The truth about Deep Brain Chain

File: 326 KB, 765x511, DQmNnxYWkka8g754y3BFjybZAd6EvtdLmwZsdvQjdiLQPqz.png [View same] [iqdb] [saucenao] [google]

The truth about Deep Brain Chain Anonymous Mon Feb 19 12:54:02 2018 No.7793111 [Reply] [Original]

By way of background, I'm a postdoctoral research fellow at Harvard developing novel neural network architectures and training methods primarily for computational medical imaging.

Deep Brain Chain aims to distribute neural network training over computers worldwide. However, one the biggest problems in distributing training, whether over multiple GPUs or multiple computers, is communication bandwidth. Even when training is done on a single machine (let's call this Distribution Level 0), data needs to be transferred between the CPU to the GPUs to transfer training data and iterative gradient updates. If your model is reasonably big (for example, I regularly work with ~10GB models), that communication is happening after each training batch and is often the main bottleneck for training time. PCIe 3.0 x16 supports ~16 GB/s max transfer rates which means I'm waiting about one second each iteration JUST for data transfer... this transfer delay is one of the primary reasons why you might hear deep learning models taking days/weeks/months to train. This is such a big problem that NVIDIA has developed a specialized I/O bus called NVLINK (which will improve those speeds 5-10x but is only currently implemented on IBM Power systems).

Parallelizing this training over multiple computers (Distribution Level 1) typically occurs in a local cluster or a cloud system that is intraconnected with 10-100 Gbit/s ethernet (AWS is 25 Gbit, or about 3GB/s). Even though you're adding more computers, this 3-5x speed transfer loss, in addition to algorithmic losses due to imperfect parallelization and asynchronous updating etc (https://en.wikipedia.org/wiki/Amdahl%27s_law)), results in a poorly-scaled system, which is a widely-studied issue in the research community (e.g. https://arxiv.org/pdf/1609.06870.pdf).). You'll still get gains better than single-machine, but they won't be linear (10 machines doesn't equal 10x compute speed), and it'll be worse the slower your transfer bandwidth is.

Anonymous Mon Feb 19 12:55:16 2018 No.7793149

>>7793111

Going from Distribution Level 1 to Level 2 - distributing this to the masses a la Deep Brain Chain (or some similar system) would further slow down the training to nearly incomprehensible levels, in my estimation. Assuming each computer is equipped with 100 Mbit/s ethernet both upload and download (which is a VERY generous assumption: http://www.speedtest.net/reports/united-states/)), that's still > 100x slower than cluster transfer rates, and >1000x slower than a single machine implementation. Also, with an institutional cluster, the machine hardware (CPU, RAM, etc...) can be controlled to be relatively similar, which makes the training more predictable and thus parallelization more efficient. However, a large set of random computers across the internet would result in potentially wildly inhomogeneous computation speeds with much less predictable collective behavior, introducing further algorithmic slowdowns. In their whitepaper they do note that they're aiming to have some advanced method of load-balancing (which wasn't pointed to solve this problem, just utilizing idle nodes), so perhaps in the future they can limit some of these losses - but even then you're still stuck with the lower bandwidth speeds that would bring training to a crawl.

If you have any thoughts on this, or have seen any (big or small) mistakes here please let me know. For example, there may be training algorithms I'm unaware of that somehow sidestep this (although I'm quite confident that what I've been describing is true for the ubiquitous stochastic gradient descent and their variations, which is like what 99% of what people use for NN training). I don't want to spread misinformation, but I just can't see how they'll get around these big issues. I'm also disheartened that these issues weren't mentioned at all in the white paper, as these are very well-known limitations of distributed training.

Anonymous Mon Feb 19 12:56:00 2018 No.7793166

well one thing I was thinking is that there will be a lot of 'low level' AI out there thanks to IoT applications, so maybe some of these shitty AI tokens are good for companies who need to fill smaller gaps, rather than dedicated centers for AI advancement/research

Anonymous Mon Feb 19 12:56:17 2018 No.7793176

>>7793149

I hope the DBC team can be as transparent as Ethereum is about the reality of their scaling problems, and be working diligently to solve them. I like the vision of DBC, but real value relies on real execution.

Everything I've described above applies to distributed training, where training a NN model requires multiple computers updating each other on training iteration results. If the scope of their training/network model is small enough and doesn't need computers talking to each other then they might be okay. However, everything in their white paper points to large, enterprise-scale datasets and large complex models.

The CEO responded to my original post via telegram. https://www.reddit.com/r/cryptocurrency/comments/7od7vg/_/ds9ecpy?context=1000 If I understand him correctly, it seems like their distribution strategy is to parallelize the training of one model over different training hyperparameters. So that indeed can be done with current hardware (like he implied, it would only be limited to bigger rigs) and bandwidth limitations. However, it's a letdown because they're not making a true distributed NN training system - parallelizing hyperparameter search is a pretty narrow offering, and is quite far from the vision they seemed to promote in their marketing which is in their own words a "decentralized neural network." This wouldn't be able to provide larger models spanning multiple nodes, or accelerating training by splitting up datasets, which are the typical benefits of distributed training. In all fairness though, searching for optimal hyperparameters is indeed an important factor of training, so there is some value there... Just not as much as it seemed they were marketing.

>>	Anonymous Mon Feb 19 13:03:17 2018 No.7793328 Unfortunately, this is probably too high level discussions for the brainlets and pajeets in /biz/

>>	Anonymous Mon Feb 19 13:03:34 2018 No.7793337 lmaoing at your life OP

>>	Anonymous Mon Feb 19 13:05:16 2018 No.7793377 just tell me if I should buy or not

>>	Anonymous Mon Feb 19 13:07:27 2018 No.7793431 >>7793377 this, you can't expect people to read all this

>>	Anonymous Mon Feb 19 13:09:47 2018 No.7793504 Nobody is going to read even 10% of that. This isn't reddit, fag

>>	Anonymous Mon Feb 19 13:10:21 2018 No.7793523 >>7793149 I’m at masters student researching deep learning, always got a good kek out of watching retards buy this. Kudos for being nice enough to warn people.

>>	Anonymous Mon Feb 19 13:10:25 2018 No.7793526 >>7793431 Its incredibly scary how much wealth is being thrown around without "investors" having any clue what they're doing. Pulling off a scam last year would have been a piece of cake.

>>	Anonymous Mon Feb 19 13:12:54 2018 No.7793580 >>7793526 so you are telling me that people are actually reading something else apart from the name of the coin?

>>	Anonymous Mon Feb 19 13:16:16 2018 No.7793662 Stay away from AI blockchain bullshit.

>>	Anonymous Mon Feb 19 13:18:55 2018 No.7793726 >>7793111 Didn't buy into this shitcoin to begin with buit still an nteresting read, thanks OP.

>>	Anonymous Mon Feb 19 13:19:37 2018 No.7793748 >>7793580 lmao fucking nerds

>>	Anonymous Mon Feb 19 13:19:51 2018 No.7793749 Nice copy pasta

>>	Anonymous Mon Feb 19 13:21:03 2018 No.7793775 >>7793580 I also count chink/no chink ratio in the devs team

>>	Anonymous Mon Feb 19 13:22:30 2018 No.7793798 >>7793523 These 'AI' coins are for flipping not investing. Matrix AI Network is next. It has all the buzzwords.

>>	Anonymous Mon Feb 19 13:26:14 2018 No.7793882 >>7793377 I wouldn't touch anything that tries to utilize blockchain mining infrastructure with massively distributed high level computation/AI.

>>	Anonymous Mon Feb 19 13:27:16 2018 No.7793905 >>7793798 As long as my /bizbros/ know this. But I know some of them don't.

>>	Anonymous Mon Feb 19 13:27:33 2018 No.7793911 File: 291 KB, 888x499, brainz26.jpg [View same] [iqdb] [saucenao] [google] Stahp, it's already dead

DBC Mon Feb 19 13:38:16 2018 No.7794185 [DELETED]

File: 10 KB, 183x275, poorr.jpg [View same] [iqdb] [saucenao] [google]

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

this is the only coin going down while most other coins are doing ok or even better. DBC is a SCAM COIN

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

>>	Anonymous Mon Feb 19 13:43:23 2018 No.7794301 >>7794185 why does dropping coin price make it a scam coin and why is your life so pathetic that you're making posts like this one

>>	DBC Mon Feb 19 13:43:24 2018 No.7794302 File: 9 KB, 251x201, bbbuum.jpg [View same] [iqdb] [saucenao] [google] DBC

>>	DBC Mon Feb 19 13:44:28 2018 No.7794328 File: 12 KB, 191x264, bbb.jpg [View same] [iqdb] [saucenao] [google] FUCK U AND SCAM COIN DBC

>>	Anonymous Mon Feb 19 13:45:26 2018 No.7794353 >>7794185 >>7794302 >>7794328 Fuck out of my thread, pajeet

>>	Anonymous Mon Feb 19 13:45:28 2018 No.7794354 File: 197 KB, 828x245, 1516467976520.png [View same] [iqdb] [saucenao] [google]

>>	Anonymous Mon Feb 19 13:45:45 2018 No.7794363 >>7793111 lol what a fuckin nerd

Anonymous Mon Feb 19 13:45:54 2018 No.7794366
File: 195 KB, 765x427, brainz13.jpg [View same] [iqdb] [saucenao] [google]

>>7794185
I'm not ever selling my DBC. You know why? I want it to stand there as a warning in my blockfolio. Like the enemy spiked on a pike that is mounted on the walls, I want to look at it each and every morning.

I want to feel the punch in my gut when I look at my NEO stack breaking through 1000$, and this shit still hovers around 0.01$. I want to kick myself when my LINK stack blows past 100$ to see that this pathetic DBC stack of mine is still in there, flaining around as its slowly dying and withering away. It shall serve as a reminder never to FOMO into stuff you don't know enough about yet, and it should be a grim reminer not to be blinded by optimism and wishful thinking.

Anonymous Mon Feb 19 13:48:30 2018 No.7794424
File: 380 KB, 763x768, 1b2c419f03b7100abfd54c973a0ddc388cb45413d16816a9f112c38bc7b507d8.png [View same] [iqdb] [saucenao] [google]

I invested in DBC at around $0.10, and I shilled this motherfucking coin like crazy. I shilled it harder than anyone else on this board, even harder than that insufferable tripfag. I created multiple threads, used 4+ unique IPs to shill it day in and day out. I told everyone it would take off after the Hyoubi listing. I got out at about $0.53 and every day I love laughing at the people I tricked to buy at the ATH. I remember how I told them it will easily be over a dollar after getting listed on Hyoubi and $10 by Feb. It makes me feel good inside.

>>	Anonymous Mon Feb 19 13:49:00 2018 No.7794436 >>7794366 >LINK ohh...

>>	DBC Mon Feb 19 13:49:30 2018 No.7794451 File: 8 KB, 263x192, tras.jpg [View same] [iqdb] [saucenao] [google] DBC holders will be like this... SCAM COIN SCAM COIN SCAM COIN DBC DBC SCAM COIN DBC SCAMMMMMMMMMMM SCAMMMMMMMMMMMM SCAMMMMMMMMMM

>>	Anonymous Mon Feb 19 13:50:20 2018 No.7794471 >>7793523 Where are you studying for this master?

>>	Anonymous Mon Feb 19 13:50:56 2018 No.7794485 >>7794424 >created multiple threads created multiple thread daily*

DBC Mon Feb 19 13:51:11 2018 No.7794488 [DELETED]

File: 16 KB, 264x191, traaaa.jpg [View same] [iqdb] [saucenao] [google]

SCAM COIN
SCAM COIN
SCAM COIN
SCAM COIN
SCAM COIN
SCAM COin
SCAM COIN
SCAM COIN
SCAM COIN
SCAM COIN
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

>>	Anonymous Mon Feb 19 13:51:47 2018 No.7794502 imagine believing this... ahahahaha

>>	Anonymous Mon Feb 19 13:52:46 2018 No.7794520 If /biz/ is fudding this hard by posting month old reddit posts about DBC you know it's about to moon.

>>	Anonymous Mon Feb 19 13:53:23 2018 No.7794538 >>7794424 enjoy the karmic consequences faggot

>>	Anonymous Mon Feb 19 13:54:09 2018 No.7794556 >>7794424 Scamming people makes you feel good on the inside? No wonder the white race is failing. At least the 3rd worlders who scam have to do it because they don't have any opportunities.

>>	Anonymous Mon Feb 19 13:54:55 2018 No.7794573 what is this absurd level of FUD jesus fucking christ

DBC Mon Feb 19 13:54:58 2018 No.7794575 [DELETED]

File: 11 KB, 194x259, poor.jpg [View same] [iqdb] [saucenao] [google]

SCAM COIN
SCAM COIN
SCAM COIN
SCAM COIN
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

DBC Mon Feb 19 13:56:40 2018 No.7794624 [DELETED]

File: 10 KB, 275x183, 1.jpg [View same] [iqdb] [saucenao] [google]

im going to make sure this piece of shit goes to 0000

f you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

f you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

DBC Mon Feb 19 13:57:46 2018 No.7794641 [DELETED]

File: 12 KB, 270x187, 2.jpg [View same] [iqdb] [saucenao] [google]

f you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

f you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

>>	Anonymous Mon Feb 19 13:58:16 2018 No.7794659 File: 35 KB, 450x124, dbccopypasta.jpg [View same] [iqdb] [saucenao] [google] >>7793111

DBC Mon Feb 19 14:03:41 2018 No.7794796 [DELETED]

File: 8 KB, 275x183, 3.jpg [View same] [iqdb] [saucenao] [google]

f you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

f you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

DBC Mon Feb 19 14:09:42 2018 No.7794938 [DELETED]

File: 8 KB, 221x228, trapp.jpg [View same] [iqdb] [saucenao] [google]

I'm trying to save any one who is thinking about buying Deep Brain Chain DBC

DON"T BUY THIS SCAM COIN
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

f you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

DBC Mon Feb 19 14:19:25 2018 No.7795188 [DELETED]

File: 9 KB, 275x183, 5.jpg [View same] [iqdb] [saucenao] [google]

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

f you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC
if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

if you want to lose all your money buy DBC..

if you want to lose all your money buy DBC

>>	Anonymous Mon Feb 19 14:20:19 2018 No.7795217 If this is not a buy signal, I don't know what is

>>	Anonymous Mon Feb 19 14:25:20 2018 No.7795337 What about the possibility of creating an AI which hijacks all resources just for the sake of cryptocurrency?

Anonymous Mon Feb 19 14:28:51 2018 No.7795440
File: 118 KB, 666x547, storage-latency-how-far-away-is-the-data.png [View same] [iqdb] [saucenao] [google]

>>7795337
The problem is synchronizing the massively distributed resources. Its extremely inefficient.

If you've studied any computer architecture, you know the slowdown rates between

>>	Anonymous Mon Feb 19 14:30:00 2018 No.7795463 File: 58 KB, 960x720, slide_9.jpg [View same] [iqdb] [saucenao] [google] >>7795440 Another for the lols

>>	Anonymous Mon Feb 19 14:32:11 2018 No.7795523 File: 22 KB, 1024x512, latency.png [View same] [iqdb] [saucenao] [google] >>7795463 Last one, couldn't help myself

Anonymous Mon Feb 19 14:34:44 2018 No.7795572

>>7793111

This copypasta is fucking autistic. I don't know what the fuck he's talking about with data having to be transfered from CPU to GPU. You can use both independently. This project never claimed to reduce computational times, even though with enough users that might be the case. The aim was to increase accessibility and reduce costs of performing deep learning.

Also, there are a few universitie that already use distributed computing to perform Protein Folding. http://folding.stanford.edu I cant remember the other ones off the top of my head.

>>	Anonymous Mon Feb 19 14:34:46 2018 No.7795573 >>7793111 Layman's term Einstein, explain it as if we don't have formal background knowledge of computers. And just a simple summary, not everyone wants sift through long winded expositions like this.

>>	Anonymous Mon Feb 19 14:35:56 2018 No.7795605 >>7793111 >>7793149 >>7793176 I'm like 99% sure this is copy pasted from a reddit post a few weeks ago. Iirc that guy even said later that's it's better than he initially thought.

Anonymous Mon Feb 19 14:42:42 2018 No.7795765

>>7793149
In a distributed system, doesn't it make more sense to have multiple iterations running at the same time? So say you have 10 computers, you could have 5 batching one iteration and the other 5 batching a different one entirely. That way you could scale more effectively. You could chunk computers together by location as well.

>>	Anonymous Mon Feb 19 14:43:10 2018 No.7795774 >>7795605 this retarded post is probably single handedly the reason why so many people think this is "a scam"

>>	Anonymous Mon Feb 19 14:44:20 2018 No.7795801 Wow thanks js sold all my DBC at a fat loss. Thanks bizbros

Anonymous Mon Feb 19 14:49:05 2018 No.7795917

This is copy pasted from a Reddit post a little while ago. I remember reading it. Nevertheless, point stands - DBC is vapourware. Bachelor’s and Master’s in ML here and this is very clearly a scheme for taking advantage of those who don’t know anything about computer science but are drawn by how ‘hot’ terms like deep learning are now.

If you have any kind of background in ML, read the whitepaper. There’s absolutely no details of the kind of neural networks or algorithms they intend to facilitate. The ‘reward function’ is formulated using game theory, maybe just to trick those with some knowledge of economics into thinking it’s legit. The whole thing’s like a schematic of a car with a great big blank space where the engine should be.

>>	Anonymous Mon Feb 19 14:51:42 2018 No.7795989 >>7795917 See >>7795572

Anonymous Mon Feb 19 14:56:14 2018 No.7796101

>>7793328
I'm an EE and work with computer vision almost exclusively, and I love that someone is here laying down hard facts instead of LARPing. Your grammar, language, field-specific knowledge and breakdown into easily digestible sentences is very appreciated. Thanks for taking the time to post

>>	Anonymous Mon Feb 19 14:59:02 2018 No.7796179 >>7793111 Now do an analysis of Oyster Pearl. And give me a nice story of why it is a shitcoin

>>	Anonymous Mon Feb 19 15:00:44 2018 No.7796224 Didn't Neo council invest 1 mil into DBC. Not saying they weren't wrong but they don't seem like they would be backing chink scams

>>	Anonymous Mon Feb 19 15:03:24 2018 No.7796299 >>7796224 >>7795188 look at this poster, do you think he's really smart enough to realize any of this

Anonymous Mon Feb 19 15:05:46 2018 No.7796358

To everyone saying "copypasta," I published this on Reddit and made a few small updates to my Reddit edits.

To everyone who is bitching and moaning and spamming this thread. Please actually dispute something I've said if you want anyone to take you seriously.

>>7795989
They aren't going to reduce computational costs, they're just jumping on the blockchain hype train (which nobody understands) and combining it with ML (which nobody understands).

It is pointless to try and "decentralize" the training algorithm for neural nets that could be done more efficiently in a data-center.

>>7795572
>I don't know what the fuck he's talking about with data having to be transfered from CPU to GPU.

Just a simple example using the two most powerful computational/processing units in a server/computer. Just look at >>7795523
if you don't understand how hardware works.

You might enjoy skimming through this: http://www.ertl.jp/~shinpei/papers/icpads13.pdf

>>	Anonymous Mon Feb 19 15:07:59 2018 No.7796409 >>7796179 Sadly not enough time in the day to analyze every shitcoin. I'm the .01% that reads the whitepaper and does real research into the team. Its scary how much dumb money is in crypto right now.

>>	Anonymous Mon Feb 19 15:08:20 2018 No.7796415 >>7795440 So I shouldn't fear Skynet?

Anonymous Mon Feb 19 15:11:03 2018 No.7796489

>>7796358
Would you already fuck off and let us lose our money?
When are fags like you going to realize that NOTHING NEEDS a blockchain.
I dont care how smart you are but I guess you havent realized mankind has flourished all the way up to 2009 and will continue to do so with or without blockchain.

>>	Anonymous Mon Feb 19 15:12:28 2018 No.7796518 >>7796409 >has been in crypto >idiot

Anonymous Mon Feb 19 15:14:50 2018 No.7796566

>>7795989
The comment
>>7795572
is insane. How do you think computers work? Input/output is handled by the operating system. The operating system runs on the CPU. If instructions are to be sent to the GPU (or any piece of hardware), they by necessity have to come from the CPU.

Folding@home from Stanford works because people are donating their computing power. The inefficiency of distributed computing is unimportant because cost is not a factor. If people sold their computing power instead, it would be far more cost-efficient for Stanford just to buy more computing clusters.

It’s really fucking hard to admit when a coin you own is shit. I’ll admit to having invested in both XRP and TRX and going through the slow realisation that I was wrong. I would find every reason to dismiss the perfectly valid criticisms I heard. You’ll probably find a way to dismiss this comment, but I hope it serves as a deterrent to those considering investing.

>>	Anonymous Mon Feb 19 15:16:59 2018 No.7796614 >>7793111 Very interesting read, thanks OP

>>	Anonymous Mon Feb 19 15:18:08 2018 No.7796641 >>7793882 You heard it folks. Time to buy

>>	Anonymous Mon Feb 19 15:23:29 2018 No.7796762 >>7796641 >actual buying starts >you done pissed off the whales

>>	Anonymous Mon Feb 19 15:25:08 2018 No.7796802 >>7793111 http://www.eecs.harvard.edu/~htk/publication/2017-icdcs-teerapittayanon-mcdanel-kung.pdf And yet, here's a Harvard research paper on Distributed Deep Neural Networks over the Cloud, the Edge and End Devices

>>	Anonymous Mon Feb 19 15:29:48 2018 No.7796895 >>7793111 What is your opinion on Golem

>>	Anonymous Mon Feb 19 15:32:57 2018 No.7796974 >>7796802 Anyone can cherrypick a research paper to support their point. http://learningsys.org/nips17/assets/papers/paper_21.pdf

>>	Anonymous Mon Feb 19 15:40:12 2018 No.7797165 File: 68 KB, 537x390, advice.jpg [View same] [iqdb] [saucenao] [google] >>7796802

>>	Anonymous Mon Feb 19 15:40:27 2018 No.7797168 OP and other ML folks on this thread are right. Like most other "AI" coins, it's all an attempt to play buzzword bingo in order to get more $$$. Getting really fucking of this bullshit ridden world.

>>	Anonymous Mon Feb 19 15:42:44 2018 No.7797226 Hahahaha nice try anon, you're stuck with those bags for at LEAST another 6 months. Try again next time.

>>	Anonymous Mon Feb 19 15:48:30 2018 No.7797368 File: 19 KB, 600x478, facepalm.jpg [View same] [iqdb] [saucenao] [google] >>7797226 >anon thinks I have dbc bags

Anonymous Mon Feb 19 15:48:36 2018 No.7797373

>>7793149
Thanks for writing this out. I’ve been warning people about this shit coin for months but never put in the effort to write a post with all the detail. (I’m also in Boston but just a PhD student) The whole project is a mediocre idea that doesn’t solve a real problem and is being marketed to people who think they are “tech savvy” by throwing a bunch of exciting buzzwords together.

Also, go watch the video on their website because it’s hilariously bad Engrish.

Anonymous Mon Feb 19 16:06:07 2018 No.7797859
File: 28 KB, 600x600, 2017100314262792-14s.jpg [View same] [iqdb] [saucenao] [google]

Whats the reason for this gay ass fud you nigger op. Anything youre hiding or are you actually trying to pretend like you know how crypto works

Its going up. Like it or not. Its still low because it hasnt gotten up yet. As simple as that.
You can go blindy buy any coin that has 1% use and you will have gains month later on

>>	Anonymous Mon Feb 19 16:17:14 2018 No.7798145 >>7797368 Are there ANY AI memecoins that are worth anything?

Anonymous Mon Feb 19 16:18:29 2018 No.7798183

>>7797859
I could care less what this shitcoin does. It might go to the moon if you guys can keep shilling your bags. It might not.

I'm here to talk about the tech behind this coin's "business"... I realize most of you are here for quick flips and don't give a fuck if you're just handing around bags, as long as your "wealth" is growing and you pass your bag before it becomes heavy.

But in my opinion, bullshit blockchain money-grabs are doing a lot of damage. Not just for crypto's image, market cap, adoption, etc., but because some uneducated kid is eventually going to end up holding worthless coins.

Anonymous Mon Feb 19 16:22:10 2018 No.7798282

>>7798145
I'm not here to FUD, shill, or offer alternatives. I'm just here to educate.

For what its worth, there ARE some incredibly cool use cases for AI + blockchain. If you're interested, check out Siraj's video https://www.youtube.com/watch?v=5Tr13l0O1Ws

Anonymous Mon Feb 19 16:23:43 2018 No.7798329

>>7797165
It may not be necessary, true, but if it works, it will still go to the moon and beyond. You don't understand how this fucking world works if you think something has to be absolutely necessary in order to make million $s off of it. Fuck, most of the stuff we produce is useless shit.

This coin is something that has an actual usecase behind it, unlike 95% of other random forks of already established coins that are literally nothing more than currencies, and will wither and die away eventually.

>>	Anonymous Mon Feb 19 16:24:38 2018 No.7798351 >>7794424 Fuck you desu. I went in early too so I only lost $30 on this shitcoin when I sold yesterday, but you're still a nigger. We could've had a dream

>>	Anonymous Mon Feb 19 16:29:13 2018 No.7798468 Link your harvard profile or else youre a dumb larping fag.

>>	Anonymous Mon Feb 19 16:40:22 2018 No.7798768 >>7798183 You yourself I suppose have never sold shitcoins before? If you have your a hypocrite with zero cridentials.

Anonymous Mon Feb 19 16:42:39 2018 No.7798834

>>7798329
Of course, you are right. But let's not split hairs. Blockchain is a data structure that will be integrated into just about everything if the opportunity cost/trade-off is good enough. But the practice of creating a token for everything and overfunding ICOs with no working product is not going to continue forever.

>>	Anonymous Mon Feb 19 16:44:42 2018 No.7798897 >>7793111 >>7793149 PhD working in Neural Nets here. Agree firmly on all points. Never will I buy an AI token.

>>	Anonymous Mon Feb 19 17:03:50 2018 No.7799385 >>7798897 All the sudden 5-10% of 4chan has PhDs in advanced machine learning. Hmmmmmm.

>>	Anonymous Mon Feb 19 17:21:18 2018 No.7799865 >>7793111 Bump for interest.

>>	Anonymous Mon Feb 19 17:28:45 2018 No.7800076 >>7799865 What do you still want to know?

>>	Anonymous Mon Feb 19 17:49:03 2018 No.7800669 File: 8 KB, 497x187, Capture.png [View same] [iqdb] [saucenao] [google] >>7799865 Always do the opposite of what /biz/ says.

>>	Anonymous Mon Feb 19 17:57:28 2018 No.7800931 >>7799865 Bump for FUD? I thought we all decided to stop caring if these coins were shit or not

Anonymous Mon Feb 19 18:00:17 2018 No.7801020

I am not an AI fag but God damn OP, read what dbc is trying to do. It's pretty clearly explained. A decentralized neural network. Many small businesses do not have money to buy expensive GPU, and many companies have GPU sitting idle. So on the Blockchain these GPUs can provide computational requirements for smaller business. And these companies can also keep their GPUs. Sounds like they redistributing the AI computing requirements/machines to users specified needs and requirements.

Harvard faggot can't read. Smd.

>>7793111
The large-scale data parallel training problem, however, is not what the deep brain chain intends to solve. To say, it is not for the same user to do the remote distributed computing where the speed is reduced significantly. The first and foremost the deep brain chain tackles with is to reduce the high-performance machine computing power costs and data privacy. In other words, how to lower the cost of neural network training and neural network computing. When an AI product has a large number of users it needs a lot of high-performance machines to meet the computing requirements, deep brain chain can execute distributed computing based on the user's specific circumstances and it is for different users. The DBC project is for global AI computing resource sharing and resource scheduling because many small businesses do not have the money to buy expensive GPU servers but many companies have a large number of GPU servers which are idle. Scheduling global resources and increasing the utilization efficiency of resources are of the positive significance. We also consider optimizing the distributed AI calculation engine and working with the underlying neural network chip makers. However, our focus is still on reducing AI business computing costs, rather than enhance the AI model of any single business model training.

Anonymous Mon Feb 19 18:00:55 2018 No.7801037

>>7793111
OP, what are you doing on /biz.? You know 99% of these pathetic brainlets can't even comprehend what you wrote?

Most are just going to argue senselessly with you or create a straw man/red herring argument because its the best they can do.

You've absolutely knocked this one out of the park though. I'd say this is nearly Assblaster level material and execution.

Grab a Trip Code so next time we'll know to trust you.

>>	Anonymous Mon Feb 19 18:03:01 2018 No.7801116 >>7801037 More like the Harvard faggot can't comprehend what DBC is trying to do. I have no clue what the op faggot said, but my post is what the CEO mentioned in his past attempt to FUD this project.

Anonymous Mon Feb 19 18:04:15 2018 No.7801145

>>7801020
>Many small businesses do not have money to buy expensive GPU, and many companies have GPU sitting idle.

Holy shit this is starting to get comical. Please tell me more about the last time a small business decided to start utilizing advanced machine learning techniques

kek

>>	Anonymous Mon Feb 19 18:04:41 2018 No.7801155 >>7800076 >>7800669 >>7800931 I hold 60k DBC and find actual discussion about the token interesting. The Telegram channel is mindless optimism and the subreddit is even worse.

>>	Anonymous Mon Feb 19 18:05:29 2018 No.7801178 >>7801116 >is what the CEO mentioned in his past attempt to FUD this project. Why would the CEO try to FUD his own project?

Anonymous Mon Feb 19 18:07:04 2018 No.7801224

>>7793111
The large-scale data parallel training problem, however, is not what the deep brain chain intends to solve. To say, it is not for the same user to do the remote distributed computing where the speed is reduced significantly. The first and foremost the deep brain chain tackles with is to reduce the high-performance machine computing power costs and data privacy. In other words, how to lower the cost of neural network training and neural network computing. When an AI product has a large number of users it needs a lot of high-performance machines to meet the computing requirements, deep brain chain can execute distributed computing based on the user's specific circumstances and it is for different users. The DBC project is for global AI computing resource sharing and resource scheduling because many small businesses do not have the money to buy expensive GPU servers but many companies have a large number of GPU servers which are idle. Scheduling global resources and increasing the utilization efficiency of resources are of the positive significance. They also consider optimizing the distributed AI calculation engine and working with the underlying neural network chip makers. However, our focus is still on reducing AI business computing costs, rather than enhance the AI model of any single business model training.

Also, remember biz isn't your friend, people only post things that will benefit them. Warning people with this extent out of kindness would be ridiculous, he only does it to benefit by keeping the price down with utter shit posts.

https://www.reddit.com/r/DeepBrainChain/comments/7sdlej/why_you_wont_find_anything_related_to_the_company/

DBC is working with the Chinese government. DYOR and don't fall for the FUD. Everyone on biz has a fucking PhD in Physics or work in AI company. Fishy, isn't it?

Anonymous Mon Feb 19 18:07:35 2018 No.7801239

>>7801145
DBC their video states there have been many small start up trying to utilize AI. The growth has been tremendous in the past 5 years. And we all know Harvard faggots don't get any funding to buy GPUs to do more research, so they need to get DBC in order to get more computing power.

There's shit load of computers and GPU sitting idle. If a start up needs AI computational shit, get dbc and use those guys idle computing power to get the shit they need.

>>	Anonymous Mon Feb 19 18:08:15 2018 No.7801260 FUCK YOU WAT DOES THIS CREEPY IDIOT SAY?? DOESNT EVEN NOW HOW BLOCKCHAIN TECHNOLOGY WORKS MAYBE YOU SHOULD LEARN WHAT A BITCOIN IS BEFORE YOU OPEN YOU STUPID MOUTH BANKS IS SCARED AND CRYPTO WILL BE WORTH MILLIONS IN 2019 IDIOT YOU NOW NOTHING !!!!

Anonymous Mon Feb 19 18:08:18 2018 No.7801262

>>7801155
There isn't really much to discuss. The team claim they can lower the price requirements for computational power needed for artificial intelligence. They have good partnerships. DBC was hit hard by the last BTC crash. Binance have started listing NEP-5 tokens. Now is a good time to buy.

>>	Anonymous Mon Feb 19 18:09:37 2018 No.7801305 >>7793111 Harvard faggot can't get funding to buy more GPUs for his AI research. So he's spreading FUD, so he can pick up DBC cheaply, and get more computing requirements for his research.

>>	Anonymous Mon Feb 19 18:12:12 2018 No.7801376 >>7801262 I'd like to hold this token until at least around the summer. The investment from NEO is the main thing keeping me positive.

>>	Anonymous Mon Feb 19 18:12:45 2018 No.7801392 >>7801224 So if they aren't solving the logistical problem of distributed computing, how can they ever hope to reach the point of having huge amounts of trained data to sell ?

Anonymous Mon Feb 19 18:14:19 2018 No.7801441

>>7801178
>>>7801116 (You)
>>is what the CEO mentioned in his past attempt to FUD this project.
>
>Why would the CEO try to FUD his own project?

Harvard faggot posted this everywhere. The CEO responded to him (Harvard faggot trying to spread FUD). Not CEO spreading FUD.

>>	Anonymous Mon Feb 19 18:16:49 2018 No.7801521 >>7801392 People set up AI requirements and computing needs, and the idling computers or people with the data can respond to those AI requirements on the Blockchain.

Anonymous Mon Feb 19 18:20:35 2018 No.7801636

>>7801020

You don't even know the lingo you're trying to use.

>It's pretty clearly explained. A decentralized neural network.

That's your explanation? Thanks bud.

You do realize that people with these "idle GPUs" are either

1) unaware they could be mining
2) found that mining was no longer profitable

DBC would need to compete with all the other coins for computational power.

The real issue here is that you seem to think

>A decentralized neural network

can be isolated from the training algorithm. You can't just "use" the idle GPU's of the world all at once for massive massive massive data sets.

You don't seem to realize that massive data sets are heavy. You can't just "mindlessly" POW. You've got to iterate through massive data sets.

>The first and foremost the deep brain chain tackles with is to reduce the high-performance machine computing power costs and data privacy. In other words, how to lower the cost of neural network training and neural network computing. When an AI product has a large number of users it needs a lot of high-performance machines to meet the computing requirements, deep brain chain can execute distributed computing based on the user's specific circumstances and it is for different users. The DBC project is for global AI computing resource sharing and resource scheduling because many small businesses do not have the money to buy expensive GPU servers but many companies have a large number of GPU servers which are idle. Scheduling global resources and increasing the utilization efficiency of resources are of the positive significance.

This entire chunk of text you pasted is describing the computational power required to do backpropagation on data.
The data privacy issue is an awesome problem to tackle.

>>	Anonymous Mon Feb 19 18:23:42 2018 No.7801730 >>7801521 >People set up AI requirements and computing needs, and the idling computers or people with the data can respond to those AI requirements on the Blockchain. Wow. Just wow. You've really jampacked that sentence with a lot of nonsense.

>>	Anonymous Mon Feb 19 18:25:32 2018 No.7801780 >>7801636 Why would dbc need to compete with other coins to provide computational requirement? Are you referring to other AI coins?

Anonymous Mon Feb 19 18:26:27 2018 No.7801803

>>7801262

>There isn't really much to discuss. The team claim they can lower the price requirements for computational power needed for artificial intelligence. They have good partnerships. DBC was hit hard by the last BTC crash. Binance have started listing NEP-5 tokens. Now is a good time to buy.

The more you distribute your computational power the more inefficient it becomes. You're becoming increasingly less efficient the more you decentralize.

Anonymous Mon Feb 19 18:29:44 2018 No.7801885

>>7801780
I think you're seriously confused.

You're seriously asking me why DBC will need to compete with other coins/tokens?

You said
>There's shit load of computers and GPU sitting idle.

I said
>You do realize that people with these "idle GPUs" are either
>1) unaware they could be mining
>2) found that mining was no longer profitable

If you have lots of "GPUs sitting idle," there is a reason for it. There are plenty of coins to mine with idle GPUs...

Anonymous Mon Feb 19 18:31:40 2018 No.7801938

>>7801803
OK I'll do an AI start up, and raise a shit load off money to buy GPUs or whatever to get my AI product up and running then. Problem is nobody going to give me the money to buy these expensive GPUs.

You going to find me Mr. Harvard with limited funding?

Or is it more cost effective to borrow GPU computing cost? And save money?

>>	Anonymous Mon Feb 19 18:42:21 2018 No.7802238 >>7799385 >GAY id lmao faggot

>>	Anonymous Mon Feb 19 19:00:25 2018 No.7802709 >>7801938 Dear god, you're a special kind of stupid aren't you? You hardly even have a tentative grasp of what a GPU is.

>>	Anonymous Mon Feb 19 19:10:16 2018 No.7803002 >>7802709 GPU Server is used to train and compute in AI such as image, voice, NLP computation and do complex science computation such as in medicine and astronomy. Straight from their AMA. Faggot. Go read their AMA and also their new videos.

>>	Anonymous Mon Feb 19 19:11:25 2018 No.7803036 There are some fucking stupid keys in this thread, goddamn

>>	Anonymous Mon Feb 19 19:14:19 2018 No.7803131 >>7803002 Read their AMA and YouTube interviews. They said they expect their existing clients to buy the dbc coins Q1 2018.

Anonymous Mon Feb 19 19:17:06 2018 No.7803220

>all problems must be massive

Plenty use cases are even solved on a single pc.
There should be enough space in between we need google's data center and we run on a laptop. Especially in cases where your issue is training as you just need more processing power to process the raw data through your trrianed nn to classify new data. In that case distribution of data for classification becomes a non-issue unless you need to protect the data itself.

Anonymous Mon Feb 19 19:19:02 2018 No.7803274

>>7803220
fml cant write.

>where your issue is training as you just need more processing power to process the raw data through your trrianed nn
Should be
where your issue isn't training as you just need more processing power to process the raw data through your trained neural net

>>	Anonymous Mon Feb 19 19:24:42 2018 No.7803416 >>7801885 Yo could ipfs/ipdb on a delegate controlled network be utilized?

>>	Anonymous Mon Feb 19 19:25:49 2018 No.7803433 What i have learned from browsing /biz/ is: never listen to /biz/.. If i think there may be something to it then i do my own research. OP probably has bags in some other AI coin

>>	Anonymous Mon Feb 19 19:29:31 2018 No.7803518 >>7793111 Nice post OP. What do you think of SingularityNET?

>>	Anonymous Mon Feb 19 19:56:49 2018 No.7804159 >>7803416 First interesting comment. I'm not sure, but this is definitely something someone above my paygrade should look into.

Anonymous Mon Feb 19 20:00:09 2018 No.7804232

>>7803274
>where your issue isn't training as you just need more processing power to process the raw data through your trained neural net

This makes sense. At the same time, deploying an Amazon cloud instance (or even a Vultr instance) takes a few seconds and is incredibly cheap. Sure, it doesn't come pre-built with the neural net and data, but neither does some random dude's retired eth mining rig.

Advanced search
Text to find
Subject [?]Search by post subject. Leave empty for any.
Username [?]Search for user name. Leave empty for any user name.
Tripcode [?]Search for tripcode. Leave empty for any.
Email [?]Search by email. Leave empty for any.
Filename [?]Search by image filename. Leave empty for any.
From Date [?]Enter what date to start searching from. Format is YYYY-MM-DD
To Date [?]Enter what date to start searching until. Format is YYYY-MM-DD
Image hash
Search in	All Posts OPs Only
Deleted posts	Show all posts Show only deleted posts Only show non-deleted posts
Internal posts	Show all posts Show only internal posts Show only archived posts
Order	New posts first Old posts first
Capcode	All Posts Only by Users Only by Mods Only by Admins Only by Developers
Results	Posts Threads
Action	[ Simple ]

/biz/ - Business & Finance