it appears to be working right okay good we've got another Twitter spaces so
that's happening on the left-hand side Alec can you please let me know if
things are sounding good here fantastic
hello Twitter people help or people
if you want to watch the present presentation if you want to see the
slides and you're in the Twitter spaces there's still a hundred slots left in the
discord unfortunately this limitations to what we can do a discord but it is our
home base is where we come from so we're you know we feel a sense of loyalty to
discord and it's a really fantastic technology so we're in here if you want
to see the presentation you'll have to come over this side I'll let people
switch over bond you're on both sides I can see all right Wow 217 234 hello
people wonderful each spread the word by the way we had to we had to stop the
other Twitter recording because it wasn't letting people join and there's now a
new spaces so if you want to join it's right there we'll say there was about a
five-second delay on the Twitter spaces so okay I don't think we're going to be
able to take questions from the Twitter spaces side yeah you can write
them in the in the chat and you can also come into the discord and type them
in there or message bond directly or al or myself on telegram or whatsapp or
send me a care pigeon we'll try to answer as many questions we can but
primarily this is gonna be a presentation a bit of a road map a rewind
of where we've been a projection of where we want to go
and we'll be starting this moment we're just gonna wait a couple minutes people
are still filing into the the Twitter spaces on the left
and it is being recorded so we will we will launch this again we will release
this if you're really busy and you got something to do you got to kids to
school put them in the bed or whatever that's fine we'll record it and and
we'll lease it release it to everyone so they can watch the whole thing with the
yeah the kids can wait all right
okay so we're gonna go live Twitter people there's still 50 spots still in
the discord it's quite cozy in there seats are filling up quite fast everyone
can see my screen and I'm going to jump over guys to a presentation everyone
can see that Alec can you give me a check mark here
can someone give me a check mark and I see the the screen
yeah we are good to go all right we can see the screen thanks bone fantastic
everyone so TJ FT we've been doing this for a long time but we've taken bit of a
break there's been a lot of innovation happening a bit tenser that has been so
much bigger than the open as a foundation and so we've taken a step
back and let a lot of teams and companies come into the ecosystem and
really own it which is in the end that's our goal we tweeted about this is
that we know the openness of foundation goal is to die talk a bit about that in
this presentation it's a bit of a paradoxical intention for a corporation
but we do really we do really believe in decentralization and that means
getting out of the way so I don't know if you guys saw GTC fantastic
presentation by Nvidia they have launched Blackwell which is the fastest
GPU in the world and they're putting that in data centers 32,000 strong 645
exa flops of AI performance behind the scenes behind bit tenser the gargantuan
company that is Nvidia that grows in our market cap every 10 minutes or drops
every 10 minutes they're fast working to build the largest supercomputers the
world's ever seen and gift them to a couple individuals this is a very
expensive machine no one in this call no it is open tensor no company in the
ecosystem will ever own 32,000 GPUs on a subnet and I want to talk about why that's
really important that's a nice chip who's who has seven trillion dollars to
buy it there's only really one man in the world and one organization in the
world that can buy 32 64,000 of Nvidia's best chips and eventually it's
going to be put behind this one individual or a couple individuals
organizations that are top-down and how could we possibly compete with that how
do we stop the inevitable outcome of seven trillion dollars worth of the
world's compute being cloistered into a single GPU cluster owned by one
individual or a number of individuals like let's take a look at this this is
open AI on the left 700 employees I'm sure not even many of them have equity
well maybe all of them have equity but maybe not very much compared to the the
the founders who situates themselves behind the scenes and offer profit and
that is controlled by a really small board on the right hand side we have
bit tensor it's hard to calculate how many people are in our ecosystem but
this is a number of non-zero wallets we are our extension at Chrome is isn't a
20,000 users we could probably be pretty sure that that's those are individuals
as a company of people that actually own a piece of this infrastructure
cryptocurrencies have the ability to exponentiate ownership of core
technology beyond anything that's possible within a centralized
organization like open AI I mean I'm doing I'm really doing them a favor by
representing it as a in this case about a thousand thoughts on the right hand
side this is a hundred thousand of bit tensor and growing because in reality
the the system that's being designed by open AI is is you know centralized or
corporation it's a normal structure it's a normal corporate structure and it
has a very top-down hierarchical design so all of these chips that are being
built they're all going to situate themselves underneath this company and
we need to stop this for a number of reasons the main reason is that all of
the value will be sucked up into the hands of these people and all of the
control will be sucked up into the hands of these people and we need to
make sure that we're building an ecosystem and an ownership layer where
anybody in the world can contribute and own these chips so the tensor is
fundamentally an open ownership layer it's an aggregation of resources layer
that's why we don't need seven trillion dollars because we don't need it seven
seven trillion dollars individually we can compile it from many many people and
stitches together with protocols it's what Bitcoin did to produce this
largest supercomputer in the world bigger than than the seven trillion
dollar ask that Sam Altman has so not only that
allocating seven trillion dollars and buying a bunch of GPUs is in the end
gonna be really inefficient it's gonna be a really inefficient way of organizing
all of that capital and bit tensor is also fundamentally a market mechanism it's
a competitive place where we only pay for what we want so that we can
allocate way more productivity we can produce so much more with so much less
and that's what I'm gonna talk about here what we've been doing over the
last four to six months primarily we've been working on allocating resources and
and showcasing what bit tensor can do what has what it has been doing where it
can go so you're only in 2023 but this is a bit of a story of where we've come
and where we've come from in 2021 we we just had a single subnet in 2022 we
also only had a single subnet around information and 2022 is around prompting
2023 early 23 it was sort of a chat GPT interface we were experimenting with the
power of ownership for first computation of incentivized computation and we
learned a lot and then in late 23 we let others contribute and own bit tensor and
innovate on the way in which we could aggregate all sorts of different digital
commodities that's what bit tensor is it's a it's a it's an aggregator it's
an optimizer of digital commodities it's a technology which can plug itself into
the servers that are owned by open AI and the 32,000 warehouses of Nvidia we
can seep behind the data center and into every crack and that means that we can
grow bigger than anything that a single company can can possibly raise even if
they're suckling off the tip of the quantitative easing tap on Wall Street
which is where you would get that kind of capital so in 23 we open it up open it
up to a lot of different individuals a lot of different teams and I want to
talk a little bit about what what happened over that those few months to
start in October you know it's February it's March now and let's talk a
little about the innovation that's occurred on on bit tensor before I do
that though I want to talk a little bit about what what exactly is the core
technology of of bit tensor it's abstract validation systems the ability to write
a digital commodity market in an ownership first manner that can validate
anything from compute to human contribution to intelligence inference
speed data collection you name it this is what subnet owners are building they're
building the validation systems those get turned into weights and inject it
onto the chain the most abstract form that we could use to represent these
type of allocation systems and then those turn into incentives incentive curves
which if they're designed properly are very competitive and drive up the thing
we're aggregating or down the thing we're trying to minimize and just like
we're training machine learning models so let's talk a little bit what's going
on I can't obviously talk about every mechanism on bit tensor and as you'll
see by the end of this call there's no way that I'll be able to even know
about them all once we get to where we want to go
optimize recall this is alpha subnet open tensors internal team which is now
external macrocosmos invented effectively the one-way function for incentive
mechanisms where miners were incentivized to scour the web and build
incredibly performant recall systems so they could answer questions it's rag on
steroids it's incentivized rag and every day this this system improves and they
have a front-end and they're building a company on top of this and we only pay
on alpha for the individuals and the computation which can effectively and
and performantly recall data recall information and synthesize it into a
reasonable response alpha could potentially go on to totally scrape the
entire internet and build miners that could rag the entire the entire world
that's just alpha but but some other subnets on bit tensor are optimizing
other things and this is like a core aspect of what bit tensor can do so we
can maximize recall ability we can also maximize the speed have performant if
we want to build really really fast front-ends search engines like civil
comm I think a million organic queries per day passing through its validation
system passing through its validators to its miners where the miners are being
incentivized to return responses that are accurate and intelligent but
incredibly fast they pinpoint this exact thing inference speed I don't know
right now if people are running you know grok chips but eventually that will be
the meta for winning on this subnet and I want to point out this one
specifically because it's probably the first of the tensor some of the truly
public-facing and it shows how we can bring the outside world into an
incentivized architecture infrastructure so that we're not just connected to
individuals that own tau and mine tau and contribute and validate but we're also
connected to the outside world and and monetizing the bandwidth onto the
infrastructure behind each subnet so this is Delta it's maximizing speed
minimizing loss we set out to be an AI company an open ownership AI company that
was split across the world between many individuals I think that we might be the
only people ever that have trained a performant machine learning model in a
distributed setting and distributed incentives incentivized setting in the
world this is this is like a holy grail I think for people in this in this
ecosystem and in this industry of doing decentralized computation just today
subnet 9 released well incentivized into production hardly we could say we owned
it it was given to the network a model that was 40% better than open AI's
equivalently sized billion primer model so and that's on wiki text we don't
know who who made this model I don't think that they're docs in any way we
paid the best price we could for the lowest loss potentially if you went out
there and and tried to hire engineers to lower the loss on a machine learning
model you might pick wrong right you might take that seven trillion dollars
you might buy the the compute cluster you might hire the engineers and they
might not be the ones you need for the job because we can represent exactly
what we want which in this case was minimizing loss we know we can get
that assuming that there's enough competition we can compete with the best
in the world and pay the minimum price minimum loss minimum price
we don't pay you any money if you can't train the machine learning model better
than the previous guy so perhaps a centralized organization would need to
hire thousands of engineers and many thousands of GPUs to accomplish the same
thing because we're a meta layer on top of all of the world's computation we
can design systems that get us exactly the thing we want and we don't even need
to know beforehand what that is we can just represent it as a metric that we're
minimizing or maximizing I mean I also mentioned new subnet as well it's
important to note that these guys the news the news research team best in the
world at fine-tuning machine learning models they built a subnet over the last
five months for a fine-tuning machine learning models specifically seven
billion primer models and they couldn't even compete with their own subnet these
are the best engineers in the world at fine-tuning language models or I mean
and they and they are but apparently there are other people out there that
they can do it when you put capital behind them they can potentially do it
better than then the best of the best so that's really showcases the ability
for bit tensor to not only procure resources but to search them out to find
the individuals that can solve the problems we don't need seven trillion
dollars worth we can optimize it we can be a lot more efficient with the way in
which we distribute capital and we can aggregate data this is data verse but
there's other aggregated aggregators on on bit tensor we don't need to go out
and make any deals with reddit or Twitter we don't need to cut any contracts
with social media companies we just point bit tensor at aggregating data
and we get 10 terabytes in a few months growing quickly I think it's a 300
megs per per day or no 300 gigs is 300 gigabytes of data added to this pile
every day and increasing as more and more individuals come in to the system to
add value to add data into the system we kick out those that cannot there's no
way that you can aggregate data faster than us and this gets pushed into bit
tensor's core ecosystem and can be used by other teams to make better and better
architectures for machine learning models we're really nascent like it's not it's
not there yet we don't have an interconnect between this yet and the
pre-training but that's coming very soon the point is that over the last
five months we've showcased that the human consensus which is bit tensor's
core technology is expressible so pardon me it can express many many many
different core digital commodities and bring them into the bit tensor ecosystem
you know at at glorious rates and incredibly if incredible efficiency we
can also accumulate the compute that we need to solve problems hey well why don't
we just point a subnet at stealing away in videos chips from them and if
anybody has any let lying around we can also use those this is neural
Internet's Senate 27 aggregating computations got over 200 GPUs on it it
only just started to click a couple couple weeks ago and and is increasing
you know an exponential rate in the amount of compute that it can
aggregate and bring into the ecosystem so if we just want GPUs yeah we can get
GPUs as well and we can apply them and give them back to towel holders so
people that owned how get access to all of these resources and our goal is to
push the price of them down so that there's no better place in the world to
go and access everything you need to be an owner of the next Industrial
Revolution of artificial intelligence
just 32 right now really just a bootstrapping phase for what bit tensor
is and can be we know there's a lot of issues in the ecosystem I know that
there's a lot of people that are trying to figure out how to use our technology
because we have open borders and the system is permeable we have to worry
about adversarial aspects of everything we do it's really like machine learning
on steroids it's way harder to build incentive mechanisms you know the the
technology that we're the substrate that we're working with here is
fighting against you and with you as soon as you can click those incentives
in place the the best teams I've talked to have focused on this core technology
I'm not putting brands on this presentation because I want to focus on
the core technology of open tensor and just you know our symbology and and
really the abstraction layer that we have built but there are a lot of
different teams that are winning at this becoming the best in the world at
incentive design and we have the best technology for them to use to build
whatever it is they want whether it's a back end for a company for a company or
it's just aggregating value into the ecosystem becoming the biggest
computational providers in the world or training machine learning models more
efficiently than open AI in a matter of months that's just 32 and we're
going to go a lot farther in 2024 we can we can add up compute as well we
can paralyze and that was it's the whole thing they kind of did but let's put
more cut of cords let's put more cut of course let's let's drive this this in
design of technology to the moon and they've exponentiated in a direction we
found that I think with bit tensor we have incentivized we're the best
incentivized computators in the world computers this is 1024 I think that we
can get there by this year 1024 unique incentive mechanisms competing in a long
tail mechanism to add value into the tensor ecosystem and specifically add
value back towards towel holders
we can find the algorithms with a long tail research because we have the people
we have so much much demand to to work with our technology that it is
becoming impossible to even get a slot on the mainnet we want to facilitate way
way way more people we think that the innovation is going to come from just
some random engineer out there in the world who picks up the bit tensor API
works with their own chain gets it on the main chain and starts getting a
trickle of incentive and finds the right incentive system to make everything
click as soon as we get these systems to click they explode seven nine some
that six some that 13 and others on bit tensor have shown what happens when we
simply make them click yes it's hard yes the infrastructure works against you
yes there's weirdos in your discord razzing you out and trying to mess with
you and and you know it's it's the dark crypto anarchistic space but we know
that we're working with the most powerful technology in the world and it's just a
matter of learning how to aggregate it and direct it properly I think that
that's where we're right we are right now at the precipice of a huge
exponential drive in what bit tensor can achieve so in order to get there we
need to remove the root network from bit tensor that's coming very soon the
goal is to build a competitive system where there can be not just 32 but
thousand at least a thousand of potentially thousands of subnets
interconnected and competing for the decisions of every single participant in
the ecosystem so not just the bootstrapping nodes the root nodes that
that have been done have been doing a decent job up until this point but we
need to go beyond them there's no better selector than towel holders there's no
more democratic there's no better democratic system than one where every
single person can have a say and a vote about in in what is incentivized and
what bit tensor produces so that we can so that we can work with really really
really small amounts of tau emission through an incentive mechanism where
people can build a $5 computing ecosystem and and and we can even work
with altruistic systems where there's no incentive and and towel holders can
make those decisions instead of a small group of individuals where you know
albeit have been I've done a good job it does lead to some cronyism and
centralization which we want to get around get away from so this is this is
this is our goal for for for 2024 it's it's in the wraps we have some papers to
show case coming out pretty soon around our proposal but of course we don't get
to decide everything so I want to talk about how how we're going to be putting
our proposal through for scaling tau before we launched this proposal we we
had the following design loaded that's going to be pushed to the main chain
which will allow any one of these groups try and for it is mostly open tensor
employees the Senate is a bit like our oligarchic plutarchic arm and the
subnets are in many ways the most broad group of individuals but in the
end all of them are being selected by a towel holders and every in any of
these groups can can propose a change to the bit tensor ecosystem and make it make
it go through so in the next two or three weeks we will have our proposal
ready but it won't be able to go through until this system is in place
one where it's a loving level playing field for anyone to participate in the
pseudo democracy of of the bit tensor ecosystem so yeah guys that's that's
a rundown of where we where we've been where we're going and why and then I
think now we're just going to transition into sort of like an open
conversation people can come on stage ask questions that's that's it thanks for
coming to the talk and let's let's do more of a open forum great other guy one
the first question is gonna be we know we mentioned proof of stake right in the
Twitter and a lot of people are curious what is that going to entail yeah it's a
really it's a good question we purposely let it left it out of this because we
believe that there are more important things that we want to decentralize
first proof of stake is a one and a half year timeframe push for the bit
tensor ecosystem this means that at the chain level itself we want to move to a
system where anybody can contribute to the actual block transition and run a
full node but you know after what we saw last week where we got ourselves into a
bit of a pickle we we don't believe that we're ready to announce proof of stake
just yet and we want to focus more on economic decentralization and emission
control and decentralization doesn't mean that's not happening it just means
that we're not ready to to present that
and so what would that essentially mean even though we're not ready to present
what would that mean on the proof of work on the subnet level oh it's not it
was never proof of sorry that it was sorry can you explain yourself bond what
do you mean in terms of proof of work right the chain has been proof of work
up until this point yeah so I think a lot of people got confused by this and our
apologies with the communication here bit tensor uses three proofs of
something so we use a POA proof of authority which means we have
a set of authority nodes that are able to produce blocks at the chain level and
those blocks run the consensus mechanisms of the individual subnets
inside the subnets themselves we use a proof of stake proof of work hybrid
where the miners are producing value which is in the form of work and the
validators are verifying that work using a proof of stake mechanism so
nothing is changing at the consensus mechanism level of subnets we were
speaking just specifically about how we want to have more participants run the
actual chain nodes that we connect to when we send transactions specifically
send way transactions yes so so hopefully that clears things up for people
and there's no more confusion between those dual hybrid incentive
mechanisms or sort of proof valid verification systems I think one thing
to add here is that effectively when we transition to a proof of stake system then
that's really sort of I think in a lot of ways the final nail in decent
foundation was sort of the foundation doesn't really own anything anymore if
the foundation were to disappear off the planet at that point wouldn't really make
a difference because everyone else would be owning all of the values and all of
everything that we don't really need to sort of exist to keep things running
yeah precisely we we still as an organization run the line share of the
then ensure block production hence why it fell on our shoulders the other week
when things didn't work and that's something that we are moving away from
at some point in q4 or q1 of next year that's the intention but where we want
to focus again I want to talk a little bit about why we focused on economic
decentralization first and foremost I think that a lot of projects jump the
gun with physical infrastructure decentralization and maybe they need
that because they're afraid of governmental top-down attacks in our case
we're in Canada we're performing in Canada we're pretty sure that we're
doing amazing work for the world and and that we're not doing anything that's
in the shady or nefarious and with the governments would want to crack down on
yet and so more importantly for us is to ensure that the token itself is
decentralized that's why we didn't do a pre-mine that's why we focused on this
long tail for subnets is why we want to get away from root centralization is
because that will ensure that the in there's a distribution a wide distribution
of those that actually hold the bit tensor tau token and from that we can
create all sorts of different types of decentralization using that as the
civil entity the foundation for other forms I mean a good example is like you
know aetherium did a massive pre-mine more than 50% of its token and sure it
has 500,000 validators on on its network but at the end of the day it's a
proof-of-stake protocol and all of the stake is controlled by a very small
group of individuals so even though they have really good infrastructure level
decentralization they don't have great economic decentralization and I think
that's the most important one and we've always been very clear about why
we think that it's more important for individuals to own a piece and make
value from and be able to access to get access to that's the that's the kind of
decentralization we focused on it doesn't mean that we want to throw proof
of stake in the bin it just means that when it comes to priorities we want to
move in it as quickly as we can to get the token into as many people's hands
in a competitive and proof-of-work way
bond is you want to read more questions from the the chat here yeah yeah so
another question is unlimited subnets or 1024 well 1024 might be effectively
unlimited we won't we're not going to do unlimited because I don't think that the
chain has infinite bandwidth you know we have some of the best people in the
world working on this problem and I mean well we've we've toyed with the idea of
potentially working with multiple chains inside of the bit tensor ecosystem
itself and and keeping things like token balances on one chain and if we are
able to accomplish this and I don't want to speak too much about it because
we're not sure we're just doing the due diligence to see if it's possible but
if we could do this we could scale it infinitely and and that's where we want
to go we want to be truly the the internet of intelligence we want to be
the neural internet so but hey we're saying 1024 we're not saying an
infinite number because I think that's a little bit too much for now all right
like that in terms of dynamic tile right without trying to bog down a specific
date when do we expect a proposal for this new fine to be able to vote on we're
actually the bit to proposal is gonna be different than dynamic tile and look
for it in the next month but but what we can't we cannot promise I always come
back to this like we cannot promise that will happen what we want to promise
first is that the proper foundation for democracy is in place right that's the
most important thing is that we have consensus amongst everybody as a
stakeholder about where businesses should go and we have our perspectives
and we want to we will forthrightly argue for them but this isn't a one-way
street so I can't say I will be here at this date people will be forewarned for
sure people will know that a proposal is going through and they'll be given
plenty of time and things will be as smooth as possible okay I you know
there's a couple questions here from the Twitter people wondering about the
exchanges obviously we had the breakdown yesterday of what happened with the with
the chain downage but can we briefly touch on that before moving on to some
further questions a post post mortem yeah sure just just just a sentence yeah
well post-mortem went really well so yeah so so effectively we introduced a panic
out-of-bounds error in the uninitialized step of block production which was a
really dangerous place for that to happen on the chain and it stopped our
validators and we needed to recompile the bit tensor binary without the error
and then basically distribute that to many people and it took it took a couple
hours for us to get that on to everybody's validators and then it took a
couple hours more for us to get most people's light nodes and Docker images
working and then it actually took a couple days for us to get the archive
nodes running again because those archive nodes have this they have to
resync a lot of the state and so it just takes a really long time when
wielding massive backups so so that was why you know it took what it took the
time it took and you know our reaction to that was that we needed to double down
on computational verification ensuring like before we run anything that the
simply mathematically impossible for there to be things like out-of-bounds and
we really think we're very close to that Russ provides us a lot of tools to make
that happen and we have a you know a full team at bit tensor right now at open
tensor part of me focusing on on ensuring that we can have that kind of
computational verification and once we have that in place we'll be pretty sure
it won't happen again and not only that but once we have that in place we
can be we can be very confident about allowing way more people to
participate in the governance of bit tensor because we can say hey these are
the rubrics this is these are the bounds like you need to be this meaning to use
this much space or you can only use this much space what you write has to be
computationally sound and provably so and and also we can try the runtime
upgrade before it happens like if we can get these things in place I think it'd be
fantastic for the ecosystem because it means we can even take a step back and
allow for there to be contributions for anyone that might even be better than
than what what Bitcoin has right was just kind of controlled by you know a
single github we can we can properly open ourselves up and be an open source
ecosystem okay that actually I'm seeing a lot of follow-up questions in terms of
liquidity exchanges I know this is something that people ask all the time
what is the open tensor foundation fans on engaging with exchanges and liquidity
related it's it's a big game that we've decided not to play too much into I
think that would that many people in this in this community have done a lot of
amazing work to get us on exchanges and that's a really really really great
thing for increased liquidity but I think that we actually have great liquidity
it's just in many ways it's it's shielded or covered up by the fact that
our you know brothers and sisters on coin market cap pay a lot of money for
market makers to to inject like false liquidity into the ecosystem and we
have a entirely well not entirely but just about entirely organic liquidity
across exchanges so our stance is we will provide as much information we can to
exchanges whether it's audits if they want a name behind something you know
open to the foundation has a name we're registered in Canada we will provide
that we'll provide our names but we are not actively pursuing that ourselves we
want to have them come to us and they are
I think one more thing to add here is that if we actively pursue exchanges then
it sort of flies in the face of our ethos in the sense that we are sort of
saying that this is our token and you need to list our token this is not
really what we're trying to do right this is everyone's token says everyone's
so this is part of the reason as well why we do not participate in exchanges or
which are the changes in the first place so we don't want false liquidity I mean I
think that's it's price manipulation you know we've we've been I think very
authentic as an organization as a network as a community and and that's you know
what will take us through and keep us alive for a long time
exactly yep the next questions are more to do with your recent up in signapore
take up constant I you know with Vitalik playing a guest speaker role do you want
to touch on anything that you discuss there or the value that you felt you
by the trade of for the bit tenser ecosystem to in in Vitalik's eyes well I
was a bit disappointed that he supports world coin I mean I've read his paper
on wanting the crypto industry to get back to its cyberpunk roots and then at
the same time he's just about promoting world coin I didn't I decided not to
bring that up on stage because I just thought you know hey he probably gets a
lot of pushback on all sorts of things and it's not my place I felt very
honored that he was he was there and and talking about decentralized
artificial intelligence and and he's an incredibly intelligent man it's just
it's just something else how deeply he engages with ideas that he only just
partially has knowledge of I don't think that he had the right ideas I
don't think he fully understood it you know he's he's lived squarely in the
crypto sphere and and so his perspective on artificial intelligence is very very
crypto e you know how can we use AI and crypto I know it's people are either on
both sides we're squarely in the middle right it's like we were working with the
the union of the two and I don't think that he exactly got that but I did
speak to him for a while and he I think changed his mind a little bit about what
the tensor was and what we were doing yeah it was it was lovely he's he's
quite a special individual very very socially awkward
well you're the social butterfly so like you can't compare in terms of in
terms of this is quite an interesting question you mentioned no government
pressure yet so this person asked are there any upcoming high-end GPU kyc laws
well you know that there are there are some governments that have been playing
with the idea of you know needing jobs to be registered with the authorities
before they can go through and and there's like the D cell army I think has
the ears of of governments and I don't know why they they do but they're very
good at getting that year and I think that's because governments love a
narrative and a fear narrative is is the most powerful so I think that there will
be pushback from governments around you know the ability to train make it to
train machine learning models we'll just go outside those about those those
borders we really can so even though we're in Canada we can move the the
the base of operations at any time and I think that we might have to do that but
we haven't seen any really real pushback from governments I think they know that
this technology is you know absolutely integral to to the innovation of their
of their economies in the expansion of their economy so they're not they're not
fighting I think they would be kind of stupid to do it but who knows you
know I haven't seen any specific legislation that I'm worried about
our next question is are subnets able to communicate with each other slash
interoperable yet or is this become they absolutely are interoperable it's just
a matter of that interoperability being built by validators and for the first
subnet to to enforce it subnet subnet six is a good example of where this is
actually coming into play it might be done even this week where validators on six
need data from 18 and so 18 validators actually open up an axon so that six
validators can communicate with some 18 to produce synthetic data that is used
for fine-tuning the models because tau is global that interconnectivity exists
it's like economic interconnectivity because with inside of each mechanism
the bandwidth you get is based on your towers or how much tau you're holding
on your validators so just by the fact that there's tau that is globally
represented everywhere means that every validator gets access and can
communicate with itself through its own staking account if that makes sense so
it is inherently connected through tau everywhere it's just it's just really
a matter of writing the Python code to make that happen and like a data versus
data is very valuable will probably be pushed into nine fairly soon I'm not
entirely sure when but but six is definitely using 18 very soon okay
this question is for ship can you expand more on the dev environment sandbox clone
of the main blockchain you referred to you on yesterday's postmortem so going
back post postmortem yeah yeah I can do so um effectively right now what we have
is a depth sandbox a test sandbox and the main chain the depth sandbox sort of
is pretty much our you know developer sort of playground we do all kinds of
stuff in there we bring it up you know bring it down almost twice three times
a day testnet is where everyone's familiar so if everyone's seen testnet
before a lot of people are deploying a testnet I'm sure in quite a few people
are actually already on testnet and about to go into mainnet and if I
don't have the mainnet so our idea after this incident is that we need a
more accurate representation of the mainnet the current testnet really is
just a brand new network that effectively runs on top
whatever they created the real it's not even close to being a representation of
the realms in creating a clone effectively which is basically a network
that is sealed off from the world but effectively does mirror the accounts and
the some of the activity in mainnet um we'll be able to actually test a lot of
really sensitive and really stressy tests things like storage migrations
things like a network stress tests and so on without actually having to worry
about breaking the network itself so that if the clone breaks slowly climb
you know what happened you know what's wrong we can fix it as opposed to for
example deploying karate on mainnet and hoping for the best so I hope that answers the
question comes from Nancy GPT what is your biggest concern and the two-part
question biggest area of excitement for the network was that for me for either
of you both of you actually should add to that yeah scaling into the long tail of
subnets so that there can be hundreds and hundreds and hundreds of teams so the
individuals can own and test out all sorts of different mechanisms at the
same time that's really where we're going to get the most innovation for a
bit tensor that's what really excites me there's only one training network
right now why there should be 30 people should run their own and I
think that that is going to be fantastic for the ecosystem and for the
community specifically to really get to know the technology and so it will make
us have to work really hard to build the best-in-class tools there's other
teams like Manifold that are building best-in-class tools for these people to
use yeah that's that's my that's my excitement really the only thing that
I'm concerned about in the ecosystem is is whether or not we can organize
ourselves and make these pushes as a community towards better and better
better and bettering bit tensor so that we're it's definitely a little worrying
to give away all this power we know we have to do it and will the ecosystem be
able to organize organize itself we we've been pushing this off for a
really long time because we know that it takes time for the community to grow
but I mean I think we can't so even though I'm worried I actually am quite
confident that that we're gonna grow into an amazing nation-state of teams that are
contributing and improving the tensor daily
yeah anything he's not happy either he's not excited he's neither nor I think
yeah no no you dropped off absolutely I'm just kind of I was kind of running
around sort of trying to answer this and that's responding to some folks here
okay you're in the chat it's partially yeah no the one thing I was just gonna
say is that what's coming in the next year and that's what they kind of keep
mine for a lot of people is I'm sensing there's a lot of expectation that a lot
of this should be ready to go and you know why is XYZ happening or how come
you guys aren't fully decentralized yada yada it's important in mind that a lot of
the work that we are coming into now is work that is not only intending to
complete a first step towards the project itself but it's also works kind of
still packed foundation so the project will be decentralized safely and so for
example because been around for like what let's say 13 years at this point
and it's still adding updates we are still rather early in the space and
we're still rather early as project also still rather early in the research on AI
and crypto so a lot of this really comes down to you know how well can we align
these incentives in this next year how well can you align everybody's bottom
lines and it's coming next year and that's really thing you ever have been
focused okay and just to say that you know we aren't avoiding any specific
questions this one's from cause I don't even know how to predict because you know
capital are you going to address the the way copy it's not cosmo the way
copying yeah the way yeah yeah totally yeah we have a commit reveal strategy
coming down the pipeline that we think will work it'll allow some of the
owners to deal with the problem you through hyper parameter changes it's not
a you know it's not an all-in-one solution it will be subnet specific I can
talk at length about how that will be designed and we have you know a lot of
thinking around it and even an implementation we're just not right yet
and it's not ready yet to implement it but yeah that will that will be
something that we we might even push that before we push out the detail
proposal that's the second question right what is the ETA ETA for dynamic tau
like I said I can't give it that's good you don't want me to be able to give it
you you will and also I would say you you don't need me to give it because
we're going to launch dynamic tau over a period of weeks so that everyone is
familiar with the way it works and there's it should be designed in such a
way that that for most people and nothing really changes and that's only if the
proposal goes through right so I can't give you an ETA is the foundation going
to start allocating tokens to new validators we don't like the foundation
itself doesn't allocate tokens individuals do right that that donate
will they delegate tokens to the foundation will those people delegate
their allocate their their tokens to new validators including myself yes I mean I
do depending on whether or not I think that the the validator needs support or
if they're building something that I that I that I want them to build that's
been bit tensors like early-day mechanism for sort of like a liquid
democracy it's still sort of in place but changing but yeah so that's my
answer to that particular question
and bomb if you should we just go down this list is Elon already known with the
answer does we know him we know Elon of course he named his child after us or
something didn't he his next child to me how everyone knows this all right yeah
you probably you went out sounds like Elon Musk son I have no idea I have no
idea I know you don't care is them is the mobile while it's still in the
pipeline yeah yeah mobile wallet yes it's coming very soon I don't want to
speak for that team though how long will the usage of intelligence be free
good good question good question the tensor is transitioning into you know
it's slowly transitioning into a more competitive system where validators will
want you to actually purchase tau and delegate it to them before they give you
that bandwidth but more on that a bit later we have some research coming down
the pipelines on that not any time in the next six months how exactly would
a sub tensor branch of code affect the main chain branches won't change affect
the main chain just it's fine make a branch don't worry when a subnet when a
subnet to make text a video I think it's happening I think omega is actually
building that right now we're gonna get like tau Sora Torah we should we'll get
Torah pretty soon well I'm not so sure if it's exactly soon but it's been
promised I know that the team is working on it this is another reason
why we want to scale subnets is because this team's are like well we're
going to train something that's just so much competition there's so much good
stuff out there on the network that you may have to compete with the the
last best subnet and it's really good right so where can they start to get
these emissions and so we want to build that long tail and allow them to to
participate or way early on are we still bullish we're always we're always
clear-headed clear-pilled do you see any potential competitors we're gonna grow
them inside of the ecosystem that's we don't focus outside we just focus on
making sure that the system is competitive at every level whether or not
it's at the government level at the the mechanism level we want the competitors
to be here inside bit tensor and focusing on building from the inside out so do I
see any competitors don't care don't know I don't think so care to explain
what happened with some of three hotkey incident oh I don't know I need some
more information on that when tau ETF is better to yeah yeah is there is
Barry on the call he can answer that question right let's go down there's a
lot of questions here this is fun this is like you know it's like rapid fire
what's when you were talking about weirdos in the channel were you
specifically talking about me yeah yeah it was I've got like several accounts on
discord right now yeah you're right that's probably that is you so what's the
utility of the token really pointed question so each subnet inside of bit
tensor produces some commodity if it's done properly it maximizes the production
of that commodity so so now 27 is incentives is taking the emission of tau
and it's turning that into as many computers as it possibly can aggregate so
that they are accessible to those that are staking inside of the subnet subnet
twenty seven thirteen is is aggregating as much you know it's scraping as much
data and putting into a shared database so that and that is given to the
people that own tau some of the subnets don't really have a very clear you could
say market mechanism utility design they're really more like bootstrapping
but for instance subnet nine is just training machine learning models but I
think that it's not too hard for people to imagine how you can go from hey we
can get the lowest loss of machine learning models but only if but only the
machine learning models we tell them to train towards something where that model
is encrypted in a way that the validators on the subnet get that big you
know or you know subnet subnet for is inference speed right like they're
they're pulling some commodity into the ecosystem and to maximize that thing
and it's available proportionally to those that own tau with inside of the
ecosystem Oh Allah just dropped unfortunately and right now that goes
to everybody who owns that goes effectively is controlled by people that
hold tau in practice it's controlled by the validators where where the tau is
delegated they have access to all of those resources in proportion to the
amount of tau they have so that's the utility of the token utility is looking
to the ability to access it bandwidth wise and you know there's more on this
coming later how we want to make that a lot more specific and directed towards
individuals so individuals can selectively pick where they want to
allocate their their tau control and extract the most amount of value I think
that's that's a requirement for bit tensor or the incentives to hold about is
it going to be the currency to use the intelligence yes when collaboration with
italic I don't have any time soon when his business her like she skipped over one
question oh yeah because Florence asked 1024 seems like considerable challenge
for quality monitoring I think you had a we had our internal discussion I think
yeah very good answer to that well this is what this is why that the real
bottlenecked bit tensor right now is the root network because it's just simply
impossible for individuals to vet all of the ecosystem participants at the same
time 32 is just way too many I mean I've done a lot of it I could take calls all
the time what is it how does it work was or is your front end below blocks
actually etc etc etc but the best way to elicit the proper estimation of the
value of things is through market dynamics through many many participants
being part of that endeavor to find what's good and allocating based on their
utility so you know we want to get there tiny steps at a time but we if we
can if we can basically leverage academic more democratic collaborative
participation of many many many different individuals that means
individual tau participants can vote and and selectively pick the subnets they
want then we can scale so it's more like it's actually an intelligence problem of
the humans within the root network it's not even the potentially cronyism or
it's not even the bandwidth and the changes in the storage and simply the
bandwidth of human time and mental space so you know that's that's why you know
in order to get to 1024 we we think we need something like detail and estelle
should I go more rapid fire your bond this is fun I hope more rapid fire more
rapid fire without getting into the one four digits types of questions I mean
just add a decimal point guys I don't people build new subnets if it
requires a lot of power register exactly well we we lower the cost and
the way we lower cost is we increase supply simple answer there's just way more
demand there is supply and right now yeah mm-hmm and that and that cost
varies right if it's dropped considerably since we last is since it
went to sweat I think 10,000 tile at one point but it's variable so if
there's demand the price the cost of registering to lock up right because
it's not you're not paying you're just locking up your tokens you get those
back but anyways when will Costco on the potential guru podcast with Keith I like
Keith I was number one fan me too soon since it's it's like we're like you know
dancers in a hall like circling each other staring each other across the
room it hasn't happened yet but it will yeah no no comment
can you clarify some tensors of validator distribution status they're fine like
how many validators oh yeah yeah we so we own the lion's share so we own 16
Ella 16 of 20 this is I'll say 16 of 20 foundry owns the other chunk and then
with the idea with the stake is that we can move that out to you know concentric
circles of more and more individuals and really carry the balance at this point
it's like people are like we don't want to do that that's expensive why would
we do that so we've been able to pump all of the value because we don't have
a we don't have a reward for running these validators we simply run them for
free and it's very expensive so we would like people to do that when proof
I question chance for custom subnet consensus very cool question yeah can we
do custom yeah this is what we really want this I mean we're not there yet
like we just we just we just unleashed the ability to write the subnet
mechanisms and it took a little while for people to understand how to use that
and put that into place and and get that working and it is about a hundred
times easier to build a subnet than it is to build a consensus mechanism but I
think people will be able to do it and really it's just a set of mathematical
functions it's a matrix multiplications that are aggregated and applied after
each other to create a you know tractor state based on speculation from the
validators and the bonds and with this and even consensus read about it and
we're pushing more and more research on this and improving it and we would love
to see people that can write their own but right now is it is it a state where
we could just throw that out to the community and they can write their own
consensus mechanisms maybe not yet maybe I'm wrong but either way we need
to build you know that computational verification layer on the chain so that
if somebody comes up with a new consensus mechanism it doesn't just broke
the chain because it takes too long or uses too much space or or you know does
it a panic out of error I think maybe there's a multi-chain future of bit
tensor just throwing that out there we don't know where people can do this
and we don't care you broke your own chain not my chain but we're all
integrated into the same ecosystem I think that might be where things go you
know not in the next couple quarters we skipped over a question so how has been
cancer going to provide services that need to treat sensible and private data
good question privacy is a difficult thing for a decentralized network that is
permissionless because the only way that you can guarantee privacy is if you
encrypt the thing that is publicly available right because it's otherwise
it's not permissionless right anybody can join that's what permissionless means
so if you're sending requests or you're asking them to do work on data that that
is encrypted it needs to be encrypted in such a way that they can they'll they
can work on it with ever not without ever ever knowing what it is otherwise
it's otherwise it's public right so we we stand at the forefront of extremely
open systems so privacy is not really one of our selling points hey we can see
everything maybe transparency is great so yeah the homework homework encryption
there are tools they slow things down and and and hey actually I'll say this
if you build decentralized permissionless systems that are actually privacy
preserving usually they have bounds that make them way more privacy
preserving so much viscerally and verifiably more privacy preserve
preserving then then a centralized entity saying we won't share your data
because we all know that is that behind the scenes is possible to some guy that
can can see your account right so we don't have it and then till we have it
it'll be better than anyone else
well if congestion on the main issue an issue over the last few months how is
scaling handle how is scaling to handle 1024 has then going to be addressed
sorry actually I didn't quite understand the question funky rooted it
what with congestion on the mainnet an issue over the last couple months how
is dealing how are we gonna have right well yeah well if you look at the
congestion that occurred it was it was actually created by a couple things that
we could easily solve so like one of them was some subnets just had miners
setting weights all the time and that's you know I'm gonna create ten
fold as much weight setting as needs be and then also validators were setting
weights very often when we didn't they didn't need to set weights as often and
we can blacklist their ability to do that and then also there was a there was a
situation with the registration where because some owners were not able to set
the alpha parameter they was it was a very sporadic was kind of like this
weird race condition we would go to like 8 tau per registration than 4 tau and so
when it was 4 everyone would register and then would go back to 8 and then
we'll go back to 4 and so on this created this this situation where people
would just run continuous registration scripts and this was basically tossing
the chain so a lot of the bandwidth and scaling problems that we're seeing
today are totally self-created and can be solved to a large degree just by
building better blacklisting at the actual chain level so if you look at how
the chain is doing now it's a lot better and that's just because we pushed a
change a couple weeks ago that that blacklisted people from setting weights
unless they had more than a couple thousand tau and so we we think that we
can look at these things and scale them now by watching you know how big these
these these blocks are and we can even increase the block side if we need to
all right no one can you guys hear me yeah I was gonna say one more thing that
here is that prior to scaling anything like you know 1024 2048 and so on a lot
of these little problems that constant just mentioned obviously have to be
resolved and so you know for anybody who's scared we're gonna scale 1024
tomorrow it's not gonna be one of those things that we're just gonna fire off
and go there's even a lot of checks and balances till we get to that point and
once we are confident then we'll incrementally grow into that point so
we're really gonna get 1024 just as a heads up because I've been hearing a lot
of people who are rather concerned about this yeah no we can't get there
maybe we can't get there until until we have a mechanism for selecting weights
that is more democratic and collaborative and open and distributed
across many people so it's gonna take some time but very soon we'll let you
know how big we're gonna go and we're gonna monitor that and yeah we should
it should be should be totally fine
this next question actually hits me personally because I'm getting a lot of
other projects from different ecosystems trying to get involved with how and it's
thoughts on utilizing polka dot scale using polka dot yeah using polka dot
scale to scale to essentially the yeah yeah like becoming a parachain yeah we
still have the parachain spot we're probably gonna well I don't know we
haven't yet seen that polka dot makes sense for us yet we haven't
actually even really considered it for more than a year I think just about one
year ago we were supposed to launch on polka dot and discover that it was just a
absolute fool's errand to attempt it and we didn't need it and I think from
what I understand there are some kind of pivoting positional things happening in
the polka dot ecosystem Kevin Woods making decisions about how he wants that
technology to work that that may might suit us not even being in parachuting
but being something else in relation to the polka dot ecosystem we we have a lot
of Carino and love towards polka dot for building substrate and we absolutely
admire the technology they built just hasn't yet worked out that it makes
sense for us to to connect in especially when it comes to scaling
because they have block limits they have time they have time limits on the
blocks so when we want to go to 1024 for instance we can't maybe increase the
block size with the block time and that's a real that's a real cost for us
so there are more important things I always come back to this it's like
economic and computational decentralization are the most important
things in of what we're doing right now and so we don't want to make
decisions that put us in a in a bad position to actually scale in in those
ways what are the next upcoming events for bit tensor in terms of in-person
events oh well that's good question we're looking to potentially be a
consensus I know she's gonna be at an AI conference in Singapore all right if
if there's any offices that people would like us to attend we're not
promising that we will attend them but we can maybe make sure some of our
delegates are there you just might be time for I'd be tough for talcon guys
I think so I think so that's that's the next step that's who would go who would
only if there's goodies if there's pizza if shrimp comes yeah if there's a booth
for manifold if there's a wrestling match between me and Artie if it's in Dubai we
can't promise anything maybe ask how often is such open talks happen on bit
sensor more gi of tea more we need more we need more we need more would it be
possible to for route to select certain members from the community to help them
that potential subnets all waiting for the changes to be implemented I think
that I think that we should have you know in the in in the interim here a call
like this potentially every week if not more often where subnets come up and
talk about what they're doing so it's like a micro TGF T for for teams I
think that would be really beneficial like to bootstrap the process of us
democratically and openly and transparently making decisions about why
you know things are things are the way they are I think it would just be great
content also really fun amazing stuff going on I only got to talk about five
and that was just the tip of the iceberg of the mechanisms that people
are doing that could be entire research departments for each of these papers of
each of these subnets and there's this just not enough currently bandwidth for us
to do it all but yeah so so I think that you know a weekly call with people up
here teams coming on and talking about what they're doing I think that would
be fantastic let's do that well you just front ran my announcement because I
was planning on announcing something similar well essentially every every
Tuesday or Wednesday for subnets to come and talk in this TGI of tea
channel or a different dedicated channel where I will host them and we'll
talk about giving them opportunity to talk about their subnets that what
they're doing their incentive mechanisms and just get eyes on what they're
building yeah but yeah I think it's a really good idea I think also be good
for the whole community kind of see what everyone's up to as well and not
sort of have everybody sell it off like this I guess that alternative would
be to put them on a call see them with some lines to see what survives right
exactly and of course the community will be able to ask them questions what are
your thoughts on the idea of reducing APY for periods of underperformance as
under the new proposal instead of putting steak towel at risk it's a great
idea it's a great idea that's all I'll say
what that's their own ledger well that's up to them will be tend to be used as a
currency within the network what it is
work on artificial intelligence anytime soon will bit do a bit tensor launch a
token anytime soon yes we do know we're still staging that though we're not
really yeah yeah the white paper soon yeah give some of these guys this is such
great feedback and an awesome questions there's a lot oh my gosh this is the
best TJ FT ever and you're in it we're in it we've been going on here for or
almost an hour and a half mm-hmm and the question is coming and coming and
coming and I'm all for it but trying to find some ones that we haven't yet
addressed it might be a good one for you guys an easy one what is your
background before creating tau scamming people no I'm just kidding I said Google
and then before that I was at a normal shipping chip company that is well I
don't want to say anything but very fucking cool a company that is now just
about to take the world by storm
the best start mining for someone on a that's an amateur extra topic but Allah
your your background here but oh sorry no I was just gonna say let's have a sexy
background classic academic AI it's homework on distributed computer VMware
and boring on the project yeah boring stuff boring smart kid stuff you know
went to school did the right thing wrote research papers made things work okay had
to be right nerd yeah cool I these questions keep on coming but
well actually I don't mind it we can stick around people are interested in
stick should we stick around as we just keep on answering questions until the
the cows come home let's go for it thoughts on existing tokenized projects
registering subnets mmm it's really interesting actually I don't I'm open to
it I don't know how it connects exactly but I but I'm open to it I just haven't
seen a like a mechanism I'm really just interested at the core of the validation
systems like do they work at maximizing digital commodity production because
that's the thing I tell us that every team I talk to you have a subnet it's
like a bucket and you want to fill the bucket up with as much digital
commodity as possible and the way to do that is to make an incredibly
competitive mechanism one that only rewards when the thing is maximally being
produced and hopefully is continuous and infinite on the in that direction so
people can just keep on getting better at producing more of the digital commodity
that is required more compute more data more loss reduction more inference speed
more whatever you want to aim it at that's the that's the thing that I'm
interested in because it's what accumulates the most amount of value for
token holders in the business or ecosystem that token you can their
websites for all the stuff is sort of irrelevant if you can't get the it's
like oh I started an AI company but my I don't know how to train a machine
learning model it's like well you know my I'm in a local minimum I just it's
collapse everything's NAN's well focus on that focus on exactly getting the
design of the incentive mechanism to work and that's what hopefully the the
community will eventually reward when when the root network is no longer in
control yeah does that need a token maybe right I don't know a lot of
innovation happening a bit tenser I'm open to people trying everything to see
if we can maximize these things
I personally speak to at least two three projects a week who are other blockchains
already and are trying to build on bit tenser and you know some do their homework
and some come to the call asking for permission to build and I tell them
it's permissionless they can do whatever they want but that's the best they're
like huh I would like to build something and then it's just silence it's
it's a silence well why are you asking me you know don't ask me
go build it on the flip side there there are some some ones that actually have
some high-quality ideas that they're looking to start building on test that
and looking forward to seeing what they can come up with but it's it's open to
all right we don't have control we don't have stay over who can or cannot build
all right and I think people don't yet fully understand that so can you maybe
just explain the process yeah well I will work but I mean I have to catch
myself because it's actually not totally true right because the root
network it's a small group of people that make a lot of really big decisions
and they're good people and they do their best but power corrupts and absolute
power corrupts absolutely and centralized power is everything we should
be it should be anathema to us so people do kind of need to get attention
and that's that's not a good game right the or if they have to get attention
they should have to get the attention from the most amount of people and
people should have to put their money where their mouth is so you know that's
that's our main focus that is the only thing I really care about in potential
right now this is what I am focusing on a hundred percent of my time and we're
getting very close so yeah like yes it's permissioned otherwise though like we
don't we don't totally control it and and people can build and I would love to
see people building on testnet and showcasing what they can build there
even though they're not a mainnet they can still really play with this
technology and get it to work
oh hey cause have you gotten any recent legal feedback on whether how open can
be considered a security someone did it they mentioned that someone did it once
and I think that the outcome according to them wasn't yeah I can maybe help on
that one if that's okay yeah with quiet a few people in the project I will not
mention names because we aren't supposed to but the main feedback has been that
we are rather high on the tree in terms of hanging fruits for the sort of
thing meaning that when it comes down to being a security versus a commodity we
are pretty close to being commodities so far according to all their legal
opinions but that's all that stuff is mostly internal to those companies who
are interested not to us specifically yeah yeah sorry no I mean we can be
optimistic also but then like you guys heard the news with the theorem right
recently today so things can flip and change and and centralized organizations
make decisions that are arbitrary we hope and we will argue that we are
becoming probably the most commodity digital currency in the space so yeah
but it's not that we're not again just ask no we're not where it's the less we
don't tell the better that's sort of why our philosophies always stay away from
exchange just stay away from speculation instead of all these things I know it's
so a lot some people really want us to kind of be more involved in that but
that's specifically why we're doing okay on that side it's because we're
it incentivize smart contracts be implemented on the blockchain and it's
so how does this affect rivals I mean we could but we're careful about giving too
many tools to people to just arbitrarily choose how tau is directed just yet
because we know that the first thing people will do is just build ponzi's and
we want to focus this technology towards production of artificial intelligence
and digital commodities and and so we don't think that people need smart
contracts to do that maybe just yet it's yeah it's like most people that want
that access are using it because they don't really want to add value into the
potential ecosystem so we're not yet ready to open that up but we want to see
proposals it doesn't mean that we're pushing behind you know pushing back on
anything and that quality we definitely want to see proposals but we'd like to
see them actually fleshed out before we said yes like we have to be careful
about how this ecosystem directs it's an inflation and submission and where value
any penalties plot for subnets locking up their regs for no reason behind or
beyond a huge paywall sorry what any penalties planned for subnets locking up
their regs for no reason hmm yeah I think I think that the community should push
back on that if it's if it's designed in such a way to you know help very a small
group of individuals and not and not produce a competitive game and do you'll
see that in whether or not that the subnet actually produces value right if
they lock up they lock up registrations it's not competitive how are they
possibly going to be able to produce the most amount of value in the secret
in the subnet people should fork what they're doing people should not should
not support that because it's not actually adding value into towards
stakeholders right so should they have penalties yeah if it's badly done if
it's not producing value absolutely and the penalties will be that you know and
the penalty has come from the community and eventually will totally come from
the community but let's um I think we should wrap it up here bond we've done a
lot of a lot of questions I think we've touched on all sorts of stuff has been
fantastic call I think people understand where we're going now it's kind of
diverging into random questions but we can do this more often we can get way
more questions I clearly it's because there's been a build-up of questions so
we'll just I think we will end it here you know thanks everyone for coming
huge round of applause for being at the biggest DJI T ever it's gonna get bigger
it's gonna get bigger and bigger bigger can't wait to see everyone here let's
get some more people on stage subnet owners if you want to get up here and
indeed be interviewed by me and we talk about your validation system let's do it
I'm so down to talk tech with you guys and see what you're building so yeah
ciao everyone lots of love yeah bye