way
so this isn't like a uh this isn't um a smart contract built on top or like a separate chart or
separate roll up or something this is just native to the chain so it's going to be completely seamless
and there'll be um synchronous composability and atomic transactions with the rest of the chain
so it's built in the cleanest way possible um the second thing we'll be supporting is optimistic
parallelization um so we'll probably talk more about parallelization later on but the nice thing about
optimistic parallelization is it lets you get the performance benefits of parallelization
without requiring developers to do any additional work there so developers can just take an existing
smart contract from ethereum l1 just deploy it to say and there will be no issues that arise so there's
complete flexibility and compatibility there um and the last thing that we're doing is cdb so one of
the issues with creating a high throughput chain is there's a ton of uh state and storage related um
issues that can come up so cdb is the way that we're combating a lot of these issues um and in cdb
uh i mean we're changing the way that state commitment state access and state storage works
in terms of the impact on this we're actually able to reduce state storage by 50 percent so full nodes
need a lot less disk space um we're able to do a 12x speed up for state sync so if you're running a new
full node you're able to state sync 1200 faster um and we also have async writes happening to the
database so through these async writes we're actually seeing 150x speed up um for commits so i'll pause
over there um a lot of stuff we have planned for saving too yeah i mean i think personally one of the
most impressive things from just watching uh say labs has really been the how quickly the team
iterates and i think as jay kind of pointed out starting with a virtual machine that is a little
bit different than the most popular virtual machine today which is the ethereum virtual machine
can be very kind of burdensome on the engineers and i think it's been amazing just to watch the say
labs team start to implement uh the new say v2 and kind of what's going on there but maybe jay to bring
the conversation up a little bit higher level at least for the beginning i i think obviously people
are kind of familiar with ethereum uh kind of the ethereum virtual machine can you talk perhaps high
level what a paralyzed evm actually impacts for the average day person yeah yeah absolutely so every evm
chain out there right now is single threaded so what this means is if you have let's say a hundred
transactions in a block they all need to be run one after the other um and this is really simple like
from a software engineering standpoint it's much more simple to implement a single threaded um
execution environment the downside here though is that you're not completely taking advantage of
modern hardware like modern hardware has several i mean most modern machines have many different
cores and if you're just using single threaded um a single threaded virtual machine then you're not
really leveraging the type of performance you can actually get with modern hardware um so what
parallelization lets you do is take advantage of uh basically what uh computers most people on this call
are listening with um take advantage of like the kind of consumer hardware that already exists
and be able to get better performance um i'll avoid going into the details around parallelization
unless um it would be helpful to just give kind of a walkthrough for like what parallelization means
um in terms of the impact on the user experience i would say there's a couple of different things
that happen um the first is that with parallelization you're able to support more transactions per second
um and this leads to lower gas costs for people like before if you were only able to get a hundred
transactions into a block now you're able to get a thousand transactions into that same block
well then suddenly there's less uh competition happening for that block space um so people are
able to pay less money to get their transactions included so this represents decrease in transaction fees
um i think this is particularly relevant right now just given how insane ethereum l1 fees are getting
and i mean even on roll-ups it's not really that much better um and the reason that things are so
expensive on ethereum is because you're not able to get very high throughput and if you're stuck with
like 30 transactions that can be processed each second then like you need to be willing to pay a lot
more money to have your transaction be one of those 30 transactions um so the first thing is just
decrease in gas costs um the second really interesting thing is that when you have greater throughput
that happens um you're able to facilitate new types of applications like right now everything
that is built on ethereum is kind of fit into this model of you're only allowed to have like 10 to 30
tps um so you're not able to build something that needs more tps than that so let's say you want to
build an order book based exchange where you want to have let's say thousands of tps or tens of
thousands of tps so that you can really have um a lot of transactions and orders being submitted
simultaneously in the case of ethereum l1 that's not possible which is why this automated market
making model started becoming more prominent um in the case of say something like this is definitely
doable so that's that's one of the things i'm really excited about like on the evm right now
you're just stuck using these kind of anti-patterns to be able to get um applications built but now
you're basically going to have this entirely new design space for evm applications like what kind of
high performance applications you want to build you can go ahead and build them on say and i think
this will lead to more interesting things that users can use as well yeah the i think for me
even outside of the increased throughput as an everyday user just having transaction fees uh i mean that
exists today on ethereum mainnet being the tens of dollars and i mean even during the bull market
hundreds of dollars kind of collapse to fractions of a penny is the extremely paramount thing and then
one thing that you also touched upon was predominantly today many l2 still use the ethereum virtual machine
that is single threaded and with that ultimately some of the frustration kind of arises with the inability to
isolate some of these hot spots that arise where one application starts consuming majority of kind of
the network resources can you talk about just maybe slightly more in depth on why uh parallelization
is allowing kind of this resource isolation and how that affects um engineers as well
absolutely so what parallelization lets you do is you're able to run multiple uh workloads at the same
time so for perhaps a more tangible example here uh there's four people that are uh speaking right now so
let's say i'm sending joe tensei um and then grover is sending logan tensei so these are two separate
transactions um it's touching separate accounts like my transaction is touching my account and joe's account
um grover's transaction will be touching his account and logan's account so these are separate accounts
there's no overlap so you can actually run them at the same time like you can have just multiple
threads that are running they're processing these transactions at the same time and then afterwards
they could just update all four of our accounts to say that like oh these many say had been uh debited
and credited from each of these accounts um when you have single threaded execution like you're obviously
not doing that um if you were able to paralyze the virtual machine um then you are able to have
these transactions get run at the same time now it sounds really simple when i explained it this
way um the thing that leads to more complications is as logan was mentioning um hotspots are like
basically a state that is being touched by multiple transactions uh that need to access it so let's say
one example over here would be i want to send everyone else on the call tensei and i only have 20 say in
my account so if i initially send it like let's say that i my first transaction is to send it
to joe so i send joe tensei second transaction in that same block is to send it to grover
uh and then the third transaction is to send it to logan so in this example i only have 20 say
and i'm trying to send out 30 say so the third person should not be getting that say because my
account should have already been uh debited 20 say and then um logan should just like that transaction
to send logan tensei should fail um with a single threaded virtual machine this is guaranteed to
happen because you run them one after the other so there's no way that you can accidentally send
logan tensei um because you see that the balance would be zero say when you try to send it to logan
um with parallelization it becomes a lot messier because if you're running things at the same time
um depending on how the operating system is giving different threads uh different resources to be
executing like this can be cpu resources like um memory like it can result in different machines in the
networks like different validators different full nodes having a different view of like what actually
ends up happening so one validator might say that okay um transaction one uh ended up getting executed
so joe received his tensei transaction two actually got run after transaction three so if that happens then
that validator might think that logan got tensei and then grover didn't get tensei another validator
might have a different view of the world so it can lead to these really complicated scenarios um where
there's non-determinism which is a fancy way of saying that like different validators or different
flow nodes just have a different view of what the state should be so that's what leads to the
complications over here um the way that we are um approaching it from say side is by using optimistic
parallelization for save v2 the way that that works is that the change is tries running every single
transaction that's there in a block in parallel so in this example let's say there's these three
transactions in the block um of me trying to send joe then grover uh then logan tensei um if there's
these three transactions then all three of them will try to get running parallel um we'd see that all
three of them are touching my account which would be the shared state over here and then from there the
first transaction would succeed and then the remaining two would need to get rerun so through this
approach um you're able to make sure that you're not having non-determinism uh there's a very clear way
that you're able to tell like okay this transaction touching this account are there any other transactions
that are touching this account if yes then i need to rerun transaction uh number two so this leads to
a uh i mean just substantially improved uh throughput scenario because now you're able to process more
transactions at the same time um and get better performance um in terms of how this impacts the
developer experience like developers are just able to play around with a lot more um they just experiment with
a lot of different types of applications um because we're using optimistic paralyzation developers
don't need to do any additional work um so in the case of other chains like solana for example
um developers will need to uh submit their transactions like developer front-end will need to
submit uh transactions to these rpc nodes and these transactions will need to contain the actual state
that is being touched uh which leads to more work that developers need to be doing um in the case of
optimistic paralyzation this just happens automatically amazing overview and for those i know we could j and
i like to nerd out on these things maybe one more kind of example and then we'll move on to some of
the things that you mentioned jay on how say is handling um kind of decreasing state but as an example in
terms of like real world user adoption i'm sure many of you were around when kind of the board eight yacht
club was minting on ethereum mainnet and say and ultimately what happens is only kind of the transaction
that uh the board eight yacht club contract would slightly increase in hotspot or those fees for that
contract would increase but all the other applications on the network would effectively
have very minimal fees and not be affected by this it's really kind of like um for example for my
fellow nerds like aws uh has large traffic because black friday deals on amazon would take down the entire
network because it could not isolate resources uh but really this is just adopting as jay mentioned
modern computing and modern computing practices by isolating these resource designs where you can
have multiple applications really live in one ecosystem or environment but maybe perhaps moving
slightly on into some of the optimizations that you've done on the state side with storing the data
ultimately these high throughput blockchains will have a lot of data and you mentioned early on jay that
uh the say labs team has done a lot of work on making sure that storage impact is as minimal as possible
can you dive a little bit deeper into that as well yes yes absolutely um so i guess just to kind of set
things at a higher level um when you increase throughput for a chain like when you have more
transactions that are getting submitted per second um this leads to more state getting created so state
is the uh active data that needs to persist on the ledger um to be able to process new transactions
and to be able to generate state hashes so that was basically a complicated way of saying that like
state will be all account balances so like logan has 10 say i have 20 say and it'll be like smart contract
uh data so like whatever data is being stored in the smart contract also needs to be part of that
uh state tree so when you have more transactions coming in there's more state that gets
created um this ends up leading to fundamentally two problems uh or two major problems the first
is state syncing and the second is state storage so state syncing is if there's like 100 gigabytes
um that is there as the active state for the network which is i mean roughly ethereum is at that size
right now um if you're trying to run and set up a new full node you need to import all 100 gigabytes
and then afterwards start processing the missed transactions um that you had while you were importing those
100 gigabytes so that becomes kind of a point of friction um and the second point is state storage
like if there's 100 gigabytes that need to be tracked then suddenly you need to have like more
um like your hardware requirements start increasing and like consumer hardware might no
longer be able to support something that's like two 300 gigabytes depending on what type of um
consumer hardware you're looking at and if you have like a super high throughput chain then like
it's going to just increase much more quickly as well so like ethereum might be at 100 gigabytes but
your chain might actually end up being like three terabytes just because you're having 100x more um
transactions being processed per second than ethereum so say's answer to this is cdb um cdb is comprised of
i would say two major things um the first part of cdb is a memory maps idl tree and the second part of cdb
is async writes um that happen to both a write ahead log and then eventually to the database um through
this approach that we're taking we're able to get a ton of benefits so the first thing that we talked
about was state storage or we talked about um state storage and state syncing so for state storage we're
actually able to decrease the amount of data that needs to be stored on each full node by 50 percent um
this is because we're able to reduce a lot of the metadata that needs to be stored um on disk and
we're able to do this because with the memory mapped file system uh you need basically just need to have
three separate files that are stored on each full node and these files would be um they basically
represent the ivl tree so this would be the key values like my account has like tensei there'd be one
file that keeps track of that there'd be another file that keeps track of the leaf nodes of this ivl tree
and then there's the third file that keeps track of all the branch nodes because it's a tree data
structure um and with this approach of these three separate files that are being tracked you are able
to eliminate a lot of the metadata that would otherwise be stored with a normal ivl tree that is
stored in your database um so for example this metadata contains things like heights versions left hash
right hash so these are all like integers that start taking up our integers or strings or some other
types of data structures um they start taking up more space on disk and they're very unnecessary
to be storing on each validator and each full node so just by doing this we're able to decrease the
amount of state on each full node by 50 percent which is massive like that's a huge huge win compared
to most other ecosystems um and this is going to dramatically help the amount of state that we're able
to store um the second thing is we are able to get a 12x speed up for state sync um so again state
sync is how long it takes to spin up a full node um because of this m mapped uh structure it's easier
to be exporting data from existing full nodes over to a new full node and it's also easier to import
data because now you have three separate files that are used to create this m mapped um kind of file
system so you're able to paralyze the importing process as well so through this you're able to get a
12x speed up compared to the save v1 approach that's also massive like it goes like in our atlantic
two testament for example it went from taking over four hours to be doing the state sync to be doing
like like getting it done in 20 minutes so that's like um phenomenal as well uh the third thing is
just speed up for commits um the way that a blockchain execution works is you have transactions getting
processed and then afterwards you have transactions that then get written to disk like whatever the
output of that state is let's say my account balance changed by five say that then needs to get
written to the database and then afterwards transaction execution is done and you can move
on to processing the next block um in say's case what we decided to do is you can have transaction
execution finished after these transactions have been updated in memory um and then afterwards you can
just move on to the next block and then what happens in the background is uh these transactions will
get written from memory to a write-ahead log and then from the write-ahead log it'll asynchronously get
written to the actual database um and by doing this you're able to substantially reduce commit time
like we saw 150x speed up for commit um which in practice means that like for higher throughput
workloads there's like 300 milliseconds that you'd be spending just committing to the database um this
then gets just avoided entirely because this all happens behind the scenes so there's i mean yeah sorry
for kind of rambling there there's a ton of innovation we've done over here and this is going to greatly
improve um basically the difficulty of running a full node on say um which will then allow us to keep
supporting greater throughput compared to other ecosystems no just i mean it's amazing again i think
the thing that i've been continuously impressed by say labs team is just the rapid number of iterations
that the team is able to make whether that's on the virtual machine being able to kind of uh add new
capabilities with evm compatibility and make that parallelize uh being able to make optimizations to how
data is stored it is all very impressive and i i think that kind of trend of shipping fast will continue
into 2024 can maybe say or jay as well comment on what are you're personally excited about in 2024
any timelines that you could high level timelines that you can share on when people in the ethereum
community will actually be able to port over applications and take advantage of all the unique
things that we've been talking about on this basis so i've been talking for a while okay i'll pass it
over to someone else all right cool i'll jump in i'll jump in so in terms of timelines uh and jay
could probably be the best one to correct me here uh testnet is pretty imminent um i think jay can
opine on the internal devnet stuff that's already happened i think it's code complete right so
like all of that's kind of um been done teams have deployed applications on uh sav2 on on internal
devnet uh testnet q1 and anticipated subject to governance mainnet q2 so moving pretty quick
um things are coming together the bd guys aren't here but they'll tell you all about the the teams
they're reaching out to and speaking to as well and uh trying to get them over here and uh uh getting
them on on sav2 asap so that's all pretty exciting this um there's a ton of a ton of development
going on on that side yeah i mean to just piggyback off what uh grover was saying before um we
actually got the audit started right now as well so we are at a code complete state in terms of getting
all this functionality working um for the next i guess until the public test that launch happens
there's going to be one priority which is working with these um external teams to deploy on the internal
devnet so we have like a private devnet that several teams have deployed to already working
with a lot more to get them uh deployed as well um and just help resolve any kind of friction points
that come up in the process um afterwards we'll be ready um in q1 and i say q1 because i want to
under promise and then over deliver i think it's probably not going to be at the end of q1 but i'll just
say q1 to keep things um a little flexible so yeah i mean it's going to be sometime in q1 the public
testnet will um come out and then assuming the public testnet uh goes smoothly then we should
be ready to work with uh governance like top the governance process go through and then work with
validators so upgrade pretty shortly afterwards amazing i i i truly think as similarly what happened to
in 2020 and 2021 with ethereum having high gas fees the inability to have these localized fee markets
has really been a challenge for the change for the ethereum chain itself and i think where i'm
uniquely excited about is say ultimately unlocking further capabilities with the ethereum virtual machine
unlocking the ability for those developers that have spent a lot of time on their applications
to ultimately build applications with larger scale uh kind of more robust feature sets because fees will
be cheaper uh they won't be impacted by other applications on the chain i'm super excited for 2024 and
ultimately really bringing the entire industry to mainstream adoption i think it's just going to be amazing to
watch uh hopefully all the normies come back but for all those that are listening i hope this has been
lots of alpha as well on the say ecosystem yeah and two thoughts to just piggyback off what logan was
saying before um the first is that we are so freaking incredibly early right now like i've had so many
conversations in the past month even with people that are extremely crypto native have been around for
several years in crypto they have no clue what paralyzation is or how it works like there's still
so much like work uh to be done to help get the word out about paralyzation i think there's going to be
so many benefits tied to the amount of kind of um performance gains we're able to get from this
improvement in user experience and improvement in new types of applications so first of all i think we're
just so freaking early still um the second uh you're talking about kind of how fast able to um move in
i think in general the way that i think about it is that time is the most critical resource that most
startups have um especially most crypto startups like after you raise money um money is not like the
thing that results in most crypto companies like most crypto projects not um being able to be very
successful i think it's taking like taking the wrong bets essentially um and instead of working
on the things that are the highest priority essentially working on things that are more
interesting technically or that kind of seem like a shinier thing to be focusing on without really
focusing on the one or two things that really move the needle and i mean i grew up in silicon valley
our team is a silicon valley team and like one of the most common um i guess cultural ethos of
silicon valley is this idea that you should ship quickly get user feedback and then iterate versus
trying to just get something perfect built up from day one um and i think that's exactly the approach
that we've been thinking like prioritizing internally it's much better to get something uh that works
get user feedback on that make sure the developers are able to use it to deploy on it like and iterate
from there versus trying to just build something that is like hyper optimized um kind of arguably premature
optimized um from day one very well said guys tons of gigabrain stuff going on here um grover i would
love to hear if you have any um anything else to chime in in regards to save v2 um and then we can kind
of move on to some community questions and some more um kind of light-hearted questions but a ton of alpha
was shared um super thankful that you guys were able to explain this in average joe terms uh i
myself have been on the city team for a long time but every time i hear jay talk i feel like i'm learning
a lot so uh yeah yeah over uh you know sure yeah no i think um i mean this is a ton of stuff uh that
we've been over maybe like if you take away one thing um it's like why what's the point of save v2
um the way i see it is it's primarily solving for one thing which is to to meet project teams where
they're building right now and helping to unblock the development of of actual killer consumer apps
so like it's all pretty clear the stuff we talked about today i think everybody's on the same page
about scalability issues design limitations for the large ecosystem and developer community right
now in crypto which is which is ethereum arguably um it's also really really important i think for
people to understand that like internally we don't see save v2 is like directly competitive with
other solutions to scaling evm right so like succeeding like say succeeding is not at the
expense of of ethereum for example like i think people get into that mentality of like blockchains
killing each other for some reason and i think we need to get past this like grow the pie sort of
mentality is super important um so yeah i mean that's uh one takeaway i'd like everybody to remember
that's kind of like a big piece of of why it's being built in the first place absolutely absolutely
i mean crypto is growing more and more from from cycle to cycle um but i i think everyone feels like
it's it's ripe for innovation and we need to continue innovating on what we are currently you know
building and using every single day so um so yeah i think i have a ton of community questions that
were kind of sourced from twitter and discord um so kind of while i prepare that i i do have a fun
one um are are you guys or are you jay grover uh logan are you dragon ball z fans um it's it's actually
the year of the dragon which is a really good sign uh for say in sayings um jay i know you're a big
one piece fan but um you know do you dabble in any uh dragon ball z yeah so fun fact about me for everyone
listening um when i was like 11 i wanted to be a mangaka so a mangaka is someone who like writes
manga like these are the comics that end up becoming anime um and that was like the main thing
that i was interested in when i was 11. like i used to just draw all the time um and i got started
with shonen jump like shonen jump is this magazine um that has a bunch of manga it's like that has one
piece naruto bleach dragon ball dragon ball z um so yeah i mean to your core point i love uh dragon
ball and dragon ball z i read it for the first time when i was 11. i yeah i mean i i just i i would
say it's probably not my favorite uh manga but i would say it's definitely top five and i think it's
just like really really well written so um yeah huge fan of db and dbz yeah uh i think when i was a kid
definitely like um played through the same story in different playstation 2 games about seven or eight
times right i think some people might relate to that on this call the same story of of goku and
frieza going through to fight cell and then fighting margin boo like over and over and over so yeah um
lots of memes infinite memes available to us here in the marketing department so very excited to uh to
be working with that i love it guys oh logan i saw you on mute i i don't know do do you guys that uh
do you guys uh you know watch dbz or or is this a little too left curve for you no no i i uh i watch
it when i was younger but i need to uh spin it back up so i can be a meme lord so i know what i'm going
to be doing uh in my very late downtime yeah i was thinking we should make it a requirement for all
uh saying you know ambassadors to watch at least like 10 episodes of dbz i think i think that would
be you know quite the bullish uh uh requirement you're uh you're joking here i'm there for it i
mean um everyone might have noticed i've put up a job advert for a uh growth meme lord uh or an
executive ship poster so uh this person should be terminally online highly skilled at meme curation
analysis and proliferation and also must answer questions about dragon ball z uh as part of the
interview so if you're interested just drop me a dm thanks absolutely love to see it um okay so i'm
gonna go ahead and just kind of ask some of the questions um that i've sourced from the community
and then if we still have some time we can kind of bring up some maybe ecosystem teams some some
admirals or um others other members of the same community to to ask themselves um but i do want to
be um you know uh cognizant of time as we're going forward because we do have a chain to build and you
know savey2 isn't going to build itself so um one of the questions i really liked was uh why isn't say
a roll up on ethereum like arbitrom or optimism um you know won't this technically scale ethereum faster
and i know you've been talking a lot jay but i'd love to hear your thoughts on this and since we
have you here i think it'd be a good opportunity yeah yeah absolutely so my goal is to scale the evm
i think there's going to be two different ways that the evm is going to scale um the first is through
roll ups and roll ups will be helping scale use cases um so people will be able to experiment very
quickly have dedicated block space and have their own custom execution layers so i think that roll
ups will be really good for um experimentation and application specific uh block space um the
other way that you're going to scale the evm is by increasing the amount of throughput you're going to
be getting and that's exactly what we're focused on um we think that there's a lot of applications that
you just can't build on the l1 right now or on any existing um ethereum roll up just because you're
bound by the amount of transactions you can get um so that that's what we're focused on right now so the
the question you were asking is like why not just build say is a roll up um the simple answer there
is bandwidth limitations um when you are building a roll up you have off-chain execution happening so
transactions are submitted to the roll-up node then afterwards um this transaction call data this
transaction data is compressed and then written as call data to the ethereum l1 um that that's how it
works right now it's written as call data with prototyping sharding there's going to be these blobs that
are created um and then you're going to be writing it uh to these blobs but it is super high level
like the amount of data that you can write on chain um it's pretty limited um if you do the math like
right now it costs 16 gas per byte of data that's written there's a target of 15 million um gas per
block on ethereum right now and i mean this can go all the way up to 30 30 million if there's like um
more congestion happening but it's and it's like dynamic for the amount of block or gas per block but
the target amount is 15 million so if you do the math around that it comes out to be around 6 000
transactions um that you can actually have happening on roll-ups that are getting built
on ethereum and using ethereum for data availability um and i mean this is assuming that all of the
ethereum l1 block space is being used purely to write this um roll-up data if you have other stuff
happening on the l1 as well like if you have like people uh uh like creating inscriptions people
trading nfts like if you have other activity happening then it becomes even lower than that 6k
number um so yeah like fundamentally if you're trying to use ethereum for uh data availability then
you just really can't get past having a low number of uh transactions you can process per second and
that's why we're building say as a um separate a one instead of trying to build this as a roll-up
one kind of analogy that i like using for a little bit more layman terms is similar to the early days
of the internet you had dial-up internet ultimately that progressed to broadband internet and then went
to fiber optics and you can kind of think of blockchains today as jay was mentioning very similarly
where these early blockchain smart contract platforms are really not propagating a lot of
data in terms of blocks or the amount of transactions that they can send in to the network and it's very
analogous to dial-up internet where there's just a fixed capacity that can ultimately be flowing through
the network at any one time and then once you have broadband type internet engineers can build
more interesting applications like youtube and i think this transition is ultimately going to be
playing out in blockchains as well as you go from these low throughput blockchains to higher
throughput blockchains engineers have the capability to build more interesting applications and applications
that scale to a much higher degree and when you kind of in my point of view are using a layer two as
jay kind of alluded to you're really using it as data compression so you're compressing data to submit
back down to that layer one but even if you're stuck in dial-up internet there's only a certain amount of
data that you can actually compress at a certain point in time you're going to need to transition to
broadband internet just because you're still going to run out of the amount of data and so
i mean my point of view say is one of the leading blockchains that is going to be propagating large
and large amounts of data in these high throughput blockchains which fund makes fundamentally makes it
much different than like an l2 and almost impossible to kind of really run on an ethereum of the world
beautiful um what one really interesting thing i like to bring up so in the last like two or three weeks
uh say has been seeing a ton of this inscription activity i think it's like 200 250 tps something
like that that's been going on constantly and so we hit i think one billion transactions on chain um on
christmas day which is which is cool um but uh that is i think uh if you look at l2 beat right now about
you know that's that's more than entirety of all these l2s are doing it together right i think it's just such
an interesting stat uh it's uh it's it's really cool i think there's another two inscription projects
coming out pretty soon as well like just uh launching a tirade of transactions let's say so
it's going to be really cool as a load test to see uh see how things play out but uh so far so good
i think no one noticed that's the cool thing um everybody's been playing with uh coins on mass report
but didn't uh didn't notice anything like related to all of that activity going on in the background so
that's uh that's really cool too it's good chain red chain red chain good uh you know super like
stoked to see the performance of of say and especially with the inscriptions and all the
activity the chain has just been skipping along without a beat so um i grover while i kind of have
you here on the mic as well i had a great question from a community member named kachu um he or she asks
uh you know how does it feel to see the transformation from kind of the insane
flood we had in august to the awesome vibes that we have now um honest uninhabited answer please um
you know uh how do you feel about the trend shift that we've had um you know it it was kind of bleak
there for a while and now um you know we're going super saiyan 2 so right uh what do you think about that
yeah and there's a good analogy of goku taking a senzu bean right but i think um i think there's a
lot of uh lessons that we learned from that like like jay was saying like you know get feedback learn
iterate move forward so both labs and foundation like a ton from that launch experience like an
unbelievable amount so i think we can see pretty clearly the path ahead from this point um personally
i'm really really happy to see a core community is forming which which are actually want to use the
chain and are invested in say taking the lead in a ton of different initiatives um you know we saw
like the saiyan guys like made an announcement about like funding tooling on say like these kinds of
like decentralized sort of communities popping up which is just really really awesome um happening
quicker than i'd actually anticipated if i'm honest uh like personally so it's really cool to see that
happening and um yeah just uh looks looks super bullish from here but um yeah that's my two cents
also quick commentary i think in general whether it's crypto or anything in the world people just
don't like new things uh and there's always i think tech particularly is really because it's software we
can iterate quickly we can move fast and it's been amazing just how the crypto community even though we
all want kind of global adoption is not very um at least has been limited open-minded to alternative
kind of scaling solutions and now i think as the narrative is really shifting to these high
throughput blockchains and say being a part of that center stage uh it's pretty funny to me how people
will switch on a dime but the vibes people generally are skeptical of new things uh but after a while of
them coming around uh people would get very excited about them and ultimately new possibilities come
about so kudos to the say team uh say labs uh and say foundation because uh the fun is hard but uh you
guys put your head down and are doing remarkable things so kudos to the entire team thank you so
much for the kind words logan and i i tend to absolutely agree like real strong community is
is absolutely forged through hardship and you know despite everything that happened with you know
the beginning of say and the launch of say the say labs team never took their foot off the gas pedal
and they continue to build they continue to innovate and they continue to grow
which kind of is a testament to the resiliency of like the say labs team so um you know you absolutely
love to see it so kind of moving on to uh you know a couple final questions maybe this might be the last
one depending on how we're doing on time um a question from lefty is if you have to summarize the
benefits and vision of say what would they be um and i think this would be kind of a good question for j
say um you know you know where do you see say in the future and um you know what are the benefits
uh of building on say so yeah so right now i think there's a fundamental kind of divide between
web 2 and web 3. um if you're trying to do anything on web 3 like doing anything at all on chain right now
on chain right now both as a developer and as a user it's super difficult and i would say it's like
orders of magnitude more difficult than doing that same thing in a web 2 way i don't think this has to
be the case at all i think there are several things around uh performance around user experience around
adoption and education um that can improve and that will result in it being a very similar experience
between web 2 and web 3. um and that's kind of what i'm working towards like i think we really need
to have better infrastructure getting built so that we can have essentially i mean ideally there would be
no difference between deploying something to an aws server which is like what happens in web 2 right now
and deploying it on chain and both the developer experience and the user experience over there
um would be the same um in terms of like what i see happening and like what i kind of see the um value
proposition for say uh being in like the next couple of years um because we're focusing on paralyzing the evm
what i see happening in the next year is there's going to be a lot more um new types of applications
getting built by crypto native developers that really need to leverage higher performance like
if you're limited by 30 tps then there's only a certain class of applications you can really build
so in the next couple of years i think there's going to be a lot more experimentation happening
um i mean right off the top of my head some things that seem like they'd be kind of no-brainers to start
building on say um one thing would be like an order book based exchange where you need to have
thousands of tps to really be offering a good user experience that's not something you can build on
ethereum l1 right now that's something that you can build on say um another type of application would
be a game where you have every single like user action that's happening getting written directly
on chain um that's not something that you can build on ethereum right now so what ends up happening is
there just ends up being uh tokens of some kind either fungible or non-fungible tokens which are
traded on chain and then most of the activity ends up happening off chain so i think there's like
applications that really haven't been explored right now um that's what people will be using
and iterating on in uh the next year or two
very well said and and i i think the entire state and community and myself agree um you know the the
entire crypto space is ripe for innovation and um you know now that the right infrastructure is being
built out i'm excited to see like the dApps that are going to be uh coming out of this kind of new era
of blockchains and you know financial infrastructure um so i think this would be a good place to kind of
wrap it up and provide some closing thoughts i apologize for kind of the say ecosystem teams and
other people that wanted to come up and ask um some questions but i do want to be cognizant of time and
make sure that our best developer um you know kind of has the time to to rest and you know uh work on
save you too so yeah if logan grover jay you have any final uh closing thoughts before i wrap it up
and we kind of rug the spaces uh please feel free to uh go ahead and say anything now i just wanted to
say thank you uh for inviting me uh super excited for say 2024 and all the things that are to come i hope
everybody had a wonderful holidays with family and friends and has a beautiful new year
thank you logan so the first thought that i had was please give logan a follow if you're not
following him already this man is machine he has a ton of really good thoughts so definitely give him
a follow um and yeah i mean besides that thank you everyone for joining uh the spaces i am sure
there's a ton of folks that had questions um if you have any outstanding questions feel free to just
dm me i'd be happy to engage on twitter um feel free to just reply to my tweets as well and i'll
try to be as responsive as i can um but yeah thank you i'll keep it very very short uh 2024 is going
to be sick thanks everyone for uh coming to the space cheers i love it i love it thank you guys for
all the alpha and thank you jay for taking the time to kind of talk to the same community directly and
explain all the awesome things that are coming for say so i won't chat too much i think this is a good
place to end it have a happy new year guys and we will see you in discord and on twitter cheers everybody
Host
Sei 🔴💨
from
Sei Network
