This discussion centers on the BitTensor ecosystem, specifically addressing criticisms regarding revenue and subsidies, the operational nuances of various subnets (Templar, Basilica, Grail), and the contentious issue of miner emissions and "burns." The conversation highlights the platform's evolution from a technically focused community to one balancing technical development with investor and market demands.
Mind Map
Cliquer pour agrandir
Cliquez pour explorer la carte mentale interactive complète
We're talking everything. We're bringing
all the order. We're talking all the
order. So, um, how do you how do you how
do you want to do this today?
>> No, I was I was I was thinking of going
through some of the the like FUD. I
think that the best critique of Bit Tensor
Tensor
um is, you know, revenue versus subsidization
subsidization
argument that the P that Pine made. And
it's interesting that that's the
argument being made because that's a
really good sign because it's not like
are you doing real things? It's not like
are you a scam? Is this a pump and dump?
It's it's you know are you able to
monetize and go to market with these
commodities that have value already and
you can give bring them to market but
like are you able to finally optimize
down the the revenue so that that it
like there's it's more than the than the
subsidy and that's what you know the
Biter community has been preemptively
focusing on this entire year. So it's
it's ironic by the way that we're going
to be talking about like Templar which
is currently more than anything a
research style subnet like big play but
for most subnets on Bit Tensor shoots as
an example
Liam Targon these other subnets that in
Basilica you know subnets that have been
focusing so much on revenue revenue
surge from Siam like when revenue from
the entire towel community from the dens
and in uh in the the price chats and
stuff like that everyone been pushing
that narrative this entire time. Not me,
by the way. Um um but other people have
been pushing that narrative. And thank
God we actually can come back and really
talk about um you know where we lie on
that curve. And you know their numbers
in the Pine article just not not
correct. They're not up to date. But the
truth is that that we you know when you
look at say like where shoots was in the
previous novelty surge um we we are
moments away from making these protocols
like fully self-sustainable in terms of
emission versus versus outflow and and
distribution which is pretty incredible.
Um and uh and that's such an in such a
short period of time and to to do all of
that on the back of actually building
permissionless protocols is you know
such a feat. So, you can come back in a
couple months and and it'll be another
bogus p piece written by Pine. I'm not
going to say bogus. I would actually I
actually appreciate that they wrote a
very good critique and I I thank them
for that. But obviously, we're going to
get past that and then there'll be
another critique of um revenue and and
inflow versus versus token distribution
and subsidy. Like there's also another
aspect to they didn't really understand
how the subsidy works, right? Like so
the the the tow emission is not the same
thing as just literally what your alpha
token holders are getting in the mining
department. Like 50% of that is gone
right away which goes to the validators.
Another 15% goes to the owners. Um most
of this is yield on the token itself. Um
the the miners are you know he was
saying that we were paying more than AWS
for our our GPUs on shoots which is just
clearly wrong. We know that that's not
true. shoots knows that. Anyways, we'll
come back to that later. That was the
only piece of like FUD that I thought
was kind of interesting and and and
worthwhile and I'm glad that they
brought that one up. Um the other one is
like um pre-mining and and like selling
to to VCs.
You know, it's interesting like if you
build a if you build a permissionless
token, you can't really choose your
bedfellows. Like people are going to
come and they're going to buy your
token. is um and like you any any
project that you build that's open truly
open is you're going to get very weird
people own your token. Not that
Polychain or TCG are weird at all by the
way they're fantastic partners. Um but
it it's an odd way to attack an open
source protocol by saying hey there's
some people that we don't like that that
own your token because like we can't
really control that and that's part of
the design of the system in the first
place. Um what else? Anything else?
But on that revenue thing like for so
again you know it's different like if
everything you um if your only object is
a hammer then everything is a nail
there's different types of protocols
some some need revenue some are able to
instantly generate revenue compute
subnets are very tricky because they can
show revenue almost instantly but it's
in most cases it's asintotic but when
people ask me about revenue for um for
for templar
I just asked them I just asked them when
is open AI or entropic when are they
going to cover their minor emissions [laughter]
[laughter]
so so because it's it's the same to to
me it's the same kind of place so you
know I'm going to ask like yeah that
that's it but again you know it's good
conversations all around and it's been I
think it's been an amazing week for Bit
Tensor and you know I think we just
carried that momentum going but I don't
know the only thing is like I'm probably
one of the key people behind it is guys
let's stop being so feral like dude like
you understand like it's there's no like
we one of the things um so one of the
things are important for any community
especially like communities like ours is
you get maxis and get people that are
absolute lunatics but I think we need to
cross from that maxi stage to the kind
acceptance and like just explain or
block because you know we're doing
amazing things like there's nowhere in
crypto that is pushing out so much [ __ ]
like bit tensor is so like you know some
people are generally curious. It's it's
hard to believe a lot of times, but
let's not attack people on the internet.
Like that's just that's just a crypto
play because again, part of our crowd is
crypto and part of our crowd is AI.
>> Yeah. And AI people already hate crypto.
So even though we're right and we get
into mud fights with like
>> what's that guy's name? It's funny like
the FUD is funny too because it's like
there's nobody that FUDs Bit Tensor than
the internal market people of Bit Tensor
like like other subnet owners coming to
ruin your day, you know, like that's
what you really worry as a subnet.
You're not worried about like Eric
be like, "Oh yeah, is this really a good
subnet?" It's it's it's like, you know,
does does Carol or fish or ref or
something like or papaya or cats or
James come in?
>> Yeah. Yeah. Exactly. Right. Like there's
there's there's no there's no healthier
system than than that which hates itself
more than anything. [laughter]
>> Um but this is we shouldn't be talking
about the tensor at whole. That's the
fire. This is the uh I mean what this at
I think at the beginning of this week uh
or or was it a bit before that like um
you guys had three subnets in the top
five of a mission
>> but they've t some people have taken it
um away from us. We we'll get it back.
We'll get it back. We we plan on but
it's it's been an amazing week. Amazing
two weeks. We were we were we were
contemplating bringing out the pseudo
key to to to destroy Templar because it
had it reached 39% emission. We were we
were concerned. I think that there's no
subnet that's reached 39% emission since
really Apex as being the only sub onens.
Um so
so that that's impressive. I I think I
think that we should we should keep like
a hall of fame for
subnets that that like based on like the
max emission that a subnet has attained.
And so 39% is the is it 39%.
>> I don't know but I know that it needs to
be hardcoded. I I'm I'm tired of
watching emissions go up and down. Just
set just templar to 35%. Let me retire.
Let or let me have a week off.
>> I think we [laughter] disagree.
>> Okay. Okay. But okay, but let's get into
it. So, so we we we Okay, so we have
Templar, Basilica, Grail. Um, but which
one should the nerds be excited about?
>> It's like asking a father which child he prefers.
prefers.
Come on. You can't You cannot tell me to
split my baby three times. We're going
to take everyone through everything. Um,
Templar has been the um Templar Templar
has pole position right now because it's
been alive the longest. It's been alive
the longest and it's it's in some
regards it's so much harder and we will
explain to you what we're doing is so
much harder. We explain to you why. But
the aim of covenant labs is to own the
endto-end intelligence continuum is to
make intelligent the creation the
co-creation the access of intelligence
is to make that it's is to democratize
that so everything it's like you know
when I started when I started templat
then I registered basilica people got
mad oh you're greedy blah blah blah blah
blah but we had to do it because we had
a vision and the vision after training
the 1.2 2 billion parameter model. We
just realized that we're going to need a
lot of compute. So we started Basilica.
Then um what happened next is you know
we used to make fun of um people doing
decentralized pro training because to
our knowledge um the the the thesis
behind um decentralized training is the
compute budget and the compute cost the
cost of training a large model is so
much that
it makes it inaccessible. The data
center is one of the massive problems.
So if we can take away the data center
then we don't need to um we can make it
cheaper. That's that's the whole that's
the whole decentralized training thesis.
You have people like Qua working on um
the context layer but for actual
digitalized training it's a
communication and networking problem
communication networking and um GPU
efficiency. So then we started grill
because um I think for Grock 4 or three
that was the first time on the frontier
that someone actually came out and said
hey guys the compute budget for RL which
is post training is as much as
pre-training like oh then I remember
looking at our scientific advisor like
it was almost at the same time we're
like we got to do we got to do this post
trading so that's it. So it's like you
know it's all part of the same picture.
Um, Templar has actually overshadowed a
massive achievement by Grill and um,
Grill and Fan. Who needs Who Leads
Grill? But anyways, we got slides.
They're not pretty. We found out we were
doing this like two hours ago. So, so
please enjoy them where we take you
through them. Cool. Um,
>> I like it. Keep it authentic.
>> Yeah. Can anyone see Can we see this?
>> Yeah. Yeah,
>> cool. So, this is Templar. It's um
sorry, with Covenant Labs, we have three
subnets. Um I I think the slogan is one
order, one covenant, many orders. So, we
we refer to each of our subnets as
orders. The Well, it's not the crown
jewel. It's just the most it's just the
oldest kid. Obviously, the oldest kid is
bigger and it's, you know, it's just
more welldeveloped. So, we have Templar.
Templar. Do people know Do people know
that the logos for the subnets are the
alpha token symbols?
>> Yes. Yes. The the the logos for the
subnets are the alpha token signals. And
I will let I'll explain how Con screws
you then he unscrews you. Because when
we had this subnet here, we literally
were waiting for which one has a nice
token. Then a couple of months ago, Cons
decides to let anyone choose their
token. But we're not we're not even
going to go into that. Like that really
that really [laughter] because I
remember we were trying to buy
something. It's like, yeah, I don't like
that slot. Okay, let's reg. Yeah. So, I
mean, this is GMA Kai. Um I think the
Basilica one is U Neor and the grill one is
is
>> which looks like the the roof of a
basilica, right? It looks like a church
and it's so there's all these double
entandros have it's multiple meanings
symmetrical meaning it's very cool and
then the grail looks like
the holy grail where it looks like a grail
grail um
um >> okay
>> okay
>> I think it's called fairun
that's the symbol but yeah cool so what
are what are our subnets so again it's
um templar decentralized training
grill decentralized post training and
basilica which is compute. So let me you
know covenant we've started a while it
was our first um it was our um template
was our first subnet um in October I
realized that const was literally went
into so I used to lead the the
blockchainf and you know I just worked
with con so much and I knew that once
prime had dropped open delo he wasn't
going to talk to anyone he wasn't going
to shower until he had solved um um
decentralized trading for bit tensor. So
I worked with him a bit on it and you
know he taught me all the ML I know left
me with GPT3 and yeah we launched the
subnets. So it was it was really
impressive because again you know we
don't I mean apart from working hard I
don't think there's anything special
about me or whatever but it's the
environment of Bit Tensor because while
others can plan and they can you know
stage test net or whatever from the 7th
of November to the till sometime in
January Templar has always had runs. So
this allowed us to really test um things
in adversarial environments. Now what's
important for us is not even as much as
you know the sensorized training because
for us um I I I I always had this
insecurity that you know I'm not an ML
guy I'm never going to be able to do
this but I think me not being an ML guy
I chose the problem that's more
important to me and it turns out that it
aligns with our goals. It aligns with
you know our aspirations it becomes okay
decentraliz if intelligence is the most
powerful thing then decentralized
training is humanity's last dance
because it's the same thing that we
fought for for ages it's what the
internet tried to do is what Bitcoin
tried to do which if we're honest it
hasn't done it is how do we reclaim
agency it's never going to be about how
Um, you know, it's it's people people
missed the we we we missed the
assignment with Bitcoin. I was like,
"Oh, no, we're going to overthrow JP
Morgan." No, we are not because unless
you're ready to die for it. You're not
that's not that's not a fight you're
already ready to do. But it was always
can we renegotiate the social contract
with the Leviathan? So, we filled with
Bitcoin. Like, you know, we we filled
with Bitcoin. But with decentralized
training, it becomes okay.
Yes, we can create. Yes, we can create
this technology. We can turn the
internet into a data center. But what is
the point? The point is not to beat open
AI. I mean, we can try, but I don't
think it's I I think you you um you miss
the forest from the trees when you try
to do that because the fight is can we
create optionality?
Can people train models cheaply? Can
people train model good models cheaply?
if we get there what what do we what do
we uncover the rewards are massive
because what people don't realize and
you know what I started to realize the
more I do to um um do into training is
the first of all let's start on a very
on um let's start like this the gap
between academia and the frontier grows
exponentially because we you know we
work with some of some of them they they
they don't have compute all they do is
train very tiny models which are not
even representative they can experiment.
The cost of training models you you say
it's 10 billion. Let's say it's 10
billion to produce um um a large model.
But what they don't tell you is the run
costs like 5% of that. There's so much
experimentation that goes through that.
So okay, if we do this then we open up
the market so much more. We open up the
surface area for
of innovation because people do not
pre-train because they say it's too
hard. So that's that's that's one of the
things to do and when we create this market
market
we will be the best at it. So that
that's that's that's what really excites
me about Templar. Grail
Grail is is one of the funniest things
because for every other for every other
um subnet I've started the subnet and
I've um you know I've started it I've
understood it and I've um you know I
could always keep an eye on it but grill
is different um I don't know I I didn't
know any RL pro training I was like you
know but you you can pick up a couple of
things so I voded the first thing then
on the A um Ora joined. I believe he
deleted my repo. He's like, "Fuck this
shit." [laughter]
So, so Grill is pretty much just an like
you know, we're very blessed to have
him. He and we've done amazing things at
Grill. Like, you know, you know what's
crazy is we there comes a point where we
realize, okay, when do we start pushing
things out? Because as you find out,
like we wrote something two months ago
and not two, no, a month ago, published
it. It's training what everyone that our
technology is training what people use
now. So again, it's, you know, we're
we're pretty happy. Now, Basilica,
that's interesting because oh my god, no
one told because I thought that okay,
you know, you've done template
decentralized training, let's pick up
infrastructure. It's a different type of
hell. I've never seen more code in my
life. Like it's it's crazy, but it is
coming together. We are moving quite
quite quickly. So what our whole thesis
on is Basil in Basilica is
um we took on the problem because
there's an econ there's a different kind
of economic problem you have to face
with compute networks is you can do
everything on them but what kind of services
services
um how do you bootstrap both economic
activity revenue and how do you become
profitable on decentralized comput
networks you don't own you don't own any
of the resources so you can't really um
uh control demand supply but One thing
that stood to us from the very start is
selling those GPUs without um renting
out those GPUs without um
it could never make economic sense. We
would have to because most compute
networks they pay $1 and they sell and
they they pay $1 and their tokens and
they and they receive 80 cents in cash
and they bank it. That's token
arbitrage. So the only way in comput
networks can ever return more of their
um more than um more than their inputs
is value added services. So we've built
you know we're doing a couple of things
we've got open core it's quite
successful we're running um we've got
our decentralized pass services
will talk through it
>> can I just happen just just for the
audience like let's talk like raw
numbers and Evan you can jump in here
too because you might know even more
than than Sam like okay so B200's H200s
H100s 4090s 3090s we have all these GPUs
CPU servers etc um you know if if comput
providers come to Basilica and they say
they they have a bunch of these and you
say we'll pay you um some sort of rate
on them by the hour because it's a
continuous network right so it's it's
it's being you're paying them in alpha
tokens as as time goes by right
continuously um what what is the markup
if you were to if you were to rent that
those machines and let's say that you
had the AP you guys do have the API it's
easy to rent SS SSH keys you can go on
the machines like you're not talking
about much markup but just what does the
market look like for these machines?
Because like B200s, for instance, if you
can give me a B200, someone's going to
rent that. There's like a shortage of
B200s. Like the price for B200s is very
high. Um H200's just about the same. So
like they're flying off they're flying
off the the racks here. You know, why do
you say that there's no money in that?
Like is it because there is just, you
know, in relative to making real money,
um the margins are pretty low? uh or or
or is it actually quite impossible? And
and why would that be? Because like
surely there's all these compute
providers that have machines that can
just bring them over to your protocol.
Anyways, Evan, speak to that like like
what are the numbers you see like just
give me some concrete if can you give me
some concrete and numbers? Let me take
let me take that because it's um it's
evan will come on but it's something
that you know we I've actually thought
about a lot is computer networks are
like on I mean things are changing now
in bit tensor but traditionally computer
networks are lazy you understand so what
happens is you are you introduce the
miner is the middleman he's not
connected from supply he's going to get
it from shade form and because your um
because your thing is lazy easy you are
not securing the best price. So you need
a mechanism to push it down. I mean
there is there is a margin but it's very
thin and you need to push your incentive
mechanism to incentivize the actual um
owners of the resources or people lower
down the value chain to be able to um to
be able to on board onto your platform.
That's a very big that's a very big
problem. and we have solutions around
that and even when they do you need to
match it with supply. So it's possible
but it's really really um it's really
really hard and take for example we we
do have solutions to get like B because
we don't have any B2 I mean we do have
B200s come on but they're very sporadic
so we we um in V2 of our incentive
mechanism which um Alex can talk about
or V3 of Alex can talk about is the need
is a being able to dynamically respond
to market conditions because we went a
bit full [ __ ] on the um on the
mechanism now and we don't pay if it's
not rendered out. So what and Basilica
is two bits. So we have a um we have a
central we have a secure cloud and we
have the boss. Now the the aim of the
>> we're not going to get rented out. You
understand? So that that's that's good
from a cost perspective, but it's not
where we want to go because the real
power becomes when you can lean into the
incentive mechanism to do something like
the the supply crunch. So if we could
dynamically say, okay, we can match
someone that wants to pay $5 above
market for B200, that is where the value
is in that. But I I think we we'll be
get um we'll be rolling out the new
incentive mechanism next. bringing this
question to to to Evan, right? Like
Evan, can you answer this question? >> Yeah.
>> Yeah.
>> Right. So the
so what happens with the with the
compute networks the innovation is not
uh is not to you know
get the GPU from somewhere and then and
then just for you know and then just
forward the you know uh the SSH access.
uh when when it comes to uh GPU networks
and compute networks in general uh the
the key components is the contracts
right so so you need to uh so when you
go and negotiate uh with uh you know
with a data center to become a minor or
the miner as Sam said who becomes the
middleman to negotiate a contract uh
they need to have some uh some
longativity right so they they need to
they need to go and and have some
security uh or or feel secure enough to
go and say, "Hey, um I'm I'm going to
buy this commodity and and I know this
commodity is going to be utilized." uh
and this is where the the innovation
comes in because uh and this is why we
see you know we see so you know such a
poor uh I guess performance from in
general from uh decentralized compute
networks is is because it's hard to
provide this uh this kind of security
this kind of enterprise security which
is a business which is a business
problem and and we think that this is
where the added services uh that we're
going to talk a lot about in in the next
slides uh are are actually you know uh
solving a lot of the problems uh because
because they guarantee they guarantee
utilization they guarantee utilization
on uh on how efficiently uh you know uh
those those uh computes uh those compute
workloads are going to run uh they they
guarantee uh they guarantee uh hype uh
and when I say hype I don't mean you
know in a token uh you know in a token
way but but for example you know we we
saw auto auto research for example uh
that you know has uh popped in left and
right uh after car path is uh you know
um uh you know broadcast and and and and
people wanted to play with this right
and if you if you go and see for example
you as a researcher to go and rent an
H200 my god this is you know this is a
this is this an expense uh you know uh
for you but if you you know if you if
you go to a compute network that says
you know what just run this and and and
we going to guarantee utilization and
then we we can go back to the you know
uh to the uh to the miners and to the
data centers providers and say you know
what we we can guarantee a huge
percentage of of this utilization. Uh
this is where the the novelty uh you
know uh comes in. Um so yeah it's a it's
more of a old school novelty uh an old
school problem uh that requires you know
new school novelty uh I would say and
and and it's mostly business so it's not
it's not it's not the kind of novelty
that you will see from templar and grail
but it's it's uh it's literally business
novelty uh in order to make it uh you
know make it a success. So right now how
does the mechanism on basilica work?
It's it's it's similar to the other
comput networks. Um but you guys are not
using TE or you are using TE and like
you know my understanding you know the
the problems that other computer
networks came against was um was really
like there's so many ways of spoofing
machines and so like did you guys solve
that in a novel way? Um it sounds like
Alec you want to jump in?
>> Yeah. Yeah. I'd love to.
>> Yeah, go. [laughter]
>> So, my my background is actually I was I
was mining in Liam. I was mining in
Shoots Targon and then Sam found me and
now I talk to Evan all day. But um the
yeah I think like the problems that we
saw with the other comput subnets was
that it's just really hard to basically
specify like a model of GPU that you
want and you know if that's the only
criteria you're looking for there's so
much sort of variability that goes into
a GPU instance right so if I just say
hey you know I want an H200 it needs to
be bare metal with TDX enabled um and
then run our you know TVM on top of that
like you know that doesn't specify the
geography that doesn't specify the type
of network that it could be on um it
doesn't specify the terms that I have it
on. So if you know we take that in and
you know say we pay $4 an hour for it
because it's super scarce, we want it
and we're like okay we have our H200 and
then you know another subnet goes and
launches their product on that. um we
have no way of ensuring that that's not
a spot instance from you know a provider
in Asia right and if that instance gets
taken back by that provider because
there's a shortage right now and we lose
all of the customers data there's like
no ability to get them a new node it's
just like a a huge problem right and
then you know sort of back to the
numbers that I think you were looking
for like you know if you take a spot
instance you can get it for 50% of the
price of an on demand instance so if
miners are going in and they're able to
buy spot from the market and then sell
it to subnets as if it's this ondemand
product. Like they're taking this huge
margin that really shouldn't be there
because like the appropriate price for
spot is what they're, you know, very
close to what they're getting it for
from the provider and then if all
they're doing is just simply reselling
us access to that node, like that's a
that's a 5 to 10% margin game that, you
know, is is fairly well trodden in the
like in the traditional market. Like you
have big providers like Hydra Host,
Fluid Stack, um, Shade Form, like
they're all out there and they've got
inroads to every one of the major data
centers and they'll they'll resell that
[ __ ] like all day long like for 5%
margin. So it's like I don't think
there's a ton of value in just focusing
like Evan said on okay here's
infrastructure as a service here's a 5%
markup you can pay in tow like I think
that's a very basic value prop that we
can offer and it's it's useful to people
but it's really that like higher layer
stuff where you know we can we can take
a bunch of infrastructure we can have a
very efficient scheduler in the middle
of all of it and drive like super super
efficient utilization that disperate
teams wouldn't be able to achieve
um independently.
>> So does that mean that you guys have
written your own form of verification or
not? It's not like that on Basilica.
>> You guys basically
>> do you have
>> the GPU verification?
>> Mhm. Yeah. [clears throat] Like those
standard questions, GPU verification,
machine, host, IP. There's some things
you do do, there's some things you can't,
can't,
>> right? Yeah. So I mean spoiler spoiler
alerts because you know we we put so
much effort for the next slides but uh
>> Can we can we follow the slides like the
whole the whole I had people working on this
this
>> I'm just I just I'm just an artist. I
like to follow my ADHD.
>> So the short answer is yes. And you know
if we haven't covered it in the in the
slides then you can ask you can ask again
again >> typical.
>> typical. >> Okay.
>> Okay.
>> Okay. Sorry. So Alex um okay cons you've
messed that slides show up. How do you
want to do it? Do you want to still go
on on Baselica or where does your autism
lead you? Just dig deep man. Where are
we going next?
>> Me or you?
>> Yeah. Which one do we want to start
with? We're on Basilica. Let's we'll do
Basilica now and then maybe we can jump
back to the beginning and then we can I
don't think it matters which order.
>> We'll let we'll let Con guide us.
>> Hi J. So right let me let me let me
start the you know the I guess the you
know go in a pitch mode I guess. Uh so
uh you know so what we what we wanted to
you know uh communicate pretty much on
Basilica uh uh is that you know is is
the next the next stage right and and
the next stage is uh agentic. So uh you
know so as we know uh you know we all
use agents right now and every agent you
know in order to do some real work uh
you know like train a model run an
inference you know run oh I'm hearing my
voice twice uh for some reason uh or or
you know run evals um it needs it needs
compute it needs GPUs right and right
now this you know is literally what we
what we were explaining uh just before
uh uh you know in order to get those GPU
use uh what happens in in pretty much
any any kind of compute network is that
you know you you know you go you know
you you go in a dashboard click around
or maybe you know have a very basic CLI
uh pick a GPU model configure the
networking and then you wait and this is
again you know an old school kind of
designed uh of of an infrastructure
provider which is designed around humans
sitting on a keyboard. Um so in Basilica
we have taken a whole different uh you
know we we have we are flipping this
completely and and and and and we do
something that we just call it you know
GPU as a code right so uh so what this
means um this means that you know uh
agents you know of course they don't
navigate dashboards they don't navigate
uh human things they uh they navigate
code so uh so so so they don't have to
you know go go into I mean we don't even
have a UI for this specific reason,
right? Like, you know, they don't have
to learn our quicks and and and and
specifics. All they do, uh, is that
they, uh, they're using our SDK, uh, and
they just and they just say, you know
what, uh, I want to run this and I need
eight 8 gigs of RAM, 8 gigs of GPU RAM,
for example. And then the Basilica has
been built in such way. We have you know
we have our novel kind ofuler uh whereby
we uh we take all the compute uh that
the miners can give us uh and then we uh
we we utilize it uh we utilize it to
place this this workload uh in you know
in in in a such in a such you know uh
major and utilized specific way. So the
agent doesn't have to think about
infrastructure doesn't have to think
about devops it just keeps building. Uh,
Alex, you you know T go.
>> Yeah. So, it's honestly like that that
focus I think on those higher level
services is is what's going to drive I
think most of the value for Basilica,
right? Like I don't think anyone in Bit
Tensor needs us to, you know, help them
make a runpod account and get access to
a GPU. Like obviously I think there's
value in that we can automate the
accounting, let you pay in tow and just,
you know, your money can sort of stay in
the ecosystem. Um but I think the real
value is in driving consumption of the
global GPU supply. So we see supply
coming from a couple different sources
and I think the biggest and you know
best opportunity we have is with the
miners. So having miners out in the
market that you know have connections to
data centers. They might have contracts
already set up like say you're at a
company and you have like a 512 cluster
or something, you know, it's going to be
idle for a month. Like giving people in
Bitensor access to that infrastructure,
getting paid sort of a fair amount for
it. Um, I think that's all like a lot of
value that we can unlock from miners and
then having all of that backstoppped by
the actual on demand market. So
something that we're able to do is go
and just go to, you know, any GPU cloud
provider, form a partnership with them
where they give us a special pricing
that's below the market rate. So if you
were to go to, let's say, mass compute
and just make an account, you know,
we've got pricing that's sort of 10 20%
lower than what you're typically getting
just if you're just on that website
renting compute. So we can still sell
people the same, you know,
infrastructure as a service for the same
price that they can get it anyways, but
we have that added benefit of there's
money there's margin in it for us. Um,
and then there's the ability for us now
to use that as a baseline for the miners
where if we need elasticity and we, you
know, we need more A100s and miners are
only able to provide it at, you know, a
$150 an hour, but there's this this
whole ondemand supply at a dollar an
hour, um, the scheduler will just
automatically start using the the
infrastructure that's most appropriate,
right? And so there's obviously a lot
more metrics than model of GPU and
price. like I think there's you know
geography there's the spec around the
machine but that concept of basically
using the you know the true on demand
market to baseline the miners and then
having an incentive mechanism that
allows us to essentially enforce an SLA
with the miners it actually allows us to
have the miners compete and bring us
what I would call like useful
infrastructure rather than just bringing
us you know GPU hours so that's concept
>> Yeah. And just just to just add to this,
>> does that mean that you guys actually
play the minor side of Basilica? So you
guys are running the minor on Basilica?
>> No. So >> no.
>> no.
>> Yeah, we we provide the integrations to
the on demand APIs. So like say like for
instance like we we're doing one with
mass compute right now. They give us a
special pricing. We make that those GPUs
available from mass compute on Basilica.
So when miners come in and they provide
their instances, they're they're in the
same pools, right? Like if miners come
in and they say, well, you know, we get
all of our A100s from mass compute. Like
we don't need somebody to go to mass
compute and rent an ondemand GPU for us.
We want them to go and forge a
relationship with a data center we don't
have access to or that doesn't have an
ondemand API and get us that commodity
that's you know like special and and
more valuable right
>> does it mean that you do the
verification through those providers in
some sense
>> we we baseline the market through those
providers right that we when miners
bring us GPU infrastructure there is
there is a technical verification that
needs to happen because like as you know
people will bring you all kinds of
different stuff right
>> yeah and how does that work? How do you
do I mean this was actually what I was
asking earlier, right? And I think it's
like really the difficult hard problem
and you you obviously know it very well.
Um how do you do that? Like you guys
build uh like I a verification script
you run on the machines. It's not TE but
it's your own custom sauce, right?
>> Yeah. So Oh, sorry.
>> So I actually wrote the
>> No, no, no. I So So again, there's a lot
of things. So most copy subnet you have
a verification package that does things.
It does asymmetric checks. So shout out
CMlex. He actually taugh taught me this.
So basically the whole game is I mean
it's not it's not foolproof and you we
have to keep adapting but the whole game
is the checks have to be quicker than
it's the proof of work. So we have
checks for compute, we have checks for
GPU, we have checks for bandwidth and we
have checks for memory. So it's X proof
of work and it has to be cheap enough
for the validator to verify in a lot
shorter time than it takes the miner to
produce it. Then we have a myriad of
other checks that Evan can talk to and
and talk about and spot checks, but
that's the main verification package. We
will be moving to TES. But the truth is
I mean we I started writing the code
like two months ago and I stopped. One
TE code is boring and two there's still
heart bleed. So almost every major T um
almost every major
um TE package now everything it's it has
a major vulnerability. I think CHS have
managed to solve it. I think John was
telling me about it but we didn't have a
solution for it. So it was about okay we
spend this much time building this thing
for it to still not be bulletproof. We
haven't done that yet but Evan you can
talk about the rest of the checks that
are not in we call it veritas that that
are not in veritas.
>> Yeah I mean so we right uh just to uh
structure my thought here. Uh so you you
mentioned GPU tech. So uh so we we we
check uh we check the whether whether
the DPU uh can produce you know uh the
work that it it is advertising CPU uh
memory uh as well uh we check uh as you
said the network uh and we also check
the network in many uh many different
levels not just uh the bandwidth uh we
are also we are also building uh when
when when the miners uh compute uh
becomes part of uh not as a rental but
uh but part of the platform uh we have
even more checks there uh whereby uh we
are we are able to uh you know uh you
know on board uh on board this uh this
minor node on our uh on our pretty much
you know custom VPN uh so we have pretty
much you know literally a global wide
network of of VPN nodes uh that that
they're coming in uh we check uh we
check their their availability uh we
check their storage uh pretty much every
every hardware that we are about to
utilize uh of course you know if it's
docker compatible if it has the runtimes
that we need and if it doesn't we we we
try to you know uh install it so so any
any kind of hardware software uh
resource that we are able to utilize uh
plus uh plus you know security scans as
well on on the OS itself
um yeah everything everything that we're
about to utilize we we we scan it and
and the scan the scan uh uh period takes
about five
five minutes maybe a bit more uh you
know for for the much larger computes
but that's about it. So, so yeah. Yeah.
>> Wow, guys. So, it is a liquid market. I
mean, like that's impressive, guys.
Like, like I hats off like to get that
working is no small feat. There's a
bunch of people have been trying to do
it and like, you know, we've there's a
variety of different ways. Liam has
their own way, TVM on Targon, but to to
also just get it done is is incredible.
So, like hats off to them. Um, I'm I'm
I'm conscious of time. Um, I really want
to speak with with an about about pulse
because this is this is what you were
talking about in terms of something that
flew under the radar. Um, and and like
it's being used by it was used by
cursor. It was mentioned by cursor in
their paper tell us
>> of course of course uh so hey guys
uh so I mean the whole purpose of pulse
is like post training for now right as
you know Sam mentioned before uh what
happened in grog tree was like they
spend the same amount of compute on like
our post training uh that they spent on
like pre-training so I really uh like I
I I think I'm I really believe that like
post training is going to be uh like
major part of like you know model
training future. It's already is all the
agent ticket stuff that you see it's
mostly like RL post training with proper
I mean uh data that they use for
pre-training [snorts] but I mean when
you come here right like u the hardest
part for post training is that you have
so many moving modules right it's not
just the pre-training some of the things
are way easier right compared to
pre-training for example u you know like
a major part of the compute in our post
training is is kind of goes through the
inference part, right? If I want to give
you like you know a quick overview like
in in like of how RL works is that you
have a trainer module and you have a
inference module right and the what
happens is that uh you know trainer
trains the model and inference just
generates like rollouts for the model
those robots are like you know when
you're talking with chat GPD it writes
thinking or something like that those
are the robots that are happening right
or when you're talking with cloud code
it executes a lot of you know code you
know write tools does search all those
stuff right so inference is doing taking
care of that right and it's kind of you
know it's uh interacting with an
environment and you know generates all
those rollouts give it to the trainer
trainer improves the model using those
robots and actually sending a new model
like updated model to the inference
module and this loops keeps happening
right until the model gets really good
right but all these have like its own
complexity. One of the major complexity
they they need to communicate with each
other, right? And the weights needs to
be sent over the network, >> right?
>> right?
>> And weights are really big. I mean, if
you look at like a one trillion
parameter model, right? That's like uh
two terabyte of like like you know like
almost like two terabyte of weights you
need to send that over the network,
right? to to uh to just like you know
for the for the um for the trainer for
the inference modules to be able to do
inferences on that to do agent stuff on
that and everything. So I mean just just
compute like how much two terabytes of
data like it's going to take time to be
sent from like one part of the globe to
another part. It's like crazy. Like it's
like 100 megabyte per second, megabit
per second kind of like
>> thousands and thousands and thousands of
iterations like this where you're you're
getting the miners to run rollouts. You
produce this large data. >> Exactly.
>> Exactly.
>> Do a training step on a large batch of
of rollouts and then you pass the model
state back to the to the miners and that
could be you know terabytes in size. So
it's very it's very large. And so pulse
is basically your guys' method for
compressing that trans that
communication between the trainer and
the inference miners. Right.
>> Yes. Yes. Exactly. And the crazy thing
about it like I'm saying like this part
is way easier than pre-training because
what we realize is that in oil post
training most of the weight updates at
least in the short term are as sparse.
In long term they're sparse but not as
sparse. So what happens that like only
1% of the weights getting updated we
don't need to send all of them but this
wasn't a known phenomena right in the in
the industry if you look at like
intellect two paper which like you know
was came out from prime intellect what
they did they spend like 14 minutes it's
like it's a lot of time like I they
spend 14 minutes sending 30 billion
parameter model weights over the network
that's a huge overhead I mean then like
you you're competing when you're, you
know, competing with OpenAI, which they
they have the O data center, you're
you're you're going to be 10 times to
like 20 times as slower than them in
training or even more. Right. Right. So,
it's going to be that's the largest
bottleneck right now. It's like when you
look at at how you guys are doing
training training in this regard,
>> the biggest bottleneck for training, the
thing that's slowing us down the most is
this bandwidth. It's getting the model
updates across the wire back and forth.
It's not even necessarily like the
rollouts the rollouts scale
horizontally. So if we can solve that
problem that means we can take way more
compute to it
>> and like for rollouts each each of the
rollouts are like a single entity right
you can stream them over the network as
well they don't need to all arrive at
the time right so it's not like
sometimes rollouts can be really big as
well because I mean if you consider
you're generating like you know 100,000
tokens or something like that and you're
having like you know thousands of those
sitting over the network that could
become pretty big as but you can stream
it. It's okay if it is off policy. You
can do a lot of things but you can do
the same with weights because it can
really destabilize the training.
So uh so like and one thing that I want
to mention like the reason that like you
know the inference part the weight sync
we call it like weight sync weight
synchronization right is the major
bottleneck in RL it goes back to another
assumption about RL that like most of
the compute is done in the inference not
in the trainer I mean if you look at the
paper sometimes 90% of the compute is
for inference only 10% is for the
trainer right it it depends like how you
do it but like especially If you go to
agentic stuff and environments become a
bottleneck as well that can even go to
95% right because you need to run your
environment and like I can't talk about
that all the like for for the whole
meeting but uh so that's why like you
know the weight still becomes a
bottleneck so in the current uh
implementation that we have we kind of
got rid of that like 100 times f it's
it's kind of like we got bandwidth
reduction by 100 times and I was telling
you about
>> and So how do you do that? Is is that by
looking at the update and trying to
figure out for the update what are the
parameters of the model that are that
are um most changed during the gradient
update step and then only basically
setting that parse compression. Is that correct?
correct?
>> Yeah. Exactly. Exactly. Exactly. And
it's lossless. So it's like it's just
like only certain number of parameters
changes not all of them. Just like natural
natural
>> and that's a quality of RL training.
It's like when we're doing RL training,
it unlike pre-training [snorts]
and fine-tuning, I suppose, it's very
much more like a Laura in the sense that
it's like a you could there's a very low
rank um representation of that update
because the especially because of the KL
minimization, etc. in the RL update
step, you you you only need to move a
couple parameters and like that's where
the like the really fine-tuned changing
comes from. So just this is incredible
by the way and like it's it was being
used by cursor which is fantastic. It's
really good news that this is going out
to people and they're obviously getting
closer to to like training like this
over the internet because it's it allows
us to scale as far as we can. Question
about where we are with Grail like what
experiments have you run on the network
here? Like what have you seen in terms
of like the amount of inferences we can
get? Like now that this is broken down,
we put this into the wild. Like how big
can we go in terms of machine learning
model that we can pre-train? Can we get
to 30 billion parameters parameters?
Like can you start you know testing
these on aine and and like and training
on RL environments using your network
because that's a big moment right for
Grail you can get incentive directly
from from 120.
Um I mean to be honest I think based on
just this researching that we've done uh
we can even go like one trillion
parameter model right but it the hardest
part then it becomes like but the problem
problem
>> wait I'm not taking that cell pressure
like let's take this off [laughter] I'm
not taking that
72 I mean theoretically we can but like
let's you know let pump our bags before
we do one trillion
>> but 30 billion barometer is going to be
pretty easy, super easy. like it doesn't
need that much uh like that much uh
effort actually if I just get like a few
B200 I can just train I start training
that the the hardest part right now is
like context length right context length
is going to get really big as well and I
want to kind of make sure that the
compute can be heterogenous right like
so it's like if there is uh if they're
using A00 it works if they're using B200
it works and if they're using like M4 on
MacBook like MacBook chip like it works
as well, right? Is it there now that
Grail's verification which is a form of
inference verification using like
similar to top lock is it that still
like the the the original design for
Grail was something along those lines.
But is it that today you guys are doing
a um you know uh a sort of a hashing of
the hidden states etc etc. Is that are
you doing that or not? And and if and if
you are why does it need to be why can't
you use heterogeneous compute?
>> That's a really good question. So the
hashing is good like we came up actually
with a new verification algorithm. I
kind of you know described in a blog as
well. They can go and look at the code
as it's super like uh efficient in a lot
of respects. But the problem is not like
verification of the rollouts per se.
It's verification of the lot
probabilities that they are sending us.
Right? So when you do RL training, we
need to get really technical here.
Right? When you do RL training, you send
the rollouts and you send the
probability of every token that has been
chosen along the network as well, right?
That kind of shows like like the
previous network that has been used for
this inference generation, what
probability it gave to every of these
tokens, right? And this probability
really helps with training you know the
model in the trainer like it's part of
the it's it's it's it's used in
something called importance sampling
ratio right and it's really important
you can you can use it in other side as
well you use the additional information
from the logic dimension per token which
is a lot more right because it
potentially it's like top k top of each
token you can use that to actually train
the model even better
>> yes yes uh like that a stabilizer
training you can get rid of that but
it's kind of it kind of get it can get a
little bit messy so just for sending
that if they're using like a really
different compute right like for example
M4 that that probability can change
drastically because if we go really deep
like it's a lot of stuff happens in
different hardware right that can result
in like different like log probabilities
and even sometimes different inference
kind of you know u uh that can come out
and that can be a stabilized training So
I I can do it but I haven't taken the
risk like I can like engineering wise I
can do it but algorithmic wise I the
hardest part is like I need to find the
like error bond over at least a lot
probability to make sure at least so
just consider we have an we have a uh
malicious agents right like malicious
agent that kind of like ones who screw
us up right they send us a good rollouts
with bad luck probabilities
Right? And if we give like too much of a
you know freedom to them to do that that
can screw us up really bad. Right? So I
need to see like if we are running it
for example or M4 M5 like I need to do
like a really big benchmark on our you
know proof and everything like to make
sure like if there is how much room for
error we give them for the probabilities
right I I I can't come up with like some
stuff like theoretical but I need to run
this larger scale experiment to make
sure like you know this kind of runs up
because it you can't get the values
theoretically for this one it needs to
be really properly benchmarked uh And it
really depends on the problem. It
depends on a lot of parameters. So it's
going to I I need to do a really large
benchmark to set that up. But it's like
it's not like it's definitely like one
of the achievable things that I'm
actually working on like building those
kind of benchmarks. Uh and that that's
kind of like so it's more like
benchmarking rather than like I'd say um
engineering kind of you know innovation
here. But it's a really big benchmark
that I need to run
>> right now. What have you trained on the
network so far? What do those results
look like? Are they significant just in
the in and of themselves?
I mean we we so far we we trained on uh
we trained on math, we trained on code
and right now we are training on GPU
kernels, right? Which is like code that
is running on GPUs, not like CPUs. It
becomes like uh way more complex.
So when we ran on math and code, we
realized like our system is working
right. It's working properly. We got
like 40 uh like for like 40% 60%
improvement like when we were actually
doing it. There's a blog about it. We
ran a lot of other experiment as well.
So they proved that system is working in
that scale like that was like a few
months ago. uh now we are like trying to
increase like this like the scale of the
system so it can handle like agentic
training right and agentic training for
GPU kernels for like for right now at
least like but I mean you can use it for
anything after you know the
infrastructure is done uh and for that
uh we got some really good results
already I mean it's working and it I
mean last time I ran it at least for
kernel correction we got like 60%
improvement I mean like from zero to
like 60% at least with like uh uh quent
tree model right tree like 8 billion
parameter model right but this this is
these are still like I would call them
like not crazy big result but because
like what we are building toward is not
just you know uh beating the benchmark
in these kind of instances we are we are
building toward building useful models
like useful open source models for these
specialized task or even like general
tasks like for post training and what
I'm like really focused on like building
the IMFra and making sure it's
bulletproof which like so far it's it's
it's proven that it is because every
training run that we run like it it
never destabilizes it runs properly and
there's actually so much more innovation
to do that like I'm really excited about
but so it kind of shows that the system
working but like it's all about the
scale we want to scale it so it can it
can actually handle uh the whole agentic
infrastructure it And it can it can
handle like like 1,000 100uh,000 tokens,
right, as a context length or even more
like it can handle like 30 billion
parameter model. Uh and it can handle it
really good like with a stabilized
trainer that like it doesn't get like uh
uh dabilize and you know diverges and
goes bad.
>> So Sam, when will the covenant subnets
be at 0% minor burn? Um,
Um,
it depends. No, no, no, no, no. That's a
good question. No, no, no. I I actually
want to talk about this because, you
know, it's it's it's so Okay, little
rant and I'm conscious of time. It's as
bit tensor, we've had to grow up. This
is my personal opinion. And we started
off very technocratic, very autistic,
minor is God. Then
DTO introduced a creature none of us
knew how to deal deal with. It was
called the investor. And instead of
minor is God or subnet owner is God or
validator is God. We just learned that
we're in this very delicate balance
between the technical people that
produce the resources and people that
produce the capital. Why? So where does
this take me to? Because you know again
somebody I respect a lot in the industry
is John Durban and he put up a post that
you know a lot of subnets burnt but I've
been on two sides of it you know first
of all you drink cons coolage you're
like no you know let it go let it you
know you know pay minus all the
incentives but one
it's not always the right thing. So as a
subnet owner your job is providing token
order value whether you like it or not
like that is your job but it's also
finding so miners are not the cost
reduction exercise they're resource
optimization exercise. So ideally you
should be finding the sweet spot in in
between how okay if your subnet is good
then if your mission is successful
you will you know it'll be worth it but
you need to lean into the miners to
understand how to to make your mission
successful. I don't believe it's paying
them I don't believe it's paying them excessively because again you know
excessively because again you know things are different and if you have to
things are different and if you have to always assume 100% 100% um sell pressure
always assume 100% 100% um sell pressure on everything you pay out. So it's
on everything you pay out. So it's really it's really it's really really
it's really you know it's it's depending on the situation. If we're trading one
on the situation. If we're trading one trillion and grill is number two
trillion and grill is number two subnets, maybe we let the burn down a
subnets, maybe we let the burn down a bit. Like it's it's it's definitely it's
bit. Like it's it's it's definitely it's definitely not, oh, I'm a minor. I
definitely not, oh, I'm a minor. I deserve 100% admission. No, go away. It
deserve 100% admission. No, go away. It doesn't it doesn't work. It it
doesn't it doesn't work. It it definitely can't work like that because
definitely can't work like that because um I can't believe I'm saying this, but
um I can't believe I'm saying this, but you need to you need to respect your
you need to you need to respect your investors, man. Especially we're
investors, man. Especially we're pre-revenue. like you know we can't just
pre-revenue. like you know we can't just go blindly go into oh no you know we're
go blindly go into oh no you know we're going to be cyber punks you know screw
going to be cyber punks you know screw investors no we bit tensor has to grow
investors no we bit tensor has to grow up everyone had to grow up one of the
up everyone had to grow up one of the problems and again I love const for
problems and again I love const for being able to interact um so one of the
being able to interact um so one of the beauties of bit tensor is const is an
beauties of bit tensor is const is an absolute lunatic and he has changed the
absolute lunatic and he has changed the protocol n to 10 times in two years so
protocol n to 10 times in two years so this lets us iterate super quick when
this lets us iterate super quick when dile launched it was perfect for
dile launched it was perfect for builders, but
builders, but people wondered why like investors were
people wondered why like investors were so pissed off. And we were like, hey,
so pissed off. And we were like, hey, you know, um we're okay, we're making
you know, um we're okay, we're making money, but we didn't understand the
money, but we didn't understand the concept of markets hate downwards
concept of markets hate downwards charts. They hate they hate down only
charts. They hate they hate down only charts. And that's how most um most uh
charts. And that's how most um most uh most of the most of the subnets were.
most of the most of the subnets were. But then, you know, Con and some other
But then, you know, Con and some other people went through it and they fixed
people went through it and they fixed it. not fix it but just made it more
it. not fix it but just made it more understanding from an investor
understanding from an investor perspective. So yeah, so anyone like
perspective. So yeah, so anyone like dude I love miners like most of my bit
dude I love miners like most of my bit tensor people are minors but anyone
tensor people are minors but anyone comes to come and disturb me oh god
comes to come and disturb me oh god owner like you know no bunnies like I I
owner like you know no bunnies like I I understand it but you know I have I have
understand it but you know I have I have a I have a I I have a duty to the
a I have a I I have a duty to the mission and it also includes optimizing
mission and it also includes optimizing resources and emissions are a resource
resources and emissions are a resource so you know the team will do the best
so you know the team will do the best they can to step forward. I think no one
they can to step forward. I think no one would disagree
would disagree that the subnet's owner's goal because
that the subnet's owner's goal because they're a line of participant
they're a line of participant is to
is to protect the subnet and and and that
protect the subnet and and and that means at least in a large part the
means at least in a large part the investors of the subnet.
investors of the subnet. um thinking from a different perspective
um thinking from a different perspective from the perspective of betensor itself
from the perspective of betensor itself and like we have rapid in the chat here
and like we have rapid in the chat here and and I know what he will say.
Does it make sense for Bit Tensor itself
make sense for Bit Tensor itself to seek to find out which subnets are
to seek to find out which subnets are burning minor emissions and to reduce
burning minor emissions and to reduce the emission to those subnets based on
the emission to those subnets based on how much they're no now obviously as a
how much they're no now obviously as a subowner you're going to say that's a
subowner you're going to say that's a bad idea right but I'm asking you to put
bad idea right but I'm asking you to put yourself in the shoes of tow
yourself in the shoes of tow the tow itself right imagine
the tow itself right imagine thinking from the perspective of the
thinking from the perspective of the investors of Tao and the emission of
investors of Tao and the emission of Tao. Now I'd like to hear your
Tao. Now I'd like to hear your perspective on that. Um right like
perspective on that. Um right like obviously the subnet but [laughter]
obviously the subnet but [laughter] >> my perspective let me talk about my
>> my perspective let me talk about my perspective you succeeded in so the
perspective you succeeded in so the beautiful thing about detail is
beautiful thing about detail is incentive alignment
incentive alignment you managed to make investors and subnet
you managed to make investors and subnet owner owners really really like you've
owner owners really really like you've cocked us so much like we're bow to the
cocked us so much like we're bow to the protocol but what about the miners and
protocol but what about the miners and validators
validators they don't they they they get a they get
they don't they they they get a they get they get a free Right. So that's one
they get a free Right. So that's one that's that's one thing and honestly
that's that's one thing and honestly like you know we all got we all got um
like you know we all got we all got um we all got like we all got less like we
we all got like we all got less like we all got fake less money like so with
all got fake less money like so with with uh with detail we are prepared to
with uh with detail we are prepared to have less but we can spend far less that
have less but we can spend far less that we could in root network which is okay.
we could in root network which is okay. So that is that is our perspective. Now
So that is that is our perspective. Now from an investor it's very different
from an investor it's very different because what is your expectation of the
because what is your expectation of the subnets? It it's it's all because market
subnets? It it's it's all because market price expectations
price expectations like take for example Templar did
like take for example Templar did something that brought Templar and also
something that brought Templar and also Tagon you know we've done things that
Tagon you know we've done things that have brought billions into the ecosystem
have brought billions into the ecosystem how should our investors look at that
how should our investors look at that look at us how should our investors
look at us how should our investors justify how much we get to spend on the
justify how much we get to spend on the team it's like because dude I I think
team it's like because dude I I think like you could try it but you know you
like you could try it but you know you know how it's going to end up if you
know how it's going to end up if you want to start doing stuff like that like
want to start doing stuff like that like you can't police it is the same reason
you can't police it is the same reason why you know you cannot police that and
why you know you cannot police that and it will it will just it will distract
it will it will just it will distract people from um it will distract people
people from um it will distract people from um from um from building it will
from um from um from building it will just distract you. You can't police
just distract you. You can't police that. So my my my challenge is if you're
that. So my my my challenge is if you're a minor and you think it's unfair dude
a minor and you think it's unfair dude minor spirit go start your own subnet
minor spirit go start your own subnet because right if we if we don't do that
because right if we if we don't do that the economic surplus on the minor side
the economic surplus on the minor side just makes it untenable. So rather than
just makes it untenable. So rather than say you know cut sub just no go miners
say you know cut sub just no go miners you made a lot of money go start your
you made a lot of money go start your own subnet but you what I find with most
own subnet but you what I find with most miners no I'm a minor of course you're
miners no I'm a minor of course you're going to be a minor because it's crazy
going to be a minor because it's crazy profitable you have subnets with where
profitable you have subnets with where striker is striker here you have subets
striker is striker here you have subets with broken incentive mechanisms that
with broken incentive mechanisms that you can make millions of like it just
you can make millions of like it just doesn't make sense so I do understand
doesn't make sense so I do understand people revoling but on templar templar
people revoling but on templar templar is very funny yeah because I was against
is very funny yeah because I was against the burn because I was I was on the cons
the burn because I was I was on the cons Kool-Aid and the biggest Templar holders
Kool-Aid and the biggest Templar holders are OG miners. They literally told me
are OG miners. They literally told me that if I didn't burn because they got
that if I didn't burn because they got out competed. You could argue skill
out competed. You could argue skill issue but we formed a community. We
issue but we formed a community. We mined early. They they they were early
mined early. They they they were early miners and they were like hey if you
miners and they were like hey if you don't start burning this new guys are
don't start burning this new guys are dumping at every epoch. If you don't
dumping at every epoch. If you don't start burning then we're gonna have to
start burning then we're gonna have to dump too. So I'll put this back to you
dump too. So I'll put this back to you cons in that situation what do you do
cons in that situation what do you do because they are by proxy also tow
because they are by proxy also tow holders
holders and if if it was a proof ofstake network
and if if it was a proof ofstake network and there was governance
and there was governance my state compared with their stake would
my state compared with their stake would trump what anyone else anyone else does.
trump what anyone else anyone else does. So I think a larger problem with and I
So I think a larger problem with and I thought I know I know I know we wanted
thought I know I know I know we wanted to solve this a larger problem with
to solve this a larger problem with detail is not even about miners. If we
detail is not even about miners. If we want miners to have that kind of voice,
want miners to have that kind of voice, let them should there must be some kind
let them should there must be some kind of way that they too are incent they're
of way that they too are incent they're aligned long term. Otherwise, let the
aligned long term. Otherwise, let the people that have the long-term view on
people that have the long-term view on the subnet, they get to decide what they
the subnet, they get to decide what they get to take the bigger bets.
Yes. I mean, so so uh this is a healthy debate. um uh you know there's there's
debate. um uh you know there's there's no there's no doubt in my mind that like
no there's no doubt in my mind that like there's no there's no um judgment
there's no there's no um judgment towards the way in which teams run their
towards the way in which teams run their systems right and your your point is
systems right and your your point is taken like what would I do I mean I'm
taken like what would I do I mean I'm not doing that on how fine but like I
not doing that on how fine but like I maybe I'm not in the same situation as
maybe I'm not in the same situation as you and and I it's it's different for
you and and I it's it's different for infant subnets and all these things um
infant subnets and all these things um so let's just say that hypothetically
so let's just say that hypothetically yes the it's the the the incentive
yes the it's the the the incentive alignment is there in the in in
alignment is there in the in in currently for a lot of subnets which are
currently for a lot of subnets which are looking at
looking at um well if for instance let's take a a
um well if for instance let's take a a comput submit like basilica right if you
comput submit like basilica right if you guys don't have demand
guys don't have demand um [snorts] and you guys turn minor
um [snorts] and you guys turn minor emission up to 100%.
emission up to 100%. Uh it doesn't find the equilibrium that
Uh it doesn't find the equilibrium that you're looking for right what what
you're looking for right what what happens where is the equilibrium if you
happens where is the equilibrium if you just turn up minor emissions to 100%.
just turn up minor emissions to 100%. Oh, computer miners are one of the most
Oh, computer miners are one of the most mercenary types of miners. They'll dump
mercenary types of miners. They'll dump everything. [laughter] They would
everything. [laughter] They would literally because also see the idea is
literally because also see the idea is miners are not dumb. It's they get a
miners are not dumb. It's they get a they get optionality over whether or not
they get optionality over whether or not your asset that token that token can
your asset that token that token can appreciate. That's all they look at. If
appreciate. That's all they look at. If they see how easy it is to get then
they see how easy it is to get then you're not a good subnet. Like it's just
you're not a good subnet. Like it's just like it's just supply demand. Like I
like it's just supply demand. Like I talked to a lot of miners, they will
talked to a lot of miners, they will just look at the opportunity. Okay,
just look at the opportunity. Okay, let's let's um let's dump. So it's it's
let's let's um let's dump. So it's it's it's very very different with compute
it's very very different with compute because quite naturally you keep the
because quite naturally you keep the margins low because they're data center
margins low because they're data center people like they will I mean some will
people like they will I mean some will keep but you know most of them will dump
keep but you know most of them will dump it. So I'm not sure that's that's that
it. So I'm not sure that's that's that that's different and there's also
that's different and there's also something to
something to >> how is that different than say shoots
>> how is that different than say shoots right? So shoots has not burned. Um uh
right? So shoots has not burned. Um uh kudos to them by the way. Uh and and in
kudos to them by the way. Uh and and in the perspective of John who's not on the
the perspective of John who's not on the call right now, uh I wish he was. If
call right now, uh I wish he was. If he's here, raise his hand. Maybe we can
he's here, raise his hand. Maybe we can take Rapido if he wants to come up here
take Rapido if he wants to come up here and speak to this to this philosophy is
and speak to this to this philosophy is that it actually just becomes surplus to
that it actually just becomes surplus to your miners. Your miners join your
your miners. Your miners join your community and then they are aligned. Um
community and then they are aligned. Um and they don't they don't sell. Um and
and they don't they don't sell. Um and and and you know shoots was the top
and and you know shoots was the top subnet for the last year and a half um
subnet for the last year and a half um in that [clears throat] in using that
in that [clears throat] in using that philosophy. I'm so happy to like like
philosophy. I'm so happy to like like like like what's your perspective here
like like what's your perspective here if you were on stage maybe you can get
if you were on stage maybe you can get up on stage and I also like to hear from
up on stage and I also like to hear from fish like about minor burn um in general
fish like about minor burn um in general like is it is it is it is is this good
like is it is it is it is is this good for bit tensor is this bad for bit
for bit tensor is this bad for bit tensor um I is it a moot point it's not
tensor um I is it a moot point it's not possible I mean like mechanistically
possible I mean like mechanistically it's almost possible to even detect
it's almost possible to even detect minor burn that's part of the problem
minor burn that's part of the problem we've spoken about this a number of
we've spoken about this a number of times Sam like you could put in um you
times Sam like you could put in um you could say hey look you know a submit
could say hey look you know a submit that burns 50% of minor emissions will
that burns 50% of minor emissions will get 50% less emission from tow we'll try
get 50% less emission from tow we'll try to incentivize people not to burn but
to incentivize people not to burn but what happens is that people just don't
what happens is that people just don't burn but they burn by sending to their
burn but they burn by sending to their own keys and then you have a different
own keys and then you have a different incentive problem where um there's just
incentive problem where um there's just simply not a very good way to measure
simply not a very good way to measure whether or not a subnet is properly
whether or not a subnet is properly using its minor mission and and also it
using its minor mission and and also it also becomes unfair because you have you
also becomes unfair because you have you have a comput subnet where like the the
have a comput subnet where like the the the minor distribution just become sell
the minor distribution just become sell pressure automatically versus say like
pressure automatically versus say like an um a research subnet that you you can
an um a research subnet that you you can turn it up all the way because they're
turn it up all the way because they're not going to sell because they don't
not going to sell because they don't have to sell. Um and and like I'm
have to sell. Um and and like I'm actually really interested to from from
actually really interested to from from like to hear from Fish um and Rapido on
like to hear from Fish um and Rapido on this if they want to come up on stage
this if they want to come up on stage in this in the time we have left on the
in this in the time we have left on the call
or even uh even uh Papaya. No, no, don't get Papaya on. Please get
No, no, don't get Papaya on. Please get fish on.
fish on. Get Fish on. He's the one that started
Get Fish on. He's the one that started burns. Get Fish on. Yes.
burns. Get Fish on. Yes. >> Hey, I'm in an airport. It might be a
>> Hey, I'm in an airport. It might be a bit loud, but I said in the chat, I
bit loud, but I said in the chat, I think it's mostly bad minor burns in
think it's mostly bad minor burns in their current form at least. Mainly
their current form at least. Mainly because of two reasons. one um it means
because of two reasons. one um it means the validation isn't ready for minor
the validation isn't ready for minor competition
competition because we see in a lot of cases I mean
because we see in a lot of cases I mean this this is a probably the best reason
this this is a probably the best reason to burn which is just that your subnet's
to burn which is just that your subnet's broken [clears throat] so if you were to
broken [clears throat] so if you were to turn on emissions the people getting the
turn on emissions the people getting the emissions would be the ones that aren't
emissions would be the ones that aren't providing something valuable so I think
providing something valuable so I think it's mostly up to the sub owners to fix
it's mostly up to the sub owners to fix that to make it so that the product so
that to make it so that the product so the validation isn't broken and that
the validation isn't broken and that minor mission is used towards something
minor mission is used towards something valuable and It's a good indicator of
valuable and It's a good indicator of which subnets are working or not if
which subnets are working or not if their emissions are.
their emissions are. >> Yeah. But do you know do you know the
>> Yeah. But do you know do you know the problem with that? Yeah. It's rooted in
problem with that? Yeah. It's rooted in philosophy that is not rooted in
philosophy that is not rooted in economic realities of running anything.
economic realities of running anything. Okay. Let me let me let me let me put it
Okay. Let me let me let me let me put it let me let me let's take a step back.
let me let me let's take a step back. Who says n why do n why why do miners
Who says n why do n why why do miners need 41%.
need 41%. like okay the way I understand it bit
like okay the way I understand it bit tensor was cyber punk miners ruled then
tensor was cyber punk miners ruled then you know we needed something we needed
you know we needed something we needed something for human consensus so we did
something for human consensus so we did the kings um the subnet owners took the
the kings um the subnet owners took the king's fifth which actually turned out
king's fifth which actually turned out to be 81%. Then we decided okay we need
to be 81%. Then we decided okay we need validators we split it between um
validators we split it between um between um miners and validators. So
between um miners and validators. So what under what economic rule is it that
what under what economic rule is it that we like every single type of commodity
we like every single type of commodity has to spend 18% on that otherwise it's
has to spend 18% on that otherwise it's not working I think I think conversation
not working I think I think conversation because
because >> yeah no but distributed obviously it
>> yeah no but distributed obviously it doesn't have to but why are you on bit
doesn't have to but why are you on bit tensor if you're not I'm not it's the
tensor if you're not I'm not it's the general you know I'm talking to you
general you know I'm talking to you specifically why are you on betensor if
specifically why are you on betensor if you're not going to turn on the minor
you're not going to turn on the minor missions that's like the main innovation
missions that's like the main innovation of bit tensor is the incentive
of bit tensor is the incentive Yes, but it's it it is. But you see,
Yes, but it's it it is. But you see, you're still on that side where
you're still on that side where >> No, no. Look, Fish, I'll actually
>> No, no. Look, Fish, I'll actually challenge you because you started the
challenge you because you started the burns on this argument and it is
burns on this argument and it is economic reality.
economic reality. >> Yeah, but I mean it's the same thing. My
>> Yeah, but I mean it's the same thing. My are yeah I am saying my subnet like yeah
are yeah I am saying my subnet like yeah it means that we're not matured enough
it means that we're not matured enough as a product and it means that we don't
as a product and it means that we don't have a good subnet as much as as shoots.
have a good subnet as much as as shoots. I mean I'll say that it's true because
I mean I'll say that it's true because we can't utilize all of the minor
we can't utilize all of the minor missions as effectively as otherwise.
missions as effectively as otherwise. >> No, but you see that's the problem is
>> No, but you see that's the problem is that that 41% is an arbitrary number.
that that 41% is an arbitrary number. They're different kind of business
They're different kind of business models. You can't p you can't say
models. You can't p you can't say everyone if you do not spend that 40 41%
everyone if you do not spend that 40 41% you're a bad subnet. Take for example
you're a bad subnet. Take for example templates probably smaller small
templates probably smaller small >> it doesn't matter if it's arbitrary not
>> it doesn't matter if it's arbitrary not >> it's just it's it's an amount that
>> it's just it's it's an amount that you're given to be able to use freely to
you're given to be able to use freely to give to the miners to create innovation
give to the miners to create innovation for anything in the world on your
for anything in the world on your subnet. And if you're not using it means
subnet. And if you're not using it means >> that that the validation isn't isn't
>> that that the validation isn't isn't validating something that that can be
validating something that that can be optimized.
optimized. No, no, no. But that's the problem. It's
No, no, no. But that's the problem. It's an arbitrary number. Like I I I
an arbitrary number. Like I I I understand it. Minor so we don't
understand it. Minor so we don't disagree. Miners are important.
disagree. Miners are important. >> Whether it's arbitrary or not, the
>> Whether it's arbitrary or not, the higher the number, the more the
higher the number, the more the optimization.
optimization. >> No, I I disagree. I disagree because you
>> No, I I disagree. I disagree because you can have a broken subnet that does 100%
can have a broken subnet that does 100% and returns no value. Like it so it's
and returns no value. Like it so it's not.
not. >> No, no, no, no, no, fish. It's not about
>> No, no, no, no, no, fish. It's not about the the outcome matters more than the
the the outcome matters more than the process. We should look at outcomes and
process. We should look at outcomes and not the process because like you know if
not the process because like you know if shoots thinks that they need to turn on
shoots thinks that they need to turn on 100% minor emissions I would argue that
100% minor emissions I would argue that shoots should be doing a lot better if
shoots should be doing a lot better if they manage if they manage if they
they manage if they manage if they manage minor emissions because you know
manage minor emissions because you know [laughter]
[laughter] miners will overpay it would never be a
miners will overpay it would never be a good idea to overpay work uh to to to
good idea to overpay work uh to to to overpay resources. It's a it's a
overpay resources. It's a it's a resource optimization.
resource optimization. >> Exactly.
>> Exactly. >> I agree. If they were over if they were
>> I agree. If they were over if they were overpaying miners that would be true,
overpaying miners that would be true, but they're not because they use all of
but they're not because they use all of the compute on their platform and
the compute on their platform and they're able to sell it all and they're
they're able to sell it all and they're able to make a product out of it. So
able to make a product out of it. So that's why that's why I think they're
that's why that's why I think they're >> as
>> as a shoots miner myself, I would challenge
a shoots miner myself, I would challenge I would continue providing the same
I would continue providing the same amount of compute to shoots if they paid
amount of compute to shoots if they paid me significantly less than they are
me significantly less than they are right now. like they could pay me half
right now. like they could pay me half of what they're paying me right now and
of what they're paying me right now and I would continue to provide exactly what
I would continue to provide exactly what I'm doing. And I think right now we're
I'm doing. And I think right now we're in a boom cycle. I think things are
in a boom cycle. I think things are going up. Everyone's loving it. They're
going up. Everyone's loving it. They're getting paid. But
getting paid. But >> the same is true when
>> we have they they have loyalty to us when it's good and then miners just like
when it's good and then miners just like we're not going to lose money for shoots
we're not going to lose money for shoots when things are bad. And I just think
when things are bad. And I just think like
like that proves the point different. Okay,
that proves the point different. Okay, >> that shows the point though that you see
>> that shows the point though that you see they leave when the price goes down. I
they leave when the price goes down. I mean I mean they're at the bare minimum
mean I mean they're at the bare minimum cost basis right now otherwise they
cost basis right now otherwise they wouldn't leave when the price goes down.
wouldn't leave when the price goes down. It's kind of contrary to what you said
It's kind of contrary to what you said but what actually happened when you saw
but what actually happened when you saw that miners pulled their GPUs.
that miners pulled their GPUs. >> I really wish Truths was here. I really
>> I really wish Truths was here. I really wish John was here because TR also has a
wish John was here because TR also has a problem like again I don't understand it
problem like again I don't understand it that much but it's more like in to to me
that much but it's more like in to to me it looks like they have a problem where
it looks like they have a problem where we have this much compute we must use
we have this much compute we must use you understand we have this much compute
you understand we have this much compute because you're forced to expand to that
because you're forced to expand to that and you're working against the tide to
and you're working against the tide to get again this is no dis like there's no
get again this is no dis like there's no sub owner I respect more than John
sub owner I respect more than John Durban yeah literally no one no one no
Durban yeah literally no one no one no one But on that it's it's more
one But on that it's it's more philosophy than practical realities.
philosophy than practical realities. Literally there's you cannot tell me
Literally there's you cannot tell me that I should pay a GPU's cost an H100
that I should pay a GPU's cost an H100 cost um cost um it's not even how much
cost um cost um it's not even how much it cost is how much demand do you have
it cost is how much demand do you have if you cannot fill your demand. And we
if you cannot fill your demand. And we we see this like you know shoot offered
we see this like you know shoot offered so many models that I'm not even sure
so many models that I'm not even sure how much traction they got but they did
how much traction they got but they did fill up their GPUs. So again, I'm I'm
fill up their GPUs. So again, I'm I'm not attacking shoots. Shoots is an
not attacking shoots. Shoots is an awesome is one of Bit Tensor's greatest
awesome is one of Bit Tensor's greatest successes. But if I talk about this
successes. But if I talk about this question of minor burns,
question of minor burns, it you cannot tell me that um you know,
it you cannot tell me that um you know, okay, because they're paying their
okay, because they're paying their miners 41%, it means their subnet is no.
miners 41%, it means their subnet is no. Look at the outcome. How is this
Look at the outcome. How is this translating to revenue? What is the
translating to revenue? What is the efficiency? Because honestly, I believe
efficiency? Because honestly, I believe if cheats was burning or at least
if cheats was burning or at least reducing how much I mean I think I think
reducing how much I mean I think I think I I spoke to John before he had he had
I I spoke to John before he had he had this um option of deferring the minor
this um option of deferring the minor payments which I thought was was good
payments which I thought was was good was good. So it's kind of like you know
was good. So it's kind of like you know they earn part of it part of his lock
they earn part of it part of his lock contract. Solutions like that are
contract. Solutions like that are absolutely necessary because I know even
absolutely necessary because I know even the most hardcore um the the most
the most hardcore um the the most hardcore um shoots u miners when things
hardcore um shoots u miners when things went sour after the initial after con
went sour after the initial after con after the first con stimulus died they
after the first con stimulus died they dumped they dumped they literally dumped
dumped they dumped they literally dumped so and this is not against them it's
so and this is not against them it's just the reality you know um I don't
just the reality you know um I don't know how like some of the template
know how like some of the template miners I don't know how they did it but
miners I don't know how they did it but they just held but if was out of I would
they just held but if was out of I would have dumped too. So while we have this
have dumped too. So while we have this thing that okay dude the investor is
thing that okay dude the investor is also so important this system we could
also so important this system we could get away with it on root network but the
get away with it on root network but the investor is such an important because we
investor is such an important because we exist at their privilege if they pull
exist at their privilege if they pull out all the money we're just there
out all the money we're just there playing simulation games so we need to
playing simulation games so we need to find and my argument is not and I do
find and my argument is not and I do agree with you fish one of the problems
agree with you fish one of the problems that is killing bit tensor is feticizing
that is killing bit tensor is feticizing of burn and not and changing it from my
of burn and not and changing it from my um changing it from a resource
um changing it from a resource optimization problem to a cost reduction
optimization problem to a cost reduction because when it becomes a cost reduction
because when it becomes a cost reduction then you're building you're building
then you're building you're building thin protocols fat companies.
thin protocols fat companies. Now the protocol is meant to be you're
Now the protocol is meant to be you're meant to enshrine all the logic in your
meant to enshrine all the logic in your protocol. The subnet is meant to be the
protocol. The subnet is meant to be the business. It's meant to be what creates
business. It's meant to be what creates creates innovation. when you strip out
creates innovation. when you strip out all the miners or because you know you
all the miners or because you know you don't want to do anything in the you
don't want to do anything in the you know you want to manage burn you want to
know you want to manage burn you want to do that then that is where detail
do that then that is where detail becomes quite toxic and also I I don't
becomes quite toxic and also I I don't know I hope maybe it changes but d also
know I hope maybe it changes but d also um it rewards stasis or it used to
um it rewards stasis or it used to reward stasis but I think constants
reward stasis but I think constants messed them up with talflow but the the
messed them up with talflow but the the the initial thing was just you know
the initial thing was just you know let's um let's do nothing test market
let's um let's do nothing test market you know investors are happy they check
you know investors are happy they check oh owner not selling oh you know minor
oh owner not selling oh you know minor bonds not on and that becomes the
bonds not on and that becomes the narrative that is perpetuated but bit
narrative that is perpetuated but bit tensor so we end up building web two
tensor so we end up building web two businesses so we're like okay we don't
businesses so we're like okay we don't spend any blah blah blah but why are you
spend any blah blah blah but why are you on bit tensor if you're not leaning into
on bit tensor if you're not leaning into your incentive mechanism so you have a
your incentive mechanism so you have a lot of subnets that literally don't even
lot of subnets that literally don't even plan on using miners or just turn on the
plan on using miners or just turn on the miners very small that is not the
miners very small that is not the direction we want
direction we want Rapido thoughts.
Rapido thoughts. >> Thanks a lot. Hello everybody. So there
>> Thanks a lot. Hello everybody. So there is many things to say. Um there are some
is many things to say. Um there are some things that uh that distributed is right
things that uh that distributed is right but there are a lot of things that I'm
but there are a lot of things that I'm I'm fully not agreed and I think every
I'm fully not agreed and I think every subnet owners that are currently burning
subnet owners that are currently burning they just forgot where they where where
they just forgot where they where where they they started. They all started as
they they started. They all started as minor as minors. Uh sorry.
minor as minors. Uh sorry. >> So my point is
>> So my point is >> so
>> so give me a second. Give me a second. Give
give me a second. Give me a second. Give me a second. Sam. [laughter]
me a second. Sam. [laughter] Um so I I think you talked earlier about
Um so I I think you talked earlier about investors. So I I can't talk on on
investors. So I I can't talk on on John's behalf of course, but I think on
John's behalf of course, but I think on shoot he's trying to build some
shoot he's trying to build some investment around the miners. Um the
investment around the miners. Um the point is to have some loyal miners that
point is to have some loyal miners that won't leave the network in case the
won't leave the network in case the miners are suffering because in the past
miners are suffering because in the past yes we we we were suffering suffering as
yes we we we were suffering suffering as minors. I was uh in in in a huge
minors. I was uh in in in a huge negative at some point myself. So you
negative at some point myself. So you know what he's trying to to to do um
know what he's trying to to to do um John is that
John is that okay some days there are good days some
okay some days there are good days some some days there are bad days because Tao
some days there are bad days because Tao is is up or to is down and and the point
is is up or to is down and and the point is we don't want shoot to be broken just
is we don't want shoot to be broken just because uh TOAO is down. You you don't
because uh TOAO is down. You you don't want every miners to to unplug their
want every miners to to unplug their machines. you want a reli reliable
machines. you want a reli reliable product and how do you build this is by
product and how do you build this is by by con building some liable miners let's
by con building some liable miners let's say and also there there there is
say and also there there there is another aspect that I'm I'm I'm fully
another aspect that I'm I'm I'm fully convinced is the fact that bit tensor is
convinced is the fact that bit tensor is designed to bring the most brilliant
designed to bring the most brilliant people in the world it can be anybody it
people in the world it can be anybody it can be any random people that are
can be any random people that are insanely brilliant. Look at fish for
insanely brilliant. Look at fish for instance. Look at look at you. So why
instance. Look at look at you. So why did you came into Bit Tensor? Because it
did you came into Bit Tensor? Because it it's designed to incentivize people to
it's designed to incentivize people to produce some
produce some commodity, some intelligence and and to
commodity, some intelligence and and to to just surpass yourself and the point
to just surpass yourself and the point to to to burn 100%.
to to to burn 100%. You're just making no chance to have
You're just making no chance to have more people brilliant like you. And that
more people brilliant like you. And that is a shame for me. Well, that's my take
is a shame for me. Well, that's my take answer.
answer. >> So that is that is interesting because I
>> So that is that is interesting because I was never a minor. I was I was never a
was never a minor. I was I was never a minor. I was never a minor. I just
minor. I was never a minor. I just couldn't do it. Like it's it's hard.
couldn't do it. Like it's it's hard. It's hard. My point is not like look at
It's hard. My point is not like look at like we can see it when what happens
like we can see it when what happens when the price of tile goes down. People
when the price of tile goes down. People shut off their machines. I know miners I
shut off their machines. I know miners I know choose miners that owe suppliers
know choose miners that owe suppliers millions. You understand? Like it's not.
millions. You understand? Like it's not. So I like John Dobin is amazing. He has
So I like John Dobin is amazing. He has a loyal following. But you cannot tell
a loyal following. But you cannot tell me ches miners are you are either
me ches miners are you are either running it because you are in a you're
running it because you are in a you're in a contract you can't get out of or
in a contract you can't get out of or you know running at a loss. It is not by
you know running at a loss. It is not by especially computer networks are the
especially computer networks are the worst for this. It is not by loyalty. We
worst for this. It is not by loyalty. We see it sometimes when like should
see it sometimes when like should service degrades miners will literally
service degrades miners will literally switch off if the tow price drops. Now,
switch off if the tow price drops. Now, there's a few people that are maxis like
there's a few people that are maxis like everyone has maxis that do the do the do
everyone has maxis that do the do the do the um do the um do the abnormal but you
the um do the um do the abnormal but you need to balance it with and we you so
need to balance it with and we you so again it if John believes this is how he
again it if John believes this is how he wants to do it he should be allowed to
wants to do it he should be allowed to you understand but I strongly believe
you understand but I strongly believe it's a resource optimization problem and
it's a resource optimization problem and there's no law there's no rule that says
there's no law there's no rule that says oh because um I dude If we did 100% burn
oh because um I dude If we did 100% burn 100 zero burn on um on um on Templar
100 zero burn on um on um on Templar even though the 72 billion parameter
even though the 72 billion parameter parameter run we would have gone to
parameter run we would have gone to zero. We would have literally gone to
zero. We would have literally gone to zero. It's just what it is. But does it
zero. It's just what it is. But does it mean our work is not valuable? Our work
mean our work is not valuable? Our work has put bits on the map. So what I'm
has put bits on the map. So what I'm saying is that okay I like we do have a
saying is that okay I like we do have a problem with burns but this law that
problem with burns but this law that everyone has to not burn 100% I think is
everyone has to not burn 100% I think is nonsense. And take for example, what if
nonsense. And take for example, what if I'm training an 8 billion parameter
I'm training an 8 billion parameter model next week
model next week um to lead up to something bigger? Do I
um to lead up to something bigger? Do I need to turn on 100% even though I even
need to turn on 100% even though I even though 10% is more is
though 10% is more is >> the question is not distributed whether
>> the question is not distributed whether or not you as a subnet owner should
or not you as a subnet owner should burn. The question is whether or not Bit
burn. The question is whether or not Bit Tensor should incentivize subnets that
Tensor should incentivize subnets that are burning. And those are not the same
are burning. And those are not the same question, right? You can see how those
question, right? You can see how those are not the same question. It's very
are not the same question. It's very clear. I understand why you want to
clear. I understand why you want to burn. I understand why fish wants to
burn. I understand why fish wants to burn. I want I understand what shoes
burn. I want I understand what shoes would probably burn. baby. Those are not
would probably burn. baby. Those are not the same question as where should Tao
the same question as where should Tao put its incentive towards sub towards
put its incentive towards sub towards people that are making things and using
people that are making things and using the the emission that they are are
the the emission that they are are producing value that is and they can
producing value that is and they can produce that that value in a way that
produce that that value in a way that they know that even though they're
they know that even though they're spending all that money that they're
spending all that money that they're producing inflow into their subnets,
producing inflow into their subnets, right, which would be a subnet that's
right, which would be a subnet that's being successful versus one that is
being successful versus one that is gaming in some sense pretending to be
gaming in some sense pretending to be productive through inflow but they're
productive through inflow but they're not actually producing any outflow. It's
not actually producing any outflow. It's a company that's just consuming emission
a company that's just consuming emission but not producing anything and the chain
but not producing anything and the chain itself can't tell. So like those are
itself can't tell. So like those are very different questions. I understand
very different questions. I understand why you don't want to you want to burn.
why you don't want to you want to burn. I understand why that is and and this is
I understand why that is and and this is so I'm asking you to take yourself out
so I'm asking you to take yourself out of your personal position and understand
of your personal position and understand the global abstract position of tow
the global abstract position of tow itself. Does that make sense? So, so I
itself. Does that make sense? So, so I think we spend a lot of time trying to
think we spend a lot of time trying to prevent instead of accelerating.
prevent instead of accelerating. You understand? Sub owners should be
You understand? Sub owners should be free to do what they want and um
free to do what they want and um eventually the market weeds them out.
eventually the market weeds them out. Toflow has been an amazing mechanism at
Toflow has been an amazing mechanism at doing this. You understand? TFL I think
doing this. You understand? TFL I think tal is amazing because you know you can
tal is amazing because you know you can pump all you can pump fake you can't
pump all you can pump fake you can't pump fake all day. I think last week we
pump fake all day. I think last week we had 50 submits that had no emissions.
had 50 submits that had no emissions. You said let the outcomes determine
You said let the outcomes determine where the emissions flow. Not our not
where the emissions flow. Not our not our philosophies are not oh we feel we
our philosophies are not oh we feel we feel this what what is that subet what
feel this what what is that subet what is that subet going to produce because
is that subet going to produce because it's almost disconnected from the
it's almost disconnected from the emissions.
emissions. The emissions are a tool for the sub
The emissions are a tool for the sub owner to to to I mean [laughter] I mean
owner to to to I mean [laughter] I mean we had some sub owners send emissions to
we had some sub owners send emissions to the um exchange. One of fish's
the um exchange. One of fish's colleagues sent emissions to the
colleagues sent emissions to the exchange but you know that's far and far
exchange but you know that's far and far between. Let us manage and you know if
between. Let us manage and you know if we're doing okay the market will react
we're doing okay the market will react to it like you can't like honestly
to it like you can't like honestly detail is very young and you can't do
detail is very young and you can't do this stuff forever. You can't fake it
this stuff forever. You can't fake it forever. You understand? So I that's why
forever. You understand? So I that's why I think as a as a tow holder you want
I think as a as a tow holder you want subnets to generate value. You do not
subnets to generate value. You do not care about that minor emission. You only
care about that minor emission. You only start to care about it. Um you you only
start to care about it. Um you you only start to care about it when there's no
start to care about it when there's no value. So let owners do like literally
value. So let owners do like literally as as they see fit. If they're not
as as they see fit. If they're not generating value, then it doesn't matter
generating value, then it doesn't matter if it's 100% emission, doesn't matter if
if it's 100% emission, doesn't matter if it's 5% emission, they get deregistered
it's 5% emission, they get deregistered or the market, you know, the the market
or the market, you know, the the market staff removes emission out of them. I
staff removes emission out of them. I think I think that works pretty well.
think I think that works pretty well. But saying that, oh no, you're not. So,
But saying that, oh no, you're not. So, dude, like you would
dude, like you would just
just you just you you would become you know
you just you you would become you know for a fact that if you do this,
for a fact that if you do this, >> but the the um the black box you open is
>> but the the um the black box you open is not um sorry the the Ptoras um Pandora's
not um sorry the the Ptoras um Pandora's box you open with this it's
box you open with this it's multi-dimensional problem because why
multi-dimensional problem because why did you why did you not do this? Because
did you why did you not do this? Because people like fish, fish has never used
people like fish, fish has never used the burn ID. He has never used the burn
the burn ID. He has never used the burn ID. You understand? We will all do that
ID. You understand? We will all do that because again about incentives and we
because again about incentives and we will go where the incentives goes. If
will go where the incentives goes. If it's possible that you can enforce 100%
it's possible that you can enforce 100% burn on 100% no minor bonds with
burn on 100% no minor bonds with everyone then okay fine because it
everyone then okay fine because it doesn't affect me. Everyone else is
doesn't affect me. Everyone else is burning the same. If it's possible that
burning the same. If it's possible that you know we've but again someone will
you know we've but again someone will figure out that minimum that games the
figure out that minimum that games the system. So rather than focus on that
system. So rather than focus on that let's focus on outcomes. Sometimes I
let's focus on outcomes. Sometimes I deliver it doesn't matter like it
deliver it doesn't matter like it doesn't m at the end of the day it
doesn't m at the end of the day it doesn't matter what your burn is. Show
doesn't matter what your burn is. Show me your outcomes.
>> I would like two points. Um so personally I have a lot of issues to
personally I have a lot of issues to sell and to to chill to when it comes to
sell and to to chill to when it comes to um okay what does this subnet most
um okay what does this subnet most especially subnets to chilling subnets
especially subnets to chilling subnets when they don't produce anything or I
when they don't produce anything or I mean not producing by something like
mean not producing by something like your subnets is I mean your subnet is
your subnets is I mean your subnet is easy to to chill because uh all of the
easy to to chill because uh all of the recent um uh let's say news and
recent um uh let's say news and achievements that you made are u insane.
achievements that you made are u insane. So congrats first of all but otherwise
So congrats first of all but otherwise in general I have a lot of issues to to
in general I have a lot of issues to to to chill the subnets uh to random people
to chill the subnets uh to random people like oh bit beat bit tensor is amazing
like oh bit beat bit tensor is amazing you you just can't believe um the the
you you just can't believe um the the whole behind the scene if I can't say
whole behind the scene if I can't say okay the miners are are producing this
okay the miners are are producing this because they are not paid so I I cannot
because they are not paid so I I cannot cheat something that doesn't exist. So
cheat something that doesn't exist. So this is this is one of the biggest point
this is this is one of the biggest point and um also my my second point is
and um also my my second point is I I really doubt the fact that so many
I I really doubt the fact that so many subnet owners would be burning if they
subnet owners would be burning if they would have also a burn cut let's say
would have also a burn cut let's say okay I know there is thousand way as you
okay I know there is thousand way as you say as you mentioned there is thousand
say as you mentioned there is thousand ways um to let's say um burn in
ways um to let's say um burn in unofficially
unofficially with with your odd keys and stuff, but I
with with your odd keys and stuff, but I think the transparency mattered the most
think the transparency mattered the most and um and I really doubt owners would
and um and I really doubt owners would would burn so much um if the
would burn so much um if the >> dude fish has never even to today fish
>> dude fish has never even to today fish doesn't use the burn ID like [laughter]
doesn't use the burn ID like [laughter] like what is like honestly like I'm very
like what is like honestly like I'm very confused by fish like honestly fish does
confused by fish like honestly fish does has never used the burn ID UID.
has never used the burn ID UID. [laughter]
[laughter] >> Nobody trusts fish anyway.
>> Nobody trusts fish anyway. >> No, no, no. But this the point. This the
>> No, no, no. But this the point. This the point. You don't have to trust fish. His
point. You don't have to trust fish. His investors loved him. They still do. You
investors loved him. They still do. You understand? Fish is still a good sub.
understand? Fish is still a good sub. Like it's sium is still the top subnets.
Like it's sium is still the top subnets. You understand? And selium at one point
You understand? And selium at one point used boasts a crazy and it still boasts
used boasts a crazy and it still boasts a crazy amount of GPUs. So, do we say oh
a crazy amount of GPUs. So, do we say oh because
because burn or I can't I can't even tell you
burn or I can't I can't even tell you the percentage of burn they have that no
the percentage of burn they have that no we must cut the emissions. It doesn't it
we must cut the emissions. It doesn't it could make some kind of sense but you do
could make some kind of sense but you do not measure measuring a subnet output by
not measure measuring a subnet output by um by minor by the amount of minor burn
um by minor by the amount of minor burn you have. It's it's killing it's killing
you have. It's it's killing it's killing it's it's splitting the baby in half. is
it's it's splitting the baby in half. is splitting the baby in half because you
splitting the baby in half because you you say okay they're not paying miners
you say okay they're not paying miners enough. How do you know what the
enough. How do you know what the subnet's task is based off some
subnet's task is based off some arbitrary numbers? But then we can go to
arbitrary numbers? But then we can go to the other side which I think is also
the other side which I think is also destruction where you know we have eal
destruction where you know we have eal then you have subnet owners set the um
then you have subnet owners set the um the percentages themselves which is a
the percentages themselves which is a solution that fish has adver has um has
solution that fish has adver has um has advertised for or has pushed. What do
advertised for or has pushed. What do you think is going to happen then when
you think is going to happen then when we when we let EAL happen? Subet owners
we when we let EAL happen? Subet owners are going to give validators a minus
are going to give validators a minus zero literally. So we can we we we can
zero literally. So we can we we we can go through permutations of this or we
go through permutations of this or we can just say dude what's your proof of
can just say dude what's your proof of work? What is your outcome? What value
work? What is your outcome? What value are you bringing back to the system?
are you bringing back to the system? Because yes I have minor bonds but I
Because yes I have minor bonds but I brought immense value to the system. I
brought immense value to the system. I so what you wanted to reduce my
so what you wanted to reduce my emissions for that because I'm not
emissions for that because I'm not paying minus 100%. It makes no sense. I
paying minus 100%. It makes no sense. I I have provided like we
I have provided like we >> Wait, wait, wait, wait. Are you saying
>> Wait, wait, wait, wait. Are you saying are are you saying that you're you're
are are you saying that you're you're accusing fish of trying to attack your
accusing fish of trying to attack your subnet on the
subnet on the >> What? [snorts]
>> What? [snorts] No, no, no, no, no, no. That's not what
No, no, no, no, no, no. That's not what I'm saying. No, no. Of course, Fish is
I'm saying. No, no. Of course, Fish is we we Leah mentioned Yeah. No, no. Yeah,
we we Leah mentioned Yeah. No, no. Yeah, exactly. No, I'm not saying that. I'm
exactly. No, I'm not saying that. I'm not saying that. I'm saying judge on
not saying that. I'm saying judge on outcomes. Judge on outcomes. This Oh,
outcomes. Judge on outcomes. This Oh, miners must be paid 100% is to me it's
miners must be paid 100% is to me it's nonsense. It's it's absolute nonsense.
nonsense. It's it's absolute nonsense. Like literally, I love miners, but
Like literally, I love miners, but honestly,
honestly, different stages of subnets require
different stages of subnets require different things. On Templar, we looked
different things. On Templar, we looked at the emissions. We looked at Okay, we
at the emissions. We looked at Okay, we looked at a couple of things. Okay, how
looked at a couple of things. Okay, how much does it take? [laughter]
Fish is sort of used as like um the personification of somebody that will
personification of somebody that will exploit the incentives if they're not
exploit the incentives if they're not correctly designed. Um which is true.
correctly designed. Um which is true. It's no longer fish by the way. It's
It's no longer fish by the way. It's it's Lazarus Group from Korea. Um which
it's Lazarus Group from Korea. Um which are much more nefarious. Um but uh and
are much more nefarious. Um but uh and and very smart. Um the so but but like
and very smart. Um the so but but like the the the if there was a way let's say
the the the if there was a way let's say to measure and then the point your point
to measure and then the point your point just taken distributed that that and if
just taken distributed that that and if the mechanism can't reliably do it then
the mechanism can't reliably do it then it's there's no point in having it and
it's there's no point in having it and this is sort of where we've been for the
this is sort of where we've been for the last little while when it comes to minor
last little while when it comes to minor burns. It's simply impossible to detect.
burns. It's simply impossible to detect. Um the first of it to burn minor
Um the first of it to burn minor emissions was fishes. Uh and he didn't
emissions was fishes. Uh and he didn't even use the burner burn minor key. And
even use the burner burn minor key. And the there's actually um the sub owners
the there's actually um the sub owners that control what is run and the
that control what is run and the validators can't reliably stop them from
validators can't reliably stop them from running whatever they want. Uh you know
running whatever they want. Uh you know perfect example is is 28. I think 28's a
perfect example is is 28. I think 28's a better example like they just send all
better example like they just send all their the current funds to their own
their the current funds to their own key. Um and uh and and this not the burn
key. Um and uh and and this not the burn key. So it's funny they actually have
key. So it's funny they actually have zero burn right now, right? So it's not
zero burn right now, right? So it's not a reliable it's not a reliable um
a reliable it's not a reliable um metric. But if if it were a reliable
metric. But if if it were a reliable metric, you you you're saying that
metric, you you you're saying that actually it doesn't you're not
actually it doesn't you're not necessarily opposed to it because it
necessarily opposed to it because it brings it makes it a fair playing field
brings it makes it a fair playing field for everyone, right? Like let's say that
for everyone, right? Like let's say that we did introduce um an adaptation to the
we did introduce um an adaptation to the mechanism inside bit tensor which which
mechanism inside bit tensor which which scaled the emissions by the amount
scaled the emissions by the amount that's burned as or up to a point maybe
that's burned as or up to a point maybe a sigmoid maybe an asmtope maybe it's
a sigmoid maybe an asmtope maybe it's allowed in somewhat but not you know
allowed in somewhat but not you know maybe it's annealed in all these things
maybe it's annealed in all these things to but but if it was even playing field
to but but if it was even playing field then it's not an issue so much
then it's not an issue so much >> from your perspective
>> from your perspective >> I think it's a waste of my time honestly
>> I think it's a waste of my time honestly because you
because you Let people build bit. Let let people
Let people build bit. Let let people build. You understand? Let's stop
build. You understand? Let's stop playing like that. We we spend a lot of
playing like that. We we spend a lot of time playing bit tester.
time playing bit tester. And if we spend a too much time playing
And if we spend a too much time playing bit test and we'll get left behind. It's
bit test and we'll get left behind. It's very it's very um you know we change a
very it's very um you know we change a lot everything you know you wake up
lot everything you know you wake up today it's tow flow you wake up today
today it's tow flow you wake up today it's tomorrow it's fine but if a system
it's tomorrow it's fine but if a system works we we can produce the good we can
works we we can produce the good we can channel what is it it is and you're
channel what is it it is and you're trying to bring this well do it but for
trying to bring this well do it but for me primary is a waste of my time because
me primary is a waste of my time because it's just going to be another experiment
it's just going to be another experiment we're doing we don't know if it works or
we're doing we don't know if it works or not we don't know the next exploit I
not we don't know the next exploit I mean if it moves the the ecosystem
mean if it moves the the ecosystem forward which I don't think yeah it will
forward which I don't think yeah it will then go for it but you know bit reduce
then go for it but you know bit reduce the barrier for people entering bit
the barrier for people entering bit tensor and yeah I don't know I don't
tensor and yeah I don't know I don't know it's because again you know I'm of
know it's because again you know I'm of the I'm of the I I'm of the opinion that
the I'm of the I I'm of the opinion that let owners optimize their resources and
let owners optimize their resources and look at outcomes but if you want to
look at outcomes but if you want to police things at that level well you
police things at that level well you know let's let's do it but it's just
know let's let's do it but it's just yeah
yeah >> it's not been that it's been interesting
>> it's not been that it's been interesting scintillating conversation
scintillating conversation um that we end ended in this way, but it
um that we end ended in this way, but it was mostly we were talking about a lot
was mostly we were talking about a lot of different things. Paul, Templar,
of different things. Paul, Templar, guys, we're at an hour and a half. Let's
guys, we're at an hour and a half. Let's let's let's let people go. Um and uh and
let's let's let people go. Um and uh and we'll bring this one back up. It's it's
we'll bring this one back up. It's it's it's on the tip of our of our tongues
it's on the tip of our of our tongues and our minds and we can think about it.
and our minds and we can think about it. Um it's it's worthy of bringing up.
Um it's it's worthy of bringing up. Rapido, thank thank you. Thank you for
Rapido, thank thank you. Thank you for having these these infographics and
having these these infographics and bringing them into the chat. Um,
bringing them into the chat. Um, everyone, uh, you thanks for coming to
everyone, uh, you thanks for coming to the stage, Alec and, um, Uran and and
the stage, Alec and, um, Uran and and and Samuel and Evan. Um, guys, it was
and Samuel and Evan. Um, guys, it was wonderful to have you up here, Rapid
wonderful to have you up here, Rapid Doofish, thank you so much for jumping
Doofish, thank you so much for jumping on the call and giving your opinions.
on the call and giving your opinions. Um, everyone else for coming uh, as
Um, everyone else for coming uh, as well, being active in the chat. Have a
well, being active in the chat. Have a good night.
good night. >> Have a good night.
Cliquez sur n'importe quel texte ou horodatage pour accéder directement à ce moment de la vidéo
Partager :
La plupart des transcriptions sont prêtes en moins de 5 secondes
Copie en un clicPlus de 125 languesRechercher dans le contenuAller aux horodatages
Collez une URL YouTube
Entrez le lien de n'importe quelle vidéo YouTube pour obtenir la transcription complète
Formulaire d'extraction de transcription
La plupart des transcriptions sont prêtes en moins de 5 secondes
Installez notre extension Chrome
Obtenez les transcriptions instantanément sans quitter YouTube. Installez notre extension Chrome et accédez en un clic à la transcription de n'importe quelle vidéo directement depuis la page de lecture.
