/biz/ - Business & Finance

BRAINLET Chain!

>>8485583
-95%

https://leakedcrypto.com/deep-brain-chain-review/

When masternode?

Good to see DBC still getting attention. Holding 60k here bought at 12 cents back on January 1st. It's been a very hard hold, but I'm in it for the long haul.

>>8485691
Give this man a brain sir

>
DBC Marines are immunized against all dangers: one may call them scoundrels, pajeets, swindlers, chinkcoins, it all runs off them like water off a raincoat. But spread out FUD and you will be astonished at how quickly they recoil, how injured they are, how they suddenly shrink back: “we've been looed out.”

>>8485942
They're consistently putting out updates now - https://medium.com/@deepbrainchain_74263/deepbrain-chain-weekly-report-3-e5f96c218296

When moon

>>8485942
underrated

mickey mouse coin, already mooned in January, stay away or regret forever...

>>8486010
Its ramping up boys, HODL to your brainz because: AI farm, binance listing, backed by China, mickey mouse, backed by Trump, huobi competition, backed by India, partnerships, backed by Hawking, briefly before he died on his deathchair, his last words were "I must now enter into my final form, the Deep Brain Chain Neural Network". Then he dropped his bags and passed away peacefully. Hawking is now officially part of DBC.

Shawn You will take the lead as our PR Director
You, graduated from Fudan University Department of Chinese, is a media professional with previous experience as a journalist at Xinmin Evening News and Wen Hui Bao Economics. You was in charge of PR for NEO and Ontology, focusing on Chinese domestic and foreign marketing, community management, and PR for Ontology Network.

>Ex head of PR for NEO
???? WHAT

>>8485583
Holding some DBC I spent too much on and I ain't selling.

>>8486236
backed by India, backed by UK, backed by Holland, backed by Australia, backed by Germany

<how deep is your FUD brainlet?

900k holding strong.

Will be laughing at the nobrains when a major chinese partnership is announced and this goes over 1 dollar.

>>8486361
Holy shit, you mad man.

>>8486361

> he unironically thinks we're still in whitepaper gains era

>DBC
>TRON
>WTC
>VEN

What could all this asian coins have in common...hmm

>>8486403
trust me on this one. I sold 100k at 55 cents which basically covered my initial investment.

This shit will go over 1 dollar. I'm not in crypto for peanuts, I'm already rich so making 400-500k is not worth cashing out to me. I will exit crypto once I hit 3-4MM.

>or you can just FUD me with your green text you stupid poor faggot

>>8485583
Your argument for buying into a 200 million dollar valuation is that it will actually turn into a product someday?

I will let you in on a little secret here, people are not selling this because they do not believe in the team, they are selling it because the idea is completely unrealistic.

Lets not church it up here, this coin performs one function, it is a marketplace token to connect people with spare GPU cycles to people that want to purchase those cycles.

Now the problem with AI is you cannot distribute the load over a network, it is advantageous to run it all on the same machine (thus the beefy neuaral net builds amd and nvidia are pumping out), so that means people are going to want to rent the highest TFLOP machines in the DBC marketplace. Now no one with a 500K-1Million dollar AI server is going to not be using it for a serious purpose, it will most likely have company IP flowing through it and thus it will be an iompossible sell to C level folks as it poses a security risk to allow anons to execute code on your racks.

So right now people with powerful AI rigs that want to sell their unused cycles will just mine shit, as it will make negligibly the same amount and poses zero security risk.

You picking up what I am laying down here? Dump your bags, this one is dead.

>>8486823
Update your pasta. The market cap is 42 million

>>8486361
Nice.jpeg megabrainmore with 900k brainz. Whale confirmed.

>>8486823
Strong FUD baby , loo levels rising

>>8485583

I might unironically throw $500 into this just for giggles

>>8486904
This the bottom floor of the bottomest, when market goes x5 later this year, DBC will x20

>>8486823

You forgot to mention your PhD in machine learning.

>>8486904
Look , that ain't pasta. I honestly had no idea it had gotten this bad.

Consistently rejecting a sane argument and just labeling it as FUD is no way to go about debating the pros/cons of a coin.

With the information and plan for the system as it exists today this is my current assessment of what will happen. Could I be wrong? absolutely, but I have been managing racks of servers for most of my adult life back to the Pentium 4 days.

I can tell you from experience, that selling crypto to cover DC expenses is hard as is. One of our COLOs we use for transcoding has a base breaker price for the power and I was able to cut costs by mining monero in our off hours. The process to convince business development folks of this was a long uphill struggle.

I cannot imagine any real business with AI servers (we have a couple but they are used for HEVC encoding) allowing actual code execution regardless of the walled garden DBC puts these clients in.

So without the real players offering the servers, what is the point of a client using the network to rent some home users 1070GTX vs just spinning up one in AWS?? I can tell you the market for AI using small scale consumer class cards is non existent.

i kind of feel like this shit is destined to explode just because everything biz realy really hates/ignores always does historically

>>8487153
>it doesn't work because I work at an IT call center

Thank you for your sane argument.

>>8487470
No, again as simple as possible one sentence:

People will not sell their cycles on this market because of the security implications of running anon code on your racks when you can just mine and only powerful enterprise grade machines are in demand for a neural net.

>>8487615
>only powerful enterprise grade machines are in demand for a neural net.

You've literally described the problem that DBC solves. AI cloud computing works.

>security implications of running anon code

Cool, so kind of like the child abuse images on bitcoin? Let's all sell our bitcoin! I don't think you know how AI block chain functions. Literally read their website.

>>8486823
>>8487153
>>8487615

>low tier IT employee is giving his "qualified" assessment.
When was the last time somebody gave a shit about the technical workings of a project? 99% of people don't even know what you're talking about.

DBC have stated multiple times that their primary concern and focus is security. It could be bullshit but I doubt anyone would take their business seriously if that wasn't the first thing they addressed. Nobody asked for your copy pasta opinion piece and nor does it make a difference, until we see some kind of usable product.

They're attending NVIDIA GTC to pitch their shit to tech companies. I'll be more interested in their opinions than yours.

200M cap for this trash lmfao
this is why u are poor

>>8487837
40mil..

Just lol if you dont hold DBC right now

>>8487805
But it does not solve that problem, did you even read the whitepaper? It is simply a marketplace, there is no concept of sharing the load across different nodes.

>>8488109
>there is no concept of sharing the load across different nodes.

Uh, yes, and also they LITERALLY say that in the video on their front page.

I can't tell if you're just trolling me, the trolls on biz are out of control. DBC is a legitimate company, it's not some small shitty team (like lamden, I hate those fuckers). It's a company with a corporate building that is expanding.

>>8488165
I am not trolling you man, that is why they have different node classes for the compute platform, you purchase nodes based on your needs for the neural net.

They toss around concurrent and decentralized, but in the end your learning task is completed on a single cluster. Unless there is a big shift in how neural nets work or the low latency bandwidth available to node operators (we are talking in the 10s of gigabits) your bottleneck will be the bandwidth and the learning task will sit idle waiting for data from the other nodes that are in different geographic regions.

All current AI has small input, big data for learning tasks stored in ram and on low latency busses then small output for the learned task.

Mainet launch, masternodes, binance listing, backed by NEO, backed Satoshi, backed by Musk, food pics, partnered with Stephen Hawking, Skynet confirmed, helped Trump win '16 election, etc.

>>8488298
Yes, data will be limited to bandwidth speed, but that's inconsequential when there are 10s of millions of computing nodes.

Cloud computing works dude

>>8485583

go to bed dave

>>8486361
wew look at all t

>>8488351
Absolutely, but these devs are not going to crack the horizontal scaling issue for neural nets.

I did not emphasize how big the big part of that data is, checkout the specs for the Nvidia DGX-1, they are not tossing a half terabyte of ram and 128GB of video ram in there for fun. It is because that is the buffer you use for learning tasks.

You while you can technically break up that data and distribute it over the internet it would be a massive waste because of the ephemeral nature of the data. There are 100s of gigabytes of data flowing through a second during a learning task.

When you are referring to cloud computing you are talking about applications specifically made to be horizontally scaled, their back-end data models will be columnar or data streams that can be consumed and represented across many frontend application servers.

Same goes for horizontally scaling large data distribution , you have metadata servers linking back to a pool of storage servers to grab reed reed solomon distributed data.

Horizontally scaling a pool of hundreds of terabytes is not possible unless you could develop some kind of marker system that would be universal for nueral net computing to ensure there was not wasted cycles and collision.

>>8488533
To add to this, if they do make a way to horizonally scale a nueral net they do not need to do any of this shit. They could directly license that patent to Amazon or Google (buildings of devs working on this there) and it would make this business look like small potatoes.

It would be along the lines of licensing LTE or CDMA like Qualcomm does, buko dollars.

>>8486823
>>8487153
>>8487615
>>8488109
>>8488298
>>8488533
>>8488600
btfo

>>8488533
>>8488600
>Absolutely, but these devs are not going to crack the horizontal scaling issue for neural nets.

I appreciate your write ups but this isn't what they're doing, and they don't need to. They're simply setting up a market place for AI cloud computing via blockchain which makes node interactions safe and anonymous. There is a huge demand for that from smaller companies with AI computing needs who don't want to invest millions in hardware specific to AI learning.

>>8488728
Yeah I literally said that:

>But it does not solve that problem, did you even read the whitepaper? It is simply a marketplace, there is no concept of sharing the load across different nodes.

Just responding to a youtube trained brainlet.

>>8488754
yeah, but that's not what I responded to. I never said that DBC was trying to solve horizontal scaling

>DBC will fail because they cannot solve the riemann hypothesis

>>8488824
Well then having 100s of millions of nodes does not matter, your task will take as long as it would to just run on a local desktop class GPU as that will be the majority of the pool.

The previous statement was to emphasize why serious AI clusters would never participate in this, they can just mine with their spare cycles and never have any kind of security risk to their local more valuable than servers IP.

*moons violently*

this and ICX are my biggest hopes for 2018

https://www.kucoin.com/#/trade.pro/DBC-ETH

The FIRE RISES brainmarines

>>8488947
no, the task will only take as long as the data can be uploaded from 100s of million of nodes, which is small quantities is instant because each calculation is tiny in itself, and it comes down to how much you're willing to pay to lease out x number of nodes. It's a game changer.

You do understand that these guys aren't a joke like most "dev teams" right? They all have tons of corporate experience and come from huge companies in china and have PHDs and MBAs, not that having degrees necessarily means anything, but it usually does. Their primary focus is B2B.

Please read up on this coin before you try and prove how smart you are.

>>8489431
backed by NEO council... these guys are Bezos and Musk at 22.

>>8489515
^.^ his brain grew hair practically, the power of DEE BEE CEE

>>8489431
Running distributed by design platforms on many commodity servers with GPUs like SETI or folding = great

Running learning sims across many commodity servers = not good

Look into Model parallelism vs Data parallelism, this should explain the bulk of it.

>>8488600
Your analysis that a marketplace will fail because people would just use Google or Amazon is flawed because it fails to account for geopolitics and hobbyist use.

> what is the point of a client using the network to rent some home users 1070GTX vs just spinning up one in AWS
Chinese want an AI solution that doesn't use western solutions.

> I can tell you the market for AI using small scale consumer class cards is non existent.
Arrogant.

increased my stack by over 3k today

>>8489983
How is that arrogant ? If you need to do machine learning renting some Slavs single GPU is not gonna revolutionize your operation.

Also The Chinese have Alibaba cloud among other services https://www.alibabacloud.com/product/gpu

It seems like you are just nitpicking a single point now because you are starting to understand the rest of the argument.

With all that said, with the amount of trash at these marketcap levels, will this likely pump and get you 5x? probably. That does not make it a sound platform though.

>>8489431
my company trains neural nets and other simpler ML models. we use clusters with GPUs. bottleneck is GPU time plus feeding training data in. not seeing how this could work well distributed without rewriting all of our code, which wouldn't be worth the cost savings.

>>8490059
Thanks just bought 100k